EFFICACY OF NEUROFEEDBACK FOR CHILDREN WITH HISTORIES OF

ABUSE AND NEGLECT: PILOT STUDY AND META-ANALYTIC

COMPARISON TO OTHER TREATMENTS

Lark Huang-Storms, B.S., M.S.

Dissertation Prepared for the Degree of

DOCTOR OF PHILOSOPHY

UNIVERSITY OF NORTH TEXAS

August 2008

APPROVED:

Eugenia Bodenhamer-Davis, Major Professor

Jerry McGill, Co-Chair
Cecilia Thomas, Committee Member
Rafael Toledo, Committee Member
Linda Marshall, Chair of the Department of
Psychology
Sandra L. Terrell, Dean of the Robert B.
Toulouse School of Graduate Studies
 Huang-Storms, Lark, Efficacy of neurofeedback for children with histories of abuse

and neglect: Pilot study and meta-analytic comparison to other treatments. Doctor of

Philosophy (Health Psychology and Behavioral Medicine), August 2008, 134 pp., 1 table, 8

illustrations, references, 174 titles.

This two-part study investigates the effectiveness of neurofeedback training for

reducing behavioral problems commonly observed in abused/neglected children, and

compares its efficacy to other treatment interventions with this population. Neuro-

developmental sequelae of early relationship trauma are explored as an etiological

framework for understanding disturbed affect-regulation, which appears central to the

behavioral and emotional difficulties commonly experienced by this pediatric population. It

is suggested that neurofeedback teaches children to self-regulate brain rhythmicity

mechanisms, which in turn affects global improvements in behavior and mood.

The pilot study utilizes records of 20 children removed from their biological homes

by Child Protective Services. Children were assessed prior to treatment using the Child

Behavior Checklist (CBCL) and the Test of Variables of Attention (TOVA), and again after

30 sessions of individualized, qEEG-guided neurofeedback training. A t-test analysis of

pre- and post-scores was computed, and indicated significant improvements following

treatment.

A meta-analysis of existing literature on treatment interventions with

abused/neglected children provides individual and aggregate effect sizes for 33 outcome

studies with this clinical population, and contextualizes the results of the present pilot study

within other empirically validated treatment modalities. Establishment of an overall effect

size for treatment for this pediatric population provides a needed method of comparing

research results across studies when control groups may not be ethical or feasible.
 Copyright 2008

by

Lark Huang-Storms

ii
 ACKNOWLEDGMENTS

Thank you to Dr. Genie Bodenhamer-Davis, who generously supported me in my

research interests and who continuously models a spiritual and curious approach to life

and learning.

Thank you to my husband, Tim, and to my son, Avery, for the daily kindnesses

and family escapades that have allowed this project to be a balanced part of our lives.

Thank you to my parents, who encourage me to compose a life on my own terms,

and for all the practical help along the way that has made the ride smoother.

Finally, thank you to all the young people Ive been privileged to know who inspire

me with their courageous growth in the face of adversity.

iii
 TABLE OF CONTENTS

Page
ACKNOWLEDGMENTS..iii

LIST OF TABLES.vi

LIST OF ILLUSTRATIONS.vii

Chapter

1. INTRODUCTION...1

2. EFFECTS OF EARLY SOCIAL TRAUMA ON NEURODEVELOPMENT....4

Attachment Relationships and the Development of Self-Regulation.......4

Neurophysiological Impact of Early Relationship Trauma..5
Animal Research on Neurodevelopmental Trauma....9
Behavioral Manifestations of Relationship Trauma...11

3. TREATMENT APPROACHES FOR CHILDREN WITH HISTORIES OF

ABUSE/NEGLECT..14

Group Therapy.14
Family Therapy15
Play Therapy18
Cognitive Behavioral Therapy...20
Neurofeedback with Pediatric Populations..21
Models Supporting Neurofeedback for Maltreated Children....22

4. PILOT STUDY................25

Method..25
Participants......25
QEEG Profiles..... 25
Instruments.......26
Procedures....28
Results..30
Child Behavior Checklist30
Test of Variables of Attention32
Discussion34

5. META-ANALYSIS...38

Method..39
Definition of Terms..39

iv
 Literature Search.....40
Criteria for Study Inclusion.41
Effect Size Computation and Analyses...44
Results..48
Participant Characteristics: Type of Abuse History...49
Participant Characteristics: Age...52
Participant Characteristics: Gender, Ethnicity, and Socio-economic
Status55
Study Characteristics: Empirical Design.56
Study Characteristics: Length of Treatment...61
Study Characteristics: Treatment Intervention...62
Summary..65

6. DISCUSSION......67
Findings Compared to Previous Research.67
Clinical Implications for Neurofeedback..69
Study Strengths and Limitations...71

APPENDIX A: INTERNATIONAL 10-20 ELECTRODE PLACEMENT SYSTEM.....75

APPENDIX B: SUMMARY TABLES OF STUDIES INCLUDED IN META-ANALYSIS77

REFERENCES......114

v
 LIST OF TABLES

Table Page

1. Number of studies included at initial and final stages of literature review....41

vi
 LIST OF ILLUSTRATIONS

Figure Page

1. Averaged raw scores on the CBCL total, externalizing, and internalizing scales....31

2. Averaged raw scores on the CBCL syndrome scales..32

3. Averaged TOVA standard scores....33

4. Effect size means for studies grouped by type of abuse..52

5. Effect size means for studies grouped by age...53

6. Effect size means for studies grouped by gender.....56

7. Effect size means for studies grouped by design..60

8. Effect size means for studies grouped by type of treatment....63

vii
 CHAPTER 1

INTRODUCTION

Each year, hundreds of thousands of children in the United States suffer some

form of childhood maltreatment. Actual figures vary, depending on how the

phenomenon is defined and the nature of the research. However, according to the most

recent federally mandated National Incidence Studies (NIS-3), approximately 1,553,800

children were abused or neglected in 1993 in the United States (Sedlack & Braodhurst,

1996). This represents a 67% increase since the NIS-2 findings in 1986 and a 149%

increase since the NIS-1 in 1980. The NIS-4 analysis is currently underway, and

maltreatment estimates are expected to again increase.

The repercussions of these numbers extend far beyond the psycho-social and

physiological injuries experienced by the direct victims of childhood maltreatment.

Indeed, child maltreatment represents a society-wide tragedy with profoundly negative

intergenerational effects on quality of life issues such as education, crime, and

economics. It is remarkable that it was only as recently as 1968 that all fifty states

adopted laws requiring health professionals to report suspected cases of abuse and

neglect (ten Bensel, Rheinberger, & Radbill, 1997). As recently as 1974, the National

Center on Child Abuse and Neglect was formed, and the Child Abuse Prevention and

Treatment Act was passed by the U.S. Congress (English, 1998).

Although theoretical knowledge has developed substantially over the past

decades with regard to treatment for abused and neglected children, little of this

information has been critically evaluated through rigorous research designs. It is evident

that the process of empirically validating treatments for this pediatric population is in its

1
early stages. Few studies have employed randomized control groups, largely due to

ethical concerns over delaying treatment for such a vulnerable population (Reeker,

Ensing & Elliott, 1997), and only a minority has utilized standardized outcome measures

and large enough sample sizes for meaningful statistical analysis. Furthermore, until the

last decade, most treatment interventions designed to address child maltreatment were

directed at interrupting maladaptive parenting practices, without addressing the abuse-

related sequelae for the child (Kaplan, Pelcovitz, & Labruna, 1999).

The quality and quantity of outcome research on the subject of therapeutic

interventions aimed at children as a whole pales in comparison to outcome research for

adults. There are, however, several hundred efficacy studies of psychosocial therapies

for children, applying different modalities and methodologies to various clinical

populations. At least four meta-analyses have been completed within the past twenty

years that encompass diverse therapeutic treatment methods for diverse

child/adolescent problems. These analyses indicate that the overall impact of child

psychotherapy is positive, with effect sizes ranging from a medium Cohens effect size

of .54 (Weisz, Weiss, Han, Granger, & Morton, 1995) to a large Cohens effect size of

.88 (Kadzin, Bass, Ayers & Rogers, 1990).

Children with histories of abuse and neglect are frequently described by child

welfare professionals as especially difficult to treat therapeutically (Chanitz, 1995;

Wilson, 2001). This is likely due to the fact that abuse and neglect are not distinct

disorders or syndromes, but rather experiences. The patterns and extent of the

symptoms have major variations, are developmentally dynamic, typically involve co-

morbid diagnoses, are behaviorally entrenched in family dynamics, and may be

2
susceptible to sleeper effects (Finkelhor & Berlinner, 1995). Thus, an overarching aim

of the present study is to establish whether treatment outcome data for this particular

population are as robust as those for the broader population of children. Specifically, the

present pilot study explores the potential of neurofeedback training to help children

recover from abuse and neglect, and contextualizes its efficacy within a meta-analysis

of other empirically substantiated treatment modalities for the same population.

3
 CHAPTER 2

EFFECTS OF EARLY SOCIAL TRAUMA ON NEURODEVELOPMENT

Attachment Relationships and the Development of Self-Regulation

Over thirty years ago, Bowlby asserted the central importance of early caretaker

relationships on the social and emotional development of children. His attachment

theory suggested that an infants ability to cope with stress is correlated with biologically

driven mother-child behavior patterns that promote primary caretaker proximity (Bowlby,

1969). Ainsworth, Blehar, Waters and Wall (1978) expanded this theory by defining four

infant attachment styles and contributing a greater understanding of the purpose behind

the primary caregivers role: to provide a secure base from which a child explores

his/her surroundings and incorporates internal working models of trusting relationships.

This seminal work was followed by a wealth of empirical research demonstrating a

correlation between level of attachment security and the development of a wide range of

psychopathology in children and adults, including mood, personality, conduct, and

anxiety disorders (Crittenden, 1995; Schore, 1994).

Attachment behaviors serve important protective functions beginning at birth, and

are believed to correspond with the onset of independent locomotion in vertebrates

(Clutton-Brock, 1991). Increasing evidence suggests that a primary function of the

attachment relationship is to develop a childs ability to self-monitor affect, self-regulate

physiological arousal level, and self-organize coping functions for stress (Cassidy, 1994;

Cicchetti, & Tucker, 1994). Kopp (1989) asserts that the development of affect

regulation proceeds from initial reliance on a caregiver, to self-soothing behaviors, and

finally to language based cognitive coping strategies. The central role of attachment in

4
the development of self-regulation may explain why relational trauma from the social

environment has been shown to have more negative impact on the rapidly developing

brains of infants and children than assaults from the inanimate physical environment

(Schore, 2001; Sgoifo, Koolhaas, & De Boor, 1999).

Because infants and young children are unable to effectively modulate affective

and physiological arousal states independently, their developing capacity to cope with

dysregulated states depends on the responses of caregivers who are psychobiologically

attuned to their needs (Schore, 1994). Caretakers externally manage infants

psychophysiologic states by responding to them in consistently sensitive ways, for

example through accurate mirroring of affect and sensitivity to gaze aversion as a signal

of over-stimulation (Field, 1994). Through such processes, children learn to develop

strategies to manage high arousal and regain a state of organization when homeostasis

has been disrupted. Research by Goldberg, MacKay-Soroka, and Rochester (1994)

showed that mothers in securely attached relationships with their infants responded

equally to their babies positive and negative affects, whereas mothers in poorly

attached relationships responded predominantly to negative affects (therefore

conditioning their children to increase negative affective behaviors). Thus, strategies

developed early in life to manage arousal can be understood as forming the

neurodevelopmental building blocks of lifelong personality structure and affective

behavior patterns (Bradley, 2000).

Neurophysiological Impact of Early Relationship Trauma

When caregivers are neglectful, inconsistent, or abusive, infants and young

children are left vulnerable to psychophysiological distress states from which they

5
cannot escape. Two interacting response patterns have thus far been identified in

children. The first is a hyperarousal (fight or flight) response, which is mediated by

sympathetic activation of the limbic hypothalamic-pituitary-adrenal (HPA) axis, and

results in increased levels of cortisol, acetylcholine, adrenaline, and noradrenaline

within the developing brain. The second response, more common in girls and younger

children/infants, is the dissociative continuum, which is mediated by the

parasympathetic activation of dorsal vagal responses (i.e., reduced metabolism, heart

rate, and respiration rate), endogenous opioids, and the dopaminergic system (Perry &

Pate, 1994). The hyperarousal and dissociative responses to stress are not discreet;

rather, when stressful situations are perceived to be hopeless or overwhelming, initial

sympathetic arousal may be followed by disengagement from external stimuli via

parasympathetic activation of the vagal and opioid systems (Perry, 1994).

Prolonged hyperarousal/dissociative states can chronically dysregulate a childs

psycho-physiological stress-response systems (e.g., the HPA axis) and patterns of

coping behaviors (e.g., withdrawal or aggression; Manassis & Bradley, 1994; Post,

Weiss, & Leverich, 1994). Chronic dysregulation is associated with sensitization of the

brain stem and midbrain neurotransmitter systems, such that early adverse attachment

experiences essentially kindle the limbic areas of the brain to be physiologically

reactive (Post, Rubinow, & Ballenger, 1984). In particular, van der Kolk and Greenberg

(1987) have suggested that the repeated trauma of child abuse may dispose the stress-

sensitive amygdaloid nuclei to develop a kindling response, by which repeated

intermittent stimulation produces increasingly greater alterations in neuronal excitability,

potentially resulting in seizures. Othmer, Othmer, and Kaiser (1999) describe the

6
kindling process as a practice effect of the brains successive experiences of

dysfunction or overarousal, and suggest that vulnerable physiologic feedback systems

tend to become more dysregulated over time when left alone. Work by Adamec and

Stark-Adamec (1989) demonstrates that kindling in the amygdala induces a defensive

personality in domestic cats, the intensity of which is mediated by both experience and

strength of neurotransmission between the basomedial nucleus of the amygdala and

the ventromedial nucleus of the hypothalamus. Because of the dense interconnections

amongst the prefrontal cortex, hypothalamus, amygdala, thalamas, cingulate gyrus, and

hippocampus, it appears likely that various patterns established through early

attachment experiences could set a precedent for the development of relatively

automatic affect-regulating mechanisms within these feedback systems of the brain

(Schore, 1994; Bradley, 2000).

In relational trauma, the developing limbic system is repeatedly exposed to high

levels of excitotoxic neurotransmitters, such as glutamate, cortisol, and NMDA-sensitive

glutamate receptor, all of which are associated with neurotoxicity and abnormal synapse

elimination in early brain development (Choi, 1992; Moghaddam, Bolinao, Syein-

Behrens, & Sapolsky, 1994). It is hypothesized that stress-induced increases in

gluccocorticoids in postnatal periods selectively induce neuronal cell death in the limbic

system and impact abnormal limbic circuitry. In particular, there is ample evidence that,

in adults, the cellular organization of the hippocampus can be dramatically affected by

levels of corticosteroids, which can exert deleterious effects on the hippocampal

pyramidal cells (Sapolsky, 1993; Watanabe, Gould, & McEwen, 1992). In separate

studies, Carrion et al. (2001) and De Bellis et al. (1999) found that children with histories

7
of trauma and post traumatic stress disorder (PTSD) symptoms had significantly smaller

total brain cerebral volume than matched control groups. In both studies, after

statistically controlling for total brain volume, no significant decreases in hippocampal

volumes were found in the PTSD child population, suggesting a more generalized effect

of the early developmental neurotoxic effects of glucocorticoids.

According to Schore (2001), attachment experiences in infancy particularly

influence the experience-dependant maturation of the right orbitoforntal cortex, which is

dominant for the processing of affect-regulation, visual emotional information, and

attachment experiences. During the first few months after birth, the right hemisphere

develops more rapidly than the left, which theoretically makes it more vulnerable to the

consequences of extreme stress and neglect (Galaburda, 1984). Furthermore, Read,

Perry, Moskowitz, and Connolly (2001) propose that abuse from 6 months until 3 to 6

years of age may have the greatest differential effect on the left hemisphere. Their

findings are supported by the work of Teicher et al. (1997), which suggests that

dendritic growth in the left hemisphere surpasses that of the right hemisphere at about 6

months of age.

Nonspecific electroencephalogram (EEG) abnormalities have been found in

populations of abused children, including psychologically abused children and physically

abused children without head trauma (Green, Voeller, & Gaines, 1981; Teicher et al.,

1997). Van Bloems (2000) work with children and adolescents diagnosed with reactive

attachment disorder (RAD) revealed specific patterns of EEG slowing in the frontal

lobes and right temporal lobe. Fisher, Turber, and Gunkelman (2005) also reported right

temporal as well as vertex slowing in quantitative electroencephalogeams (QEEGs) of

8
children with RAD. Ito, Teicher, Glod, and Ackerman (1998) observed childhood

physical and sexual abuse to be associated with an increased prevalence of left-sided

EEG abnormalities (particularly fronto-temporally). In terms of cortical EEG coherence,

Teicher et al. (1997) found that a group of sexually and physically abused children had

greater average left hemisphere hypercoherence than normal children, but comparable

right hemisphere coherence patterns (indicating diminished left hemisphere

differentiation in the abused group). Their work highlights the relationship between

limbic system dysfunction and reversed left/right hemispheric asymmetry, asserting the

possibility that early abuse may impede hemispheric integration and the establishment

of normal left cortical dominance. In contrast, in a small clinical sample of women,

Black, Hudspeth, Townsend and Bodenhamer-Davis (2002) found histories of childhood

sexual abuse to be associated with hypocoherence in left frontal regions in the theta

and beta bands and hypercoherence in posterior central regions across all bands;

however these findings were not wholly replicated with a high functioning non-clinical

college sample of women with sex abuse histories (Black, 2005).

The question naturally arises whether these brain abnormalities may be at least

partially the result of genetic factors, intergenerational effects of parenting, or learned

stress-coping behaviors. Indeed, in addition to evidence of intergenerational

transmission of child abuse in humans (Kaufman & Zigler, 1989), there is evidence in

group-living pigtail macaques of genealogical and demographic influences on maternal

neglect and abuse of offspring (Dario, Wallen, & Carroll, 1997). Because it is ethically

difficult to design studies to tease apart these hypotheses, there has not been definitive

research to clarify these questions. However, there are a number of studies of

9
neglected and abused children in orphanage settings that have found dramatically

smaller frontal-occipital head circumferences (38% below the third percentile), as well

as computed axial tomography (CT) and magnetic resonance imaging (MRI) findings of

enlarged ventricles and cortical atrophy in this population (Perry & Pollard, 1998; Rutter

et al., 1998). Using functional magnetic resonance imaging (fMRI) with a population of

Romanian orphans, Chungani et al. (2001) found these children had decreased

metabolic activity in the orbital frontal gyrus, intralimbic prefrontal cortex, amygdala,

hippocampus, lateral temporal cortex, and brainstem.

Animal Research on Neurodevelopmental Trauma

Animal research with rats provides further support for the neurodevelopmental

impact of neglect/physical abuse, specifically on hemispheric laterality (Denenberg,

1983), hippocampal shrinkage (Meaney, Aitken, van Berkel, Bhatnagar, & Sapolsky,

1988), and alterations of neuro-endocrine stress response systems (Fride, Dan, Feldon,

Halevy, & Weinstock, 1986). Research conducted with chimpanzees and gorillas

(Davenport & Rogers, 1970) and with rhesus monkeys (Harlow & Harlow, 1965), though

not neurophysiological in focus, demonstrates the profoundly negative behavioral

outcomes of severe social deprivation during the first year of life (similar to behaviors

seen in neglected and abused children). Intriguingly, there is even evidence that

maternal neglect in invertebrate wolf spiders results in decreased central nervous

system (CNS) development, as observed in decreased brain weight and number of

brain cells (Punzo & Ludwig, 2002). Spiderlings removed from their mothers also show

less ability to hunt and learn maze navigation, as compared to spiderlings who remain

with their mothers.

10
 Recent observations of elephant behavior in South Africa, India, and Southeast

Asia have brought international attention to the dramatic increase in incidents of juvenile

elephant attacks on villages and elephant killings of humans and rhinoceroses (Siebert,

2006). Ethologist Gay Bradshaw and colleagues describe contemporary elephant

populations as suffering from chronic stress due to decades of poaching and habitat

loss, which have disrupted the intricate web of familial and societal relationships

governing the rearing of young elephants (Bradshaw, Schore, Brown, Poole, & Moss,

2005). Like humans, elephants are known for their close social relationships, and young

elephants are traditionally raised in a matriarchal society that includes multi-

generational allomothers. Bradshaw et al. have observed that calves witnessing culls

or raised by isolated mothers are high risk candidates for inability to regulate stress-

reactive aggressive states and for an array of behavior patterns similar to human PTSD

(i.e., abnormal startle reflex, depression, asocial behavior, and hyper-aggression). This

appears to be especially relevant for male calves, who traditionally go through a second

phase of all-male group socialization and neurodevelopment during adolescence. The

critical role of older males in their normal social development was demonstrated when

researchers re-introduced older bulls to successfully quell male juvenile hyper-

aggression.

Behavioral Manifestations of Relationship Trauma

Neurodevelopmental research has established that, because of the brains

extreme malleability and sensitivity to experience in early childhood, traumatic events in

the first few years of life can have long-term impacts on socio-emotional and cognitive

functioning. This is particularly likely if events are severe, unpredictable, or ambient all

11
of which describe relational trauma in the form of neglect or abuse (Perry, 1994). Child

neglect and abuse have been shown to have an etiological role in a remarkable range

of behavioral disorders affecting children, including attention deficit disorders, mood and

anxiety disorders, conduct and oppositional defiant disorders, RAD, learning disabilities,

PTSD, eating disorders, substance abuse, and dissociative disorders (Beitchman et al.,

1992; Boney-McCoy, & Finkelhor, 1995). Although estimates vary widely, some

researchers propose that up to 80% of abused children display symptoms of severe

attachment disturbance or RAD (Hall & Geher, 2003).

Numerous studies demonstrate that children who have been mistreated or have

had multiple, inconsistent caregivers are unlikely to develop secure attachment styles

(Egeland & Sroufe, 1981). Insecure or disorganized attachment (George & Main,

1979) is suggestive of a lack of pattern/strategy for regulation of affect; thus it is not

surprising that children with histories of disrupted attachment commonly experience

externalizing problem behaviors, sleep and eating irregularities, and attentional

difficulties indicative of poor self-regulation functions (DeGangi, 2000). Behaviorally,

limbic dysregulation may inhibit a childs capacity to cope with stressors by maintaining

heightened arousability (e.g., a child may rapidly escalate from feeling slight anxiety to

terror) and supporting chronically heightened arousal states (e.g., a child may not

maintain a focused state for academic learning due to hypervigilance to threat). Foster

and adoptive parents and professionals at Child Protective Services (CPS) commonly

voice concerns that these children demonstrate peer aggression, stealing, food

hoarding/gorging, destruction of property, poor impulse control, limited cause-and-effect

thinking, inappropriate sexual behavior, school failure, and hyperactivity (Iwaniec,

12
1995). Difficulty forming trusting relationships, indiscriminate affection with strangers

concomitant with refusal to give affection to family members, limited ability to

empathize, and poor social skills are also prevalent in this population of children,

particularly those diagnosed with RAD (Hall & Geher, 2003).

Child maltreatment research suggests that the developmental and psychosocial

consequences of childhood abuse/neglect have implications across the lifespan. Early

maltreatment is associated with adult involvement in coercive relationships (Malamuth,

Sockloskie, Koss & Tanaka, 1991), diagnoses of personality disorders (Barach, 1991),

and high incidences of adult depression, anxiety, and suicidality (Brown & Smails, 1999;

Dube et al., 2001). A longitudinal study of victims of childhood maltreatment by Widom

(1999) found childhood experiences of abuse and neglect to be strong predictors of

PTSD. Specifically, Widoms interviews with victims approximately twenty years

following substantiated maltreatment revealed that 37% of childhood sexual abuse

victims, 33% of physical abuse victims, and 31% of neglect victims met criteria for

PTSD as defined by the revised third edition of the Diagnostic and Statistical Manual of

Mental Disorders (DSM-III-R).

13
 CHAPTER 3

TREATMENT APPROACHES FOR CHILDREN WITH HISTORIES OF

ABUSE/NEGLECT

Treatment modalities for maltreated children fall into five primary categories

group/peer therapy, play therapy, cognitive-behavioral therapy (CBT), family therapy,

and experimental therapies (such as eye movement desensitization and reprocessing

[EMDR], neurotherapy, and holding therapy). Sometimes, children in research sample

groups are involved in multiple intervention modalities (e.g., both group play therapy

and family therapy). This complicates the interpretation of discrete treatment outcome

data, and was handled in the present meta-analysis by creating a separate category of

treatment, i.e., multi-modal or combined therapy.

Group Therapy

Based on the quantity of literature on the subject, group therapy appears to be

the most popular treatment approach for working with sexually abused children and

adolescents, in particular. Group therapy is distinctive in that it centrally addresses two

hallmark effects of abuse and neglect feelings of isolation and social stigmatization

while also being cost-effective and labor conserving (Silovsky & Hembree-Kigin, 1994).

Additionally, group therapy can decrease resistance to treatment through mutual self-

disclosure and fostering of a positive peer milieu (Carozza & Heirsteiner, 1982). Groups

for this particular population may be open or closed, and may take a process-oriented

approach or, more commonly, a structured psycho-educational approach.

Issues frequently encountered in groups for maltreated children include feeling

damaged, feeling guilty and responsible, feeling fearful and betrayed, feeling

14
depressed, and experiencing low self-esteem (Porter, Blick, & Sgroi, 1983). Other

common group themes for this population are social and communication skills,

separation and abandonment difficulties, blurred role boundaries, and inappropriate

sexual behaviors (Gil, 1992). Alternatively, some groups de-emphasize abuse themes

and instead aim to provide normalizing social experiences for participants (Borgman,

1984). Because trust forms the foundation for willingness to self-disclose, trust-building

exercises are generally introduced early in group treatment (Grammer & Shannon,

1992).

Psycho-educational components of group treatment for maltreated children

frequently include stress management, anger management, assertiveness training,

safety, appropriate touch, and individualized coping/problem solving strategies. A

parallel group for non-offending parents is often advocated (Damon & Waterman, 1986).

It is not uncommon that parents have their own past abuse issues that re-emerge when

they become aware their children have been victimized. In particular, mothers of

children who have been sexually abused may have emotional reactions that mirror their

childrens, such as fear, anger, and feelings of over-responsibility (Hagans & Case,

1988). Parallel treatment for non-offending parents is also believed to increase parental

commitment to therapy and prepare parents for changes in their childs behavior, such

as increased assertiveness (Silvosky & Hembree-Kigin, 1994).

Family Therapy

Family therapy uniquely offers a systems-based approach to the treatment of

abused and neglected children, whereby family members are involved in exploring the

significance of individual roles and interpersonal dynamics within the family that

15
contribute to the mental health of the child (Goldenberg & Goldenberg, 1996). The

family systems model considers the child to be the symptom bearer of dysfunction in the

family system. As such, the focus of intervention is the family unit rather than the

individual child (Wagner, 2003). In highly dysfunctional families, children are often

removed from the home, and intervention with the biological family is less pertinent than

family interventions with foster and adoptive families. As is frequently cited in the

literature, family therapy for maltreated children may consist of parent-child dyads

excluding the offending parent, either for legal, logistical, or safety reasons. Family

therapy is also often utilized as an adjunctive treatment to individual child therapy

interventions.

There is a great deal of diversity in terms of family therapy approaches and

interventions for maltreated children. Psychodynamic family systems therapy relates

intra-psychic functioning to interpersonal relationships to explain family members

behaviors within the context of the family system. Specifically, object relations family

therapists address transference issues and help the family develop nurturing holding

relationships (i.e., environmental extensions of an emotionally responsive maternal

presence) (Scharff & Scharff, 1987). Multigenerational family systems therapists

additionally focus on the transmission of family patterns over multiple generations

(Bowen, 1996) a perspective that is particularly relevant in the case of

intergenerational family violence and maladaptive parenting schemas. Bowen describes

how stressed families are composed of interlocking emotional triangles whose

composition changes as family members realign to reduce tension. Interventions are

aimed at disengaging from dysfunctional emotional triangulation by using such

16
strategies as taking personal responsibility for ones actions and using I statements to

express feelings without criticism.

Humanistic family therapists encourage family members to be genuine with each

other, to be honest about their feelings, and to resolve conflicts through effective

communication. There is an understanding that families operate through largely

unspoken family rules that govern family members behaviors and communication styles

(Satir, 1972). In abusive families, a typical unspoken family rule might be that one child

is the source of all problems encountered by the family. Families are encouraged to

examine their spoken and unspoken rules, and to take risks to negotiate rules that

enhance the self-worth of all family members. Dysfunctional families also frequently

demonstrate unhealthy communication styles, including double-bind messages that

leave the listener feeling confused (Bateson, Jackson, Haley, & Wakeland, 1968). Satir

(1972) observed that family members respond to stress in predictable ways, such as

placating, blaming, or distracting, which serve to maintain the familys unhealthy

homeostasis.

Structural family systems therapy shares a focus on family rules and patterns of

homeostasis, and focuses on boundaries and levels of authority within the family.

Minuchin (1974) described family boundaries as dictating who interacts with whom and

in what manner, and labeled them as either clear (i.e., healthy), diffuse, or rigid.

Families with diffuse or rigid boundaries sacrifice the members autonomy for the family

unit, resulting in either overly enmeshed relationships or overly detached ones.

Structural family therapy is present-oriented and, like humanistic therapy utilizes live

enactments to clarify family behavior patterns.

17
 Most clinical settings that treat the families of maltreated children practice time-

limited therapy that relies heavily on cognitive-behavioral, strategic, and solution-

focused family systems approaches (Wagner, 2003). These modalities adopt a

pragmatic approach in which the focus of treatment is the resolution of family problems,

usually without emphasis on insight into the causes behind the presenting concerns

(Sharf, 2000). Psycho-education, parent training, establishment of behavioral

contingencies, and therapy homework are typically involved. Solution-focused therapy

with families of maltreated children typically strives to restructure common dysfunctional

beliefs (e.g., the problem is destined to continue, or the problem is beyond the familys

control; De Shazer & Molnar, 1984).

Play Therapy

Play therapy provides a nurturing relationship between a child and a trained

provider in which therapeutically derived play activities occupy the central aspect of

treatment (Phillips, 1985). It may be directive, non-directive (i.e., client-centered), or filial

(i.e., involving training parents in therapeutic play). In all cases, the child has the

opportunity to have a corrective emotional experience with an adult who is positive and

consistent, and who allows the child to have an influential role in shaping the

relationship (Reams & Friedrich, 1994). The adult models the core conditions for

healthy relationships (i.e., congruence, acceptance, and understanding), fostering the

childs feelings of self worth and innate value. Through unconditional acceptance of the

child, careful selection of toys, and the use of reflective remarks, children are believed

to grow by reconstructing their world and managing emotional and cognitive material

through the safe medium of play (Landreth, 2002).

18
 Play therapy approaches for maltreated children vary widely, but several abuse-

specific approaches appear more commonly in the literature. Trauma focused play

therapy is a dynamic model of therapy that uses the medium of play to enhance

emotional mastery, lend insight to cognitive distortions, experience emotional catharsis,

and process abuse related material (Gil, 1998). Trauma-focused eclectic integrative

therapy advocates the use of play to enhance the childs insight, and addresses specific

issues of shame, self-blame, and understanding of the self in relationships (Friederich,

1995). This technique typically combines parent-child interactive therapy, in which

directed play activities are designed to correct cognitive distortions and process abuse-

related material. Landreth (2002) has suggested that play therapy may be particularly

beneficial in helping abused and neglected children deal with ambivalent feelings

toward the aggressor. Targeted feelings include fear and negativity, as well as positive

attachments toward the perpetrator. This state of ambivalence often brings about

anxiety in maltreated children, which play therapy may help alleviate.

Extensive research on the common play characteristics of abused and neglected

children has identified several behavioral patterns in this population. These include

developmental immaturity, oppositionality and aggression, withdrawal and passivity, self

deprecation and self harming, hypervigilance, age-inappropriate sexuality, and

dissociation (White & Allers, 2001). Recurrent play themes have also been observed in

the literature, such as unimaginative, inhibited, and literal play (Jacobson & Straker,

1982). Repetitious play with a compulsive dynamic is also frequently noted by clinicians,

including ritualized behaviors and mechanistic play behaviors that appear to serve

stress reducing functions for the child (Terr, 1981).

19
 Cognitive-Behavioral Therapy

Cognitive-behavioral approaches with children who have experienced abuse

and/or neglect tend to be individually-administered or shared between the child and a

non-offending family member. Typical goals of treatment include providing education,

teaching coping and safety skills, and conducting gradual exposure and processing

exercises to help the child overcome harmful, oppositional, or anxious/avoidant

behavior patterns stemming from maltreatment (Deblinger & Heflin, 1996). With very

young children, CBT interventions typically incorporate the use of play or expressive

arts as a medium of communication.

Treatments are based on the application of social learning principles and

cognitive theories, designed to reduce childrens negative emotional and behavioral

responses related to the abusive or neglectful situations. Trauma focused CBT

assumes that symptoms develop and are maintained at least in part by conditioned and

learned behavioral responses and maladaptive cognitions. Interventions are thus

designed to target trauma-induced thoughts, behaviors, and feelings through the use of

psychoeducation, gradual exposure, cognitive reframing, and stress management

(Cohen & Mannarino, 1993; Deblinger & Heflin, 1996). As evidenced in the above

descriptions of group, family, and play therapies, CBT interventions and approaches are

frequently incorporated fluidly into other treatment modalities.

The role of the child in cognitive-behavioral therapies is one of active

involvement. Ronen (1998) describes the child in cognitive-behavioral therapy as an

active partner who learns how she or he behaves and how to acquire knowledge in

order to change. Therapists impart this knowledge as they teach children how to use

20
techniques such as stress management or self-monitoring, but it is ultimately the child

who actually evaluates and reinforces the application of these strategies outside the

therapy session. Homework is typically an integral part of cognitive-behavioral

treatments because it gives children an opportunity to practice skills learned in therapy

sessions.

Neurofeedback with Pediatric Populations

Neurofeedback, or electroencephalographic (EEG) biofeedback, is among the

most promising modalities for the treatment of child and adolescent psychological

disorders. Interest in neurofeedbacks potential as a therapeutic intervention for

pediatric populations is heightened by concerns regarding the uncertain long-term

effects of psychiatric and stimulant medications on the developing brain (Wilens, 2004).

Furthermore, for children with histories of relational trauma, attachment disorders, or

reactive attachment disorder (RAD), conventional relationship-based therapies

(including cognitive therapies and play-based therapies) have demonstrated limited

clinical success (Chanitz, 1995; Wilson, 2001).

There has been increasing clinical and empirical support for the efficacy of

neurofeedback interventions for adolescent and child applications (Hirshberg, Chui, &

Frazier, 2005). The most substantial literature supports neurofeedback training to

decrease slow wave activity while increasing the power of 12-15 hertz (Hz) and 15-18

Hz activity for improvement of attention deficit hyperactivity disorder (ADHD) symptoms

in children (Fuchs, Birbaumer, Lutaenberger, Gruzelier, & Kaiser, 2003; Linden, Habib &

Radojevic, 1996; Lubar, Swartwood, Swartwood, & ODonnell, 1995; Monastra,

Monastra, & George, 2002; Rossiter, 2004; Rossiter & La Vaque, 1995). In addition, the

21
literature includes some clinical evidence that neurofeedback benefits children with

autism spectrum disorder (Jarusiewicz, 2002; Sichel, Fehmi, & Goldstein, 1995),

asthma (Tansey, 1992), pediatric stroke (Ayers, 1995), and pediatric migraine

(Siniatchkin et al., 2000). As of yet, no study has investigated the effectiveness of

neurofeedback with children and adolescents with histories of relational trauma, though

there are a few unpublished clinical reports of success with this pediatric population

(e.g., Fisher et al., 2005).

Models Supporting Neurofeedback for Maltreated Children

There is an increasing number of voices in the behavioral sciences asserting the

prominent role of arousal dysregulation in the development of psychopathology

(Grotstein, 1986; Le Doux, 1996). This has fostered a meaningful integration of

biological and psychosocial approaches regarding the conceptualization and treatment

of mental health and illness. Based on work by Gorman, Liebowitz, Fryer, and Stein

(1989), Bradley (2000) proposes a general model for the development of affect

regulation in which the pre-frontal cortex integrates experiential learning and cognitive

schemas with the output of the limbic and reticular activating systems. Similarly, Schore

(1994) describes a model of arousal regulation in which the orbitofrontal cortex sits at

the apex of a fluctuating sympathetic-parasympathetic autonomic nervous system

(ANS) balancing system. As outlined above, early experiences of relational trauma

produce deficits in the ability to regulate affect and manage arousal through processes

of limbic kindling, dysregulation of neurochemical ANS stress-response mechanisms,

and maladaptive internalized coping strategies. In addition, the fact that relational

trauma is a risk factor common to a wide array of psychological disorders points to the

22
centrality of physiological and emotional self-regulation as a key variable in

psychopathology.

Othmer et al. (1999) propose a dysregulation model of psychopathology based

on a failure of control systems and feedback loops in the brain, specifically in terms of

the regulatory EEG activity of the central nervous system (CNS). This model contributes

the addition of the electrical frequency domain to the traditional neurochemical

conceptualization of CNS functioning, and asserts that rhythmicity is the basis for the

organization of brain function. Rhythmic activity of the brain is produced by complex

interactions of feedback loops (or servosystems) that are potentially vulnerable to

setpoint errors, instabilities, and over/under-shooting in response to stimuli.

Thompson and Thompson (2003) describe three basic resonant cortical loops that

affect EEG rhythmicity (local, regional, and global), all of which can operate

spontaneously or may be driven by subcortical thalamic pacemakers. As explained by

Lubar (1997), changes in the patterned functioning of cortical loops (as a result of

learning, for example) can affect the intrinsic firing rate of thalamic pacemakers, which

is ultimately associated with changes in mental state.

As conceptualized by Sterman (1996) and by Othmer and colleagues (1999),

EEG biofeedback is a means of operantly conditioning the rhythmic electrical

manifestations of CNS regulatory function by challenging the brain to shift away from

unstable firing patterns toward more homeostatic ones. The addition of an external

feedback loop via biofeedback encourages the brain to alter its prevailing rhythmicity

through a repetitive process of imposed disequilibrium and attempted return to baseline.

This regulatory challenge model of neurofeedback facilitates improved ability to

23
maintain homeostasis and improved stability of the regulatory CNS system itself.

Although factors other than early relational trauma can disrupt normal brain

development and regulation patterns, it would appear that children with histories of

traumatic attachment struggle fundamentally with self-regulatory challenges. EEG

biofeedback may teach children to self-regulate brain rhythmicity, which in turn may

impact their ability to adaptively manage physiological and emotional arousal states.

Because poor self-regulation of arousal is central to the vast majority of the behavioral

difficulties experienced by children with traumatic attachment histories, neurofeedback

has the potential to affect global improvements in the typical problem areas of attention,

aggression, impulse control, hypervigilance, classroom learning, empathy and trust

formation, sleep, hyperactivity, etc. The mechanisms and objectives of EEG

biofeedback would seem to be expressly appropriate for the behavioral and

neurophysiological needs of this population. The present pilot study using a clinical case

series represents an initial step in the empirically-based investigation of neurofeedback

treatment for children with reactive attachment disorder (RAD) symptoms.

24
 CHAPTER 4

PILOT STUDY

Method

Participants

Data for this study were obtained from files of 20 children treated in a private

neurotherapy practice affiliated with the University of North Texas. All of the children

had documented histories of removal from their biological home by Child Protective

Services (CPS) and were currently living with adoptive parents. Participants included 9

girls and 11 boys, all of whom were brought by their adoptive parents between 2002

and 2005 for treatment of behavioral problems associated with attachment difficulties.

Many of the children had previously been diagnosed with reactive attachment disorder

(RAD) and all of the children manifested many of the typical symptoms of RAD, as

described above. The children ranged in age from 6 years to 15.5 years, with a mean

age of 10.43 years (SD = 2.66). Fifteen of the participants were Caucasian, two were

Latino, and three were African-American. All 20 children had experienced multiple

previous therapies for RAD related symptoms, with little or no success. Each child

participated in approximately 30-40 neurofeedback sessions (M = 38) over the course of

2 to 8 months (M = 5); however, all post-treatment assessments used in the analysis for

this study were collected at the end of 30 sessions. The majority of the participants were

taking medications concurrent with neurofeedback training, including serotonin reuptake

inhibitors, amphetamine, methylphenidate, atomoxetine, ziprasidone, and risperidone.

Quantitative Electroencephalogram (QEEG) Profiles

All 20 of the childrens files were included in a separate study at the University of

25
North Texas investigating QEEG findings for adults and children with abuse histories.

QEEGs of the 26 children comprising the latter study (i.e., the present sample plus six

additional children) were compared to the NxLink (John, Prichep, & Easton, 1987) and

NeuroGuide (Thatcher, 1998) databases of normal controls. The results showed that, in

general, the abused children had significantly decreased delta in frontal sites with a Z-

score range of -2.05 to -2.31. All 26 children in this larger sample had increased theta in

at least one frontal International 10-20 site (Jasper, 1958), and 21 of these had relative

power Z-scores in excess of 1.20. (See Appendix A for a graphic depiction of the

International 10-20 Electrode Placement System). NxLink maturational lag scores were

also computed to express, in years, how much the children in the sample lagged behind

children in an age-matched reference group. All of the maturational lag averaged values

except those at O1, O2, T5, and T6 were in the negative direction. Left fronto-central lag

scores for the group ranged from -.05 at C3 to -1.07 at Fp1, indicating that the left

anterior relative power values of the abused/neglected sample were more consistent

with values seen in normal children one year younger. Seventy-five percent of the

QEEG records showed significant coherence Z-score abnormalities (either

hypercoherence, hypocoherence, or both). Although these QEEG findings refer to a

slightly expanded sample of 26, they are descriptive of the children comprising the

neurofeedback treatment group in this study.

Instruments

All QEEG assessments and neurofeedback treatments were conducted by a

licensed counselor who is also a Biofeedback Certification Institute of America (BCIA)

certified electroencephalogram (EEG) biofeedback provider. EEG data was digitally

26
recorded at 19 scalp electrodes contained within a stretchy QEEG cap and referenced

to linked ears using either a Lexicor digital EEG system (NRS-2D) or a Deymed

TruScan 32 digital system. Electrode resistances were kept below 5 Kohms and equal

to within + 1 Kohm between leads. On the Lexicor equipment, bandpass filters were set

at .5-30 Hz, and the sampling rate was set at 128 samples per second. On the Deymed

equipment, bandpass filters were set at .5-80 Hz. Because the Deymed samples 4,096

times per second during the recording process, the sampling rate was decreased to 128

samples per second when exported for analysis. Neurofeedback sessions were

recorded and conducted using either Lexicor Biolex or BrainMaster equipment and

software. Settings on the BrainMaster equipment were set for single channel training at

the third filter order with peak to peak amplitude scaling and manual thresholding. The

Lexicor equipment also was set for single channel training, utilizing digital filters set in

the eighth order with manual thresholding. In all sessions, EEG was recorded from a

referential montage, with a single-electrode placement referenced to the ipsilateral ear

and a ground on the opposite ear, using a sampling rate of 128 Hz.

The Test of Variables of Attention (TOVA; Greenberg, 1987) was used to assess

self-regulatory changes in attention following neurofeedback training. The TOVA is a

computerized visual continuous performance test in which two easily discriminated

visual stimuli are presented for 100 milliseconds every 2 seconds for 22.5 minutes. The

variables assessed by the TOVA (i.e., errors of omission, errors of commission,

response time, and response time variability) have been shown to be significantly

different between pre-treatment and methylphenidate treatment conditions, and are

reported to be invulnerable to test-retest practice effects (Greenberg, 1987).

27
 The Child Behavior Checklist (CBCL) is widely utilized in both clinical and

research applications (Achenbach, 1991). It was used in the present study to assess

behavioral and emotional changes following neurofeedback training, as reported by

adoptive parents. The CBCL is a four-page form designed to obtain descriptions of the

competencies and behavioral-emotional problems of children that impact successful

adaptive development. Parents with a reading ability at the fifth grade level or greater

typically complete the form in 10 to 15 minutes. The CBCL includes competence scales

(i.e., activities, social, and school), as well as eight syndrome scales. The syndrome

scales include an internalizing grouping (i.e., withdrawn, somatic complaints, and

anxious/depressed), an externalizing grouping (i.e., delinquent behavior and aggressive

behavior), as well as scales for social problems, thought problems, and attention

problems.

Procedures

All procedures for this study were approved by the Committee on the Use of

Human Subjects at the University of North Texas. All participants gave informed

parental consent and/or assent for inclusion in the study. The pre-treatment CBCL was

completed by an adoptive parent prior to or concurrent with the childs first intake

session. Child clients completed the pre-treatment TOVA during the initial session. All

but three participants in the study completed a post-treatment TOVA following the 30th

neurofeedback training session. Several of the children continued for about 10

additional sessions, but only the post 30 session assessment data was used for

treatment outcome analysis. The same adoptive parents who completed the pre-

treatment CBCLs also completed the post-treatment CBCLs at 30 sessions of training.

28
QEEG data collection and analysis occurred prior to neurofeedback training, so that

individualized treatment protocols could be developed.

Neurofeedback training consisted of 30 minutes of auditory and visual feedback

per session, while the child was seated comfortably in a quiet room. Although each child

was treated with an individualized protocol, there was nevertheless much similarity

among these protocols due to the common features seen in the groups QEEGs. The

training site locations were derived from the individual NeuroRep weighted average

QEEG topographies (not from the relative power Z scores reported earlier from the

larger university sample of children with relational trauma). The weighted average

topographies showed excess slow wave amplitude in the frontal and/or central vertex

locations for all subjects treated. This pattern was consistent with the RAD QEEG

patterns reported by Van Bloem (2000) and by Fisher et al. (2005). In all cases,

feedback initially was contingent on the reduction of 2-7 Hz activity at CZ or FZ based

on the International 10-20 electrode placement system (Jasper, 1958). Training at the

frontal sites usually commenced by rewarding the reduction of 2-7 Hz activity. In a few

cases, both the CZ and FZ sites were trained within the first 20 sessions, targeting FZ

first for 10-15 sessions and then moving to CZ for 5-8 sessions. In cases where the

QEEG showed excessive fast frequency activity at FZ or CZ as well, a second inhibit

filter band of 20-32 Hz was also utilized in the initial protocol. No frequencies were

enhanced during the initial frontal training period; all protocols were inhibiting only.

This protocol was followed until the individual subjects training records indicated

consistent ability to maintain targeted thresholds, i.e., consistently reducing 2-7 Hz and

20-32 Hz activity below baseline levels, and also when changes in behavioral symptoms

29
were observed. In the majority of cases, guided by subjects individual QEEGs and/or

behavioral changes, the next stage of training targeted right hemisphere protocols

enchancing 12-15 Hz and reducing 2-7 Hz at T4, P4 or C4. Often, this right side training

was initiated following treatment at frontal and vertex sites if the clinician observed or

parents reported increased agitation in the child. Right side training reliably reduced

such responses. For those children not requiring this right side training focus, second

stage protocols targeted various QEEG-determined sites and frequencies, most often

reducing 8-12 Hz Alpha activity at C3, CZ, C4, P3, PZ, P4, T5, or T6. Again, no

frequencies were enhanced in these alpha reduction protocols.

Results

Child Behavior Checklist

Differences in CBCL raw scores before and after neurofeedback training were

analyzed using the Statistical Package for the Behavioral Sciences (SPSS). Eighteen of

the 20 participants adoptive parents completed valid pre- and post-CBCL parent forms,

which were included in the analysis. Raw scores were used for analysis in accordance

with CBCL manual recommendations for research (Achenbach, 1991). A two-tailed

paired t-test analysis yielded meaningful statistical differences in 10 of the 14 assessed

competence and syndrome scores, indicating that significant behavioral improvements

were observed by adoptive parents following neurofeedback training. Total syndrome

scale scores decreased an average of 23.05 points (SD = 21.44) with a 95% confidence

interval of 12.73-33.39, and were significant at t(18) = 4.69, p < .001. Externalizing scale

scores decreased an average of 7.32 points (SD = 8.40) with a 95% confidence interval

of 3.27-11.36, and were significant at t(18) = 3.799, p = .001. Both total and

30
externalizing scale score changes represent large effect sizes (d = .78 and d = .94,

respectively). Internalizing scale scores decreased an average of 4.74 points (SD =

7.26), with a 95% confidence interval of 1.24-8.24, and significance at t(18) = 2.84, p <

.01. This represents a medium effect size (d = .59). Figure 1 presents these results

graphically.

Figure 1. Averaged raw scores on the CBCL total, externalizing, and internalizing scales
before and after neurotherapy treatment. All differences are statistically significant at
p < .05.

Within the CBCL syndrome scales, six of the eight scale scores significantly

improved: social problems [t(18) = 3.59, p = .002], aggressive behaviors [t(18) = 3.72, p

= .002], thought problems [t(18) = 3.33, p = .004], delinquent behavior [t(18) = 2.82, p =

.01], attention [t(18) = 2.82, p = .01], and anxious/depressed [t(18) = 2.80, p = .01]. All of

these changes represent medium effect sizes (d >.55). Somatic complaints and

withdrawn scale scores also improved, but not to statistical significance. Figure 2

presents these results graphically. Of the three competence scales, there was a

31
statistically significant improvement on the social scale [t(12) = 2.67, p = .021] with

medium effect size (d = .68). Activities and school scale scores also showed positive

trends following treatment, but not to statistical significance.

Figure 2. Averaged raw scores on the CBCL syndrome scales (social problems,
delinquent behaviors, aggressive behaviors, somatic complaints, attention problems,
anxiety/depression, withdrawn, and thought problems) before and after neurotherapy
treatment. An asterisk indicates significance at p < .05.

Test of Variables of Attention

Pre- and post-TOVA standard scores were also analyzed using SPSS. Ten

children completed valid pre- and post-TOVAs, while the remaining 10 protocols were

not included due to incompletion or one or more invalid quarter scores, (e.g., due to

>10% anticipatory errors). Two-tailed paired t-test analysis found significant differences

post-treatment for three of the four TOVA variables: omission errors [t(9) = -2.37, p =

.04], commission errors [t(9) = -3.16, p = .011] and total variability [t(9) = -2.39, p = .04].

32
 On the TOVA, lower scores represent more problematic behavior, with average

standard scores ranging from 85-115. Omission errors improved an average of 17.3

points (SD = 23.0) from an average pre-treatment score of 71.5 to an average post-

treatment score of 88.8 (95% CI 0.8-33.8). Commission errors improved an average of

16.4 points (SD = 16.4) from an average pre-treatment score of 91.5 to an average

post-treatment score of 107.9 (95% CI 4.7-28.1). Total variability scores increased an

average of 12.3 points (SD = 16.3) from 67.1 at pre-test to 79.4 at post-test (95% CI

0.7-23.9). Omission, commission, and total variability scale score changes all represent

medium effect sizes (d > .60), with average omission and total variability scores moving

from abnormal to normal ranges following treatment. Total response time standard

scores decreased slightly (from an average of 71.9 to 70.0). These results are depicted

in Figure 3.

Figure 3. Averaged TOVA standard scores (omission errors, commission errors,

response time, and response time variability) before and after neurotherapy treatment.
An asterisk indicates significance at p < .05.

33
 Discussion

The CBCL and TOVA score improvements observed in this study suggest that

neurofeedback is an effective treatment for children with behavioral problems

associated with histories of neglect and/or abuse. Specifically, aggressive, delinquent,

and socially problematic behaviors (e.g., lying, fighting, demanding attention, etc.)

appear to be strongly impacted. Behaviors associated with attentional problems,

anxiety/depression, and thought problems also appear to be significantly reduced.

These data are encouraging because these problematic behaviors are not only common

and disabling for children with traumatic attachment histories, but are also frequently

resistant to change via traditional therapeutic interventions. On the other hand, because

these behaviors are often the primary impetus for referral, they are also potentially

vulnerable to overly-optimistic behavior ratings from parents seeking to justify treatment.

It should be noted that the CBCL activities and school scales may not have shown

significant changes because many children received treatment during summer vacation

and, as several parents reported, it was difficult to assess these behaviors during

summer months when school was not in session and social activities were less

frequent.

Improvement of self-regulatory behaviors characterizes the changes in CBCL

and TOVA Scale scores observed in this sample. For the majority of the children, CBCL

and/or TOVA scores shifted from clinical/borderline ranges into normal ranges,

indicating meaningful behavioral changes occurred. The statistical significance and

medium to large effect sizes of score changes following treatment is notable, especially

considering the small sample size of the study. If, as has been suggested,

34
neurofeedback training teaches children to self-regulate brain rhythmicity, the broad

changes observed in the CBCL scale scores suggest a general decrease of

physiological and emotional arousal dysregulation. CBCL externalizing scale score

changes, as well as decreases in thought problems (e.g., repetitive actions,

distractibility), attention problems (e.g., poor concentration, impulsiveness),

anxiety/depression (e.g., nervousness, crying), and somatic complaints (e.g., tiredness,

headaches) may all be expressions of improved stability of the regulatory CNS system.

Improvements on TOVA omission, commission, and total variability scores imply

decreased impulsivity and improved attention modulation, and corroborate positive

changes on the attention scale of the CBCL. (The slight decrease in response time

scores on the TOVA is commonly seen immediately following neurofeedback, as

individuals become more focused and deliberate in their responses.)

Although encouraging, this pilot study clearly represents only a preliminary

clinical investigation of the effectiveness of neurofeedback treatment for this population.

Some of the major caveats of the study include the small and non-randomized sample,

the lack of control group, and the failure to control for potential effects of medications

and concurrent additional therapies. The use of an external behavioral rating (CBCL)

and a standardized performance test (TOVA) provides a multidimensional method of

evaluation, but with considerable limitations. In addition to the potential vulnerability of

the CBCL to low reliability and validity due to parental investment in positive change, the

CBCL is designed to be given at intervals of 6 months or more. In the present study,

some children were re-evaluated as early as 3 months following the initial session.

Standard deviations of certain CBCL and TOVA scores were also large, thus should be

35
interpreted with some caution in spite of the robustness of t tests to violations of

normality. These data imply that there is a high degree of variability between children in

terms of their initial presentation and their patterns of improvement following

neurofeedback training.

It should be noted that randomized controlled trials, while strong in terms of

internal validity, have been criticized by some for their lack of external validity and

practical application to the broader clinical population. Recent studies cited by

Hirshberg, Chiu, and Frazier (2005) indicate the results of non-randomized

observational studies are generally similar to those of randomized controlled trials; thus,

the authors assert the conservative emphasis on randomized controlled trials to

demonstrate efficacy of treatments may be misguided. Specifically, Hirshberg et al.

(2005) suggest that effectiveness research i.e., formal measurement of outcomes

from treatments administered in typical clinical practice may provide empirical support

that is better aligned to accurately represent treatment outcomes for a general clinical

population. The authors further note that the American Psychological Association (APA)

and the EEG biofeedback professional organizations have considerably more stringent

parameters for evaluating the efficacy of clinical practices than the American Academy

of Child and Adolescent Psychiatry (AACAP), which gives more weight to the

knowledge base that emerges from shared clinical experience. In this light, the lack of a

randomized control group for the present neurotherapy pilot study (and indeed for the

vast majority of studies designed to treat maltreated children) may potentially be

considered a less serious design liability.

36
 It is interesting, both clinically and etiologically, that the majority of the childrens

initial QEEGs (specifically the weighted average amplitude topographies in NeuroRep)

indicated use of similar treatment protocols (inhibiting 2-7 Hz activity at CZ or FZ

followed by inhibiting 2-7 Hz and enhancing 12-15 Hz at T4, P4 or C4). Other clinicians

have reported combining the slow wave inhibits at Fz and Cz with enhancement of 15-

18 Hz (Fisher, Turber, & Gunkelman, 2005), which is apparently an effective alternative

to the inhibit-only focus of the treatment reported in this study. Although T6 has been

mentioned in the literature as a focal site in RAD (Schore, 1994), T6 was not found to

have as much slow wave activity as the Cz and Fz sites in the pre-treatment QEEGs of

the current sample. It should be noted that some of the children who comprised the

sample had never received a formal diagnosis of RAD. Further analysis of QEEG

patterns in this population may clarify areas of the developing brain that are particularly

susceptible to early relationship trauma, as well as help establish more standardized

neurofeedback protocols for this population. In the future, it would also be beneficial to

analyze the actual power changes in the EEG frequency bands as a result of

neurofeedback. Long-term follow-ups, additional assessment measures, larger,

randomized samples, and control of additional therapies were regrettably not possible

for the present study. Nonetheless, the findings clearly support the use of

neurofeedback training for children with histories of relationship trauma and serve to

clarify directions for further research.

37
 CHAPTER 5

META-ANALYSIS

The second part of the present project aims to contextualize the neurofeedback

pilot study in terms of its practical efficacy. Due to the fact that it was not possible to

include a control or comparison group in the study, the results, though significant, lack

meaning in terms of their realistic clinical application. Thus, for the purposes of further

evaluating the clinical and statistical relevancy of this pilot study, comparison of its

outcomes to those of other treatments was deemed necessary. Initial attempts to find

empirical literature documenting the efficacy of various treatments with pediatric

populations suffering from abuse and/or neglect were unfruitful. It quickly became

apparent that no such comparative or meta-analytical research had yet been conducted;

thus, there was no existing literature that could provide general guidelines for the

expected treatment outcome efficacy for this population.

The following meta-analysis is intended to remedy this gap in the research

literature by providing a comprehensive overview of the outcome research with this

population. The use of a meta-analysis (rather than a qualitative review of the literature)

provides a statistically meaningful method to measure effect sizes (i.e., outcome

efficacy) across different studies. Meta-analysis creates a uniform yardstick to

measure the effectiveness of one treatment study against another, in spite of their

potentially differing outcome measures. Ultimately, a quantitative estimate of net benefit

aggregated over all the included studies can provide a basic comparative guideline for

the pilot study outcomes, as well as for other studies with this population.

38
 Method

Definition of Terms

Following the guidelines of Weisz, Donenberg, Han, and Weiss (1990),

therapeutic treatment was defined for the present study as any intervention designed to

alleviate psychological distress, reduce maladaptive behavior, or enhance adaptive

behavior through counseling, structured or unstructured interaction, training programs,

or predetermined treatment plans. Treatments were excluded from the meta-analysis if

they were described purely as reading interventions (e.g., bibliotherapy), were

psychopharmacological in nature, involved teaching or tutoring intended only to

increase knowledge about a specific subject, predominantly involved intervention via

relocation (e.g., foster placement), or were purely preventative in nature (e.g., school-

based family education for the prevention of child abuse). For meta-analysis inclusion,

the primary recipient of intervention was necessarily the child, although adjunctive

family interventions were commonplace and did not exclude studies from inclusion.

The working definition of abuse for the present study included severe emotional,

physical, and sexual maltreatment; and the working definition of neglect included

caretaker failure to provide basic physical, emotional, and medical/educational needs.

Diagnosis of failure to thrive was not considered relevant to the study unless it occurred

specifically as a result of neglect. Although abuse and neglect (and indeed each of their

subtypes) are likely associated with different risk indicators, the vast majority of studies

reflect the reality that there is a high level of co-occurrence in clinical child populations

(Allin, Wathen, & MacMillan, 2005; Scher, Forde, McQuaid, & Stein, 2004). Thus, there

39
was no attempt to separate these sub-groups in terms of their discrete outcomes

following therapeutic interventions.

Literature Search

Attempts were made to locate all completed, English-language outcome studies

of treatments for abused and/or neglected children and adolescents ages 3-17. The

computerized databases PsychInfo, EBSCO Host, Social Work Abstracts, and Medline

were searched from 1970 to 2006, using the key words child, adolescent, youth, abuse,

neglect, sexual abuse, maltreatment, treatment, therapy, counseling, child protective

services, and intervention. From this search, approximately 670 articles and abstracts

were found, of which approximately 120 were more closely examined for possible

inclusion in the present study. Ultimately, twenty research articles from databases were

found to meet inclusion criteria for the present meta-analysis.

Secondly, relevant child therapy meta-analyses (e.g., Weisz et al., 1995; Casey

& Berman, 1985), child therapy review articles (e.g., MacMillan, 2000; Kadzin et al.,

1990), and recent volumes of pertinent journals (e.g., Child Maltreatment) were

individually examined for possible inclusion articles. This yielded another 28 articles,

which were more closely examined for potential inclusion, ultimately yielding five studies

meeting criteria.

Finally, dissertations were utilized to represent completed studies not subject to

publication bias (i.e., an unrealistic predominance of significant findings). Dissertations

written between 1970 and 2006 involving treatment outcome studies with

abused/neglected children were identified through the Dissertation Abstracts

International (DAI) database for possible inclusion in the present meta-analysis. Twenty-

40
six dissertations and theses were examined for potential inclusion, of which eight

ultimately met criteria and were included in the meta-analysis. Table 1 depicts the

stages of literature review and the total number of studies involved in each stage of the

inclusion process.

Table 1

Number of Studies Included at Initial and Final Stages of Literature Review

Initial Final inclusion

literature following review
search
Computerized data bases 670 20
PsychInfo, EBSCO Host, Social Work Abstracts, and
Medline (1970 to 2006).
Key words: child, adolescent, youth, abuse, neglect,
sexual abuse, maltreatment, treatment, therapy,
counseling, child protective services, and
intervention.
Hand searches 28 5
Literature reviews, meta-analyses, recent journals
(2005-2007)

Dissertations 26 8
Dissertation Abstracts International (DAI) database
(1970-2006)

Total 724 33

Criteria for Study Inclusion

The criteria for the inclusion of studies in the meta-analysis was as follows: (1)

studies were designed to examine the effectiveness of a treatment intervention for

children and/or adolescents ages 3-17 who had experienced abuse and/or neglect; (2)

results were based on empirical measures, as opposed to clinical impressions or

unstructured interviews; (3) sufficient statistical information was reported in the study to

allow for calculation of effect size estimates; and (4) sample sizes were greater than or

41
equal to four (N 4). In two cases where inclusion status was not straightforward, an

impartial second review was sought from a clinical psychology doctoral student who had

experience doing meta-analyses, but no investment in the outcome literature with this

population.

The most common reasons for rejection of studies for the present meta-analysis

included parent rather than child-focused interventions and outcome measures,

insufficient statistical information (e.g., failure to report standard deviations, confidence

intervals, t-scores, or adequate raw data to extrapolate key statistics), lack of clear

empirical outcome measures, and single case study design. Unfortunately, there were

over 20 older empirical studies that reported significant p values without providing

further statistical data, thus precluding the calculation of effect sizes. Attempts to

contact the authors for additional data were not successful.

Due to the limited quantity of outcome research with abused and/or neglected

children, the present meta-analysis did not require that studies utilize a randomized

control group design for inclusion. This leniency is discrepant with the gold standard for

meta-analysis (e.g., Weisz et al., 1995), but due to the current state of research with this

pediatric population, this represents a necessary and reasonable design compromise.

Due to the paucity of controlled experimental design studies for this population, a more

inclusive meta-analysis provides a means of comparing research results across

different studies, with the end result of supporting stronger future research on the

subject.

It should be noted that some authors suggest that inclusion of single treatment

group pre- post-designs in a meta-analysis may inflate effect sizes. Although the reason

42
for this potential inflation is debatable, Lipsey and Wilson (1993) speculate that it may

be a result of the confounding of maturational effects with treatment effects. Alternately,

slight effect size inflation may be due to the fact that single group repeated measure

designs do not control as well as matched-control designs for regression to the mean

(Reeker, Ensing, & Elliott, 1997). Indeed, this concern is relevant to any meta-analysis

of therapeutic interventions because these studies necessarily utilize predominantly

clinically distressed samples.

In a number of studies reviewed for potential inclusion in the current meta-

analysis, single data sets were found to be included in multiple research papers. For

example, data from a sample group of child participants in an initial pilot study might

again serve as part of a larger sample group in a follow-up study by the same author.

To maintain independence between included data sets, information was sought in the

articles and through personal communication with the authors to determine whether the

same data sets had been used in multiple studies. If the data sets overlapped, the study

with the largest sample size and most complete set of outcome measures was included

while the others were excluded.

On the other hand, there were also a number of empirical studies that compared

two or more different therapeutic interventions utilizing two distinct sample groups within

one study. For example, one study was designed to evaluate two randomly selected

samples of abused children from a community mental health center who received either

group or individual therapy. The two groups were compared pre- and post-treatment on

behavioral outcome measures. In this case, a single empirical study essentially

provided distinct intervention outcome data for two treatment approaches, each with

43
separate and non-overlapping participant pools. Because the goal of the present meta-

analysis was not only to determine an overall effect size for therapeutic interventions

with abused/neglected children, but also to attempt to separate out the utility of different

treatment approaches, studies involving more than one intervention approach with

distinct sample groups were especially valuable. In these cases, as long as the samples

did not overlap, multiple treatment outcome data were separated out for analysis of

specific treatment approaches before being recombined to calculate the aggregate

effect size for the study.

Computation and Analyses of Effect Size

Effect sizes for each within-group design study included in the present meta-

analysis were calculated using reported pre-post group means and standard deviations

of the dependent variables. Specifically, Cohens d-values (Cohen, 1988) were

computed by finding the difference between the pre-treatment and post-treatment mean

scores and dividing by the pooled standard deviation of measures in that study (see

Equation 1 depicting Cohens d calculation for within groups studies).

(1)

Cohens d calculation was followed by correcting for sample size using the

Hedges formula. Imitating a similar model by Reeker and colleagues (1997), this

correction was utilized because the variance of an effect size is partially a function of a

studys sample size. According to Weisz et al. (1995), small sample sizes (N < 15)

commonly result in slightly greater effect sizes than larger sample sizes. Hedges and

44
Olkin (1985) have noted that a problem with the Cohens d statistic is that the outcome

is heavily influenced by the denominator of the equation. If one standard deviation is

larger than the other, the denominator is weighted in that direction and the effect size is

more conservative. However, large standard deviations are associated with studies

utilizing larger sample sizes, which logically should be more heavily weighted for their

statistical importance. Hedges incorporates sample size by computing a denominator

that incorporates the sample sizes of the respective standard deviations while also

adjusting for overall effect size based on this sample size (see Equation 2 depicting the

formula for Hedges sample size adjustment).

(2)

For empirical studies that utilized an independent groups design without reporting

standard deviation values, effect sizes were alternatively calculated using the F-test

statistic, mean score, and standard deviation/mean squared error. Equation 3 depicts

the Cohens d calculation for between group studies using F-test data. This calculation

method is recommended by Thalheimer and Cook (2002) for studies utilizing between-

group designs, such as randomized controlled studies. Two studies included in the

current meta-analysis utilized this method of effect size calculation.

d = xt xc / MSE(nt + nc 2/nt + nc) (3)

45
 In cases where independent group design studies reported t-test values but no

standard deviation or standard error values, Thalheimer and Cook (2002) recommend

that effect size may alternatively be calculated using number of subjects (n) and the t-

test statistic. Equation 4 depicts the Cohens d calculation for between group studies

using t -test data. This alternative approach was utilized for four studies included in the

current meta-analysis.

d = t (nt + nc/ntnc)( nt + nc / nt + nc 2) (4)

Three publicly available effect size calculators were utilized to cross-compare

effect size calculations and ensure accurate Cohens d statistics for meta-analysis

inclusion. These included a frequently cited on-line effect size calculator created by Lee

A. Becker at the University of Colorado at Colorado Springs (Becker, 2000), an on-line

Excel spreadsheet calculator created by Samathna Cook of Harvard University and Will

Thalheimer of Work Learning Research (Thalheimer & Cook, 2002), and a

downloadable effect size calculation program created by Grant Devilly of Swineburn

University, Austria (Devilly, 2004).

Analysis of pre-post change effect sizes within each study allowed the use of the

largest number of studies, although, as described above, the effect size values can be

expected to be only roughly equivalent to the between group effect sizes described in

most meta-analyses (e.g., Lipsey & Wilson, 1993). Rosenthall (1991) suggested that

because the paired t-test value takes into account the correlation between the two

scores, paired t-test values will be larger than a between groups t-test values. Thus,

effect sizes computed using paired t-tests will always be larger than effect sizes

46
computed using between-groups t-tests. Dunlop, Cortina, Vaslow and Burke (1996)

argue that this discrepancy between within-group and between-group designs is best

managed by using the original standard deviations to compute effect sizes, rather than

the paired t-test value or the within-subjects F value. Thus, the present study attempted

to use standard deviations for all within-group effect size calculations, as described

above. In two cases, due to constraints with paucity of reported data, this was not

possible.

To control for non-independence of findings within studies that employed multiple

measurements of the same subjects (i.e., conceptually replicated designs), an overall

effect size was computed for each study. As recommended by Hunter and Schmidt

(1990), individual d-values for each relevant outcome measure were averaged together.

For fully replicated designs based on separate groups of subjects, averaging of effect

sizes was not necessary.

In a few cases, outcome measures were described in the methods section, but

the follow up data were not presented in the results section. In this situation, an effect

size value of zero was assigned to that measure. Zero was also assigned as an effect

size when results were reported as non-significant and no further data was provided.

Casey and Berman (1985) recommend this approach due to the fact that excluding non-

significant or non-reported statistics would result in an artificial inflation of the overall

effect size. Assigning an effect size of zero is thus a conservative procedure to

counterbalance investigator bias toward exclusively reporting statistically significant

data. This approach was also used in a meta-analysis of child and adolescent group

treatment by Hoag and Burlingame (1997). After calculating an average effect size for

47
each study, all study effect sizes were averaged to derive a total overall mean effect

size for treatment of children with histories of abuse and/or neglect.

Results

The grand total average Cohens effect size for all 33 studies included in the

meta-analysis for children with histories of abuse and neglect is medium (d = .70; SD =

.45; 95% CI .54-.85). It is inevitably complex making generalizations across 33 studies

with so much diversity in terms of research designs, treatments, outcome measures,

sample age, etc. Nonetheless, it is hoped that the following examination of key variables

will help to contextualize and integrate the broad effect size results. Using the Statistical

Package for the Social Sciences (SPSS), appropriate statistical analyses have been

employed to give a rough sense of the contributions of various factors. However, due to

the small size and heterogeneity of the present meta-analysis, an individualized and

narrative approach also is helpful in interpreting the significance of effect size findings.

When asserting the overall effectiveness of various treatments, concerted

attempts have been made to incorporate effect size results with relevant variables, such

as design strength, sample size, breadth of outcome measures, and treatment duration.

Information regarding these details is summarized in table form for clarity and ease of

accessibility in Appendix B. To the extent made available by the studies authors,

Appendix B displays each studys essential data regarding treatment setting, sample

demographics, sample sizes, specific treatment interventions, treatment duration, client

contact time, treatment outcome measures, measure response perspective, research

design, as well as individual outcome effect sizes and overall study effect sizes.

Presentation in this way is intended to facilitate easy reference and to allow readers to

48
compare characteristics of particular interest across studies (for example, length of

treatment or Child Behavior Checklist effect sizes).

Participant Characteristics: Type of Abuse History

Looking broadly at the studies meeting criteria for meta-analysis inclusion, it is

immediately apparent that significantly more empirical research has been conducted on

the treatment of sexually abused children than on the treatment of physically abused

and neglected children. Of the 33 studies included, 24 involved interventions designed

specifically for children with histories of sexual abuse. Of these, five solely investigated

outcomes of individual therapy interventions and 11 looked solely at group therapy

outcomes. In both individual and group studies, about half included parallel parent

therapies (either group, individual, or family treatment). Deblinger et al. (1996)

compared the effectiveness of individual therapy to family therapy, and Baker (1985)

compared group therapy to individual therapy, with similar effect size outcomes in all

cases. A study by Nolan et al. (2002) investigated the effectiveness of combining

individual and group therapy; Rust and Troupe (1991) investigated the full combination

of group, individual, and family therapy; and Clendenon-Wallen (1991) investigated

group therapy with the addition of music therapy. These studies collectively suggest that

adding individual, family, or music therapy to group therapy does not significantly add to

the benefits of group therapy alone for sexually abused youth.

Of the studies designed to treat children with histories of sexual abuse, the

largest effect sizes were consistently seen in correlation with (1) parallel group therapy

treatments for children and non-offending caretakers (e.g., Hall-Marley & Damon, 1993;

Humberson, 1998), and (2) structured child group therapy with a specific educational

49
component for the treatment of sexual abuse (e.g., De Luca et al., 1995; Verleur et al.,

1986). Treatment duration for these studies ranged from three months to a year, with

longer durations almost always resulting in stronger results on behavior outcome

measures. One notable exception is the study by Perez (1987), which found strong

results and large effect sizes on self-concept and locus of control measures using group

play therapy and individual play therapy with sexually abused children after three

months of weekly sessions. Also pertinent is the finding by Deblinger et al. (1996) that

family dyad cognitive behavioral therapy had very large effects on measures of child

post traumatic stress disorder (PTSD) symptoms after three months of treatment.

For samples of sexually abused children, cognitive-behavioral individual child

therapy was, in general, not as effective as group cognitive-behavioral therapy. Non-

structured or supportive individual therapy was less effective still (e.g., Cohen &

Mannarino, 1996; 1998). The least effective treatment for children with sex abuse

histories in the present meta-analysis was individual non-directive play therapy (e.g.,

Cotton-Cornelius, 2000; Zion, 1999). It should be noted, however, that the play therapy

study by Perez (1987) referenced above supports the benefits of play therapy for this

population, particularly in a group modality. Additionally, Baker (1985) found large effect

size outcomes for sexually abused teenagers using both client-centered group and

client-centered individual interventions. The mean Cohens effect size for the entire

group of 24 studies geared toward the treatment of sexually abused children and teens

was medium (d = .71; SD = .41; 95% CI .54 -.88).

A total of nine studies included in the meta-analysis focused on samples of

children with physical abuse and neglect histories. In all cases there was no separation

50
of the two forms of maltreatment, and groups were made up of children who may have

experienced one or both maltreatment types (and/or unspecified sexual abuse, as well).

Two studies utilized group treatment, one of which utilized parallel group therapy for

caregivers. As with the sexual abuse studies, group therapy was associated with

medium to very large effect size changes on global measures of behavior, with longer

treatment durations producing stronger effects (McGain & McKinzey, 1995; Stauffer &

Deblinger, 1996). Neurotherapy with this population was associated with comparably

large effect sizes (Huang-Storms et al., 2007).

Three studies with physically abused and/or neglected samples utilized play-

based therapies. Educational theraplay and social-modeling play therapies were

associated with large effect size outcomes on developmental inventories (Fantuzzo et

al., 1988; Stubenbort, 2000), while structured individual play therapy was associated

with only small effect size changes on outcome measures (Reams & Friedrich, 1994).

Least effective of the included study treatments for abused and neglected children was

the use of outdoor adventure groups (Tucker, 2006), though it should be noted that this

was the only abuse/neglect sample comprised primarily of adolescents.

Overall, the mean Cohens effect size for the treatment of physically abused and

neglected children was medium (d = .66; SD = .56; 95% CI .23 -1.09), slightly lower

than the effect size for the treatment of sexually abused youth. A two-tailed t-test for

independent samples indicated that the difference in mean effect sizes for sexually

abused and physically abused/neglected youth was not statistically significant, t(31) =

.29, p = .77. This suggests that children who have experienced these different

51
categories of abuse are similarly responsive to treatment interventions, overall (see

Figure 4).

2.5

2.0 30
23

1.5

1.0

.5

0.0

-.5
D

N= 24 9

sexual phy/neglect

ABUSE

Figure 4. Effect size means for studies grouped by type of abuse.

Participant Characteristics: Age

Age characteristics of the samples included in the meta-analysis varied greatly,

with some studies including participants ranging in age from pre-school to high school

and other studies geared toward single-aged participants. Studies were categorized into

four age groups described below. A one-way ANOVA indicated that age groupings did

not differ significantly from one another in terms of effect size results, F(3,29) = .85, p =

.48 (see Figure 5).

52
 2.5

2.0 30

1.5

1.0

.5

0.0

-.5
D

N= 10 6 8 9

3-9 9-13 ad/teen mixed

AGE

Figure 5. Effect size means for studies grouped by age.

Five of the studies focused on pre-school and kindergarten aged children (3 to 6

years old), and five more included a wider age range from age 3 to 9 years. All of these

studies involved play and/or group therapies, and two included additional milieu or

multi-modal family support. Of these ten studies, seven demonstrated medium to large

effect sizes, making for a medium to large total effect size average for the 3-9 year old

group (d = .77; SD = .51; 95% CI .40-1.14).

Six studies focused on late grade-school through middle school-aged children

(approximately 9 to 13 years old), and primarily utilized group and/or cognitive-

behavioral approaches for treatment. The two strongest studies in terms of outcome

measures for this age group utilized family (dyad) cognitive-behavioral therapy

(Deblinger et al., 1996) and eye movement desensitization and reprogramming (EMDR;

Nasrin et al., 2004) for the treatment of PTSD symptoms. Of the six total studies for this

53
age group, four demonstrated large or very large effect sizes, making for a large total

effect size average for the 9 -13 year old group (d = .90; SD = .60; 95% CI .27-1.5)

Another age category was comprised of adolescents and teenagers, consisting

of eight studies included in the meta-analysis. All of these studies utilized group therapy

in some form, often with adjunctive therapies (e.g., music skills training, perpetrator

group therapy, or individual therapy). In general, groups meeting for 6 months or more

demonstrated large effect size changes on outcome measures (e.g., Rust & Troupe,

1991; Verleur et al., 1986), while small effect sizes were observed in group treatment

durations of 12 weeks or less (e.g., Clendenon-Waller, 1991; Thun et al., 2002). An

exception is a study by Baker (1995) using a 6 week unstructured group therapy

intervention with sexually abused teen girls that found large effect size changes on a

self concept measure. Also of note is the apparent success of inclusion of a sex

education component to a girls group therapy program for increasing levels of self

esteem (Verleur et al., 1986). Altogether, the mean Cohens effect size on outcome

measures for the adolescent and teen studies is medium (d = .59; SD = .40; 95% CI

.25-.93)

Finally, nine studies included in the meta-analysis utilized heterogeneous

samples of mixed-age youth, ranging from young children to teen-agers. The total mean

Cohens effect size for the group is medium (d = .58; SD = .25; 95% CI .38-.77). Two

exceptions with large effect sizes include the study by Sullivan et al. (1992) utilizing

individual psychotherapy conducted in sign language for 35 sexually abused deaf youth,

and the study by Huang-Storms et al. (2007) utilizing neurofeedback with maltreated

youth. As a group, the mixed-age studies described heterogeneous treatment

54
interventions and intervention combinations. It is notable within this group that longer

treatment durations did not generally correlate with better outcomes on global measures

of behavior. It appears that the most effective treatments of the mixed-age studies

utilized more focused treatment interventions (rather than multiple treatment

approaches), and/or were geared toward a more specific population of youth (e.g., deaf

students).

Participant Characteristics: Gender, Ethnicity, and Socio-economic Status

The majority of the studies in the meta-analysis included both boys and girls in

their samples. These 24 mixed studies were diverse in participant demographics and

treatment interventions, and have an overall medium Cohens effect size (d = .67; SD =

.24; 95% CI .47-.87). Eight of the studies included in the meta-analysis had female only

samples, and one utilized a boys-only sample. The female-only studies were all

designed for older children (between the ages of 9 and 17 years) who had experienced

sexual abuse. All but one of these utilized a group treatment intervention, including one

drama therapy group (Mackay et al., 1987). The mean Cohens d effect size for the all-

girls studies is large (d = .84; SD = .32; 95% CI .57-1.10). However, this is not

statistically significantly higher than the mixed-gender studies, F(1,30) = .88; p = .36.

The single all-male study by Hack et al. (1994) was also designed for group treatment of

sexual abuse, and had a very small sample size and overall small effect size (d = .15)

on outcome measures (see Figure 6).

Eighteen of the studies included in the meta-analysis provided statistics

regarding the racial/ethnic makeup of the participants, and seven provided socio-

55
economic information. There are no discernable patterns in terms of effectiveness of

treatment based on socio-economic or racial-ethnic distinctions among the studies.

2.5

2.0 29
22

1.5

1.0

.5

0.0

-.5
D

N= 8 1 24

girls boys mixed

SEX

Figure 6. Effect size means for studies grouped by gender.

Study Characteristics: Empirical Design

Just under half of the empirical studies included in the meta-analysis (13 studies)

utilized a single-group repeated measure design. In other words, one sample group of

children was initially assessed pre-treatment using a standardized measure(s), then the

entire group received the treatment intervention(s), and finally the group was re-

assessed for changes on the standardized measure(s) post-treatment. Sometimes,

long-term follow up post-testing was also part of the methodology (i.e., longitudinal

design). This design methodology is typical in clinical settings where control groups of

children are not readily available, and it enables researchers to investigate whether

56
treatment is associated with changes on specific outcome measures. Of course,

although correlation may be inferred using this design, causal relationships between

treatments and changes in targeted outcomes cannot be established without a control

group. Confounding variables such as maturational and placebo effects cannot be ruled

out as alternative explanations for differences between pre- and post- measures.

However, due to ethical concerns regarding denying or delaying treatment for children

identified as maltreated, single group repeated measure designs currently represent a

compromise between clinical pragmatism and research vigor.

The group of 13 single group repeated measure studies included in the meta-

analysis represents a heterogeneous mixture of treatment interventions and sample

characteristics, as well as an array of effect size outcomes ranging from negligible to

very large. Again, group therapy interventions generally found stronger results on

outcome measures for sexually abused children, while theraplay and neurofeedback

demonstrated the strongest effect size outcomes for abused and neglected children.

The mean Cohens d effect size for the entire group of 13 is medium (d = .59; SD = .5;

95% CI .29-.90).

The second research design category represented in the meta-analysis is quasi-

experimental, or comparison group design. Specifically, theses studies utilized two or

more sample groups (sometimes randomly assigned) to assess the differential effects of

two or more treatment interventions on a specific pre-post- measure(s). Because they

lack a control group, these studies are similar to single group repeated measure studies

in that they can assert a correlation between treatments and targeted outcome

measures, but cannot assert causation. Ten of the studies included in the meta-analysis

57
utilized this design methodology. Again, treatment approaches and participant

demographics were heterogeneous, and effect size outcomes varied from small to

large. The mean Cohens d effect size for the group is medium (d = .60; SD = .29; 95%

CI .40-.80).

Some interesting tentative findings arose in the comparison group studies

regarding the effectiveness of differential treatment interventions. For example, Berliner

and Saunders (1996) unexpectedly found that adding a stress inoculation training

component to structured group therapy for sexually abused children did not improve the

participants outcomes on measures of fear or anxiety. On the other hand, Cohen &

Mannarino (1996; 1998) found significantly better outcomes on global behavior and

anxiety/depression measures using sexual abuse-specific cognitive-behavioral therapy,

as compared to using non-directive supportive therapy with individual children. Nasrin et

al. (2004) found significantly more improvement on PTSD symptom outcomes for a

small group of sexually abused Iranian girls using EMDR, as compared to individual

manualized cognitive-behavioral therapy. Clendenon-Wallen (1991) found that the

addition of music therapy to a mixed-gender sexual abuse therapy group did not

enhance results on an outcome measure of self-esteem. In another study, structured

individual play therapy did not improve childrens scores on developmental inventories

above and beyond changes observed in a multi-modal therapeutic daycare treatment

setting (Reams & Friedrich, 1994). And finally, Nolan et al. (2002) found no significant

gains on depression/anxiety or global behavior measures with the addition of non-

manualized group therapy to supportive individual therapy for mixed-gender samples of

sexually abused children.

58
 The third and last research design category represented in the meta-analysis is

experimental control group design. These 10 studies represent the most desirable of

empirical designs, as their utilization of a control group allows for a level of causal

inference about the relationship of the treatment intervention to changes on targeted

outcome measures. In particular, five studies represent the gold standard in that they

utilized a randomized control group and had sample sizes of at least 12 participants per

group (Deblinger et al., 1996; Fantuzzo et al., 1988; King et al., 2000; Perez, 1987;

Sullivan et al., 1992). Specifically, Deblinger et al. (1996) and King et al. (2000) both

investigated cognitive-behavioral family therapy and cognitive-behavioral/trauma-

focused individual therapy for mixed-aged survivors of sexual abuse. The studies used

a variety of outcome measures, and both found medium to large effect sizes for both

treatment interventions. Fantuzzo et al. (1988) and Perez (1987) both investigated play

therapy interventions (group, individual, peer-dyad, and adult-child dyad), and both

studies found medium to large effect size outcomes overall for young, maltreated

children. Finally, Sullivan et al. (1992) found large effect size outcomes for individual

psychotherapy in sign language for a mixed-age sample of sexually abused youth.

Other control group designs in the meta-analysis utilized matched, waitlist, or

non-randomized community control groups. Three investigated group therapy

treatments (De Luca et al. 1995; McGain & McKinzey, 1995; Verleur et al., 1986), and

two investigated multi-modal treatment interventions (Culp et al. 1991; Rust & Troupe,

1991). Taken as a group, the mean effect size for the control group design studies is

quite large (d = .92; SD = .45; 95% CI .60-1.25). It is encouraging that among the three

research design categories included in the meta-analysis, the most stringent

59
demonstrates the strongest treatment results on outcome measures. Differences in

effect size between the three research design groups were not robust enough to reach

statistical significance, however, F(2,30) = 2.0; p =.15 (see Figure 7).

2.5

2.0 30
23

1.5

1.0

.5

0.0

-.5
D

N= 13 10 10

repeated measures comparison group control group

DESIGN

Figure 7. Effect size means for studies grouped by design.

Study Characteristics: Length of Treatment

Length of treatment is a particularly relevant study variable, as it has practical

implications regarding the cost and feasibility aspects of treatment intervention. Large

effect sizes on outcomes for interventions of shorter duration may well be more readily

utilized by families and clinicians than those with similar change outcomes but more

extensive treatment requirements. Indeed, the importance of efficiency and cost

effectiveness is the likely reason why there has been more empirical research on group

treatment interventions than on individualized treatments.

60
 In terms of total treatment duration, the 33 studies included in the meta-analysis

ranged from 2 weeks (Kruczek & Vitanza, 1999) to 1 year (Humberson, 1998), with a

mode of 12 weeks and a mean of 14.7 weeks. Incidentally, Humbersons (1998) study

of parallel group therapy for 4-8 year olds and their non-offending caregivers boasts the

largest effect size of all the studies included in the meta-analysis (d = 1.95). The second

largest effect size outcome (d = 1.90) was found in McGain & McKinzeys (1995)

matched control group study, which was the second longest duration in the meta-

analysis at 6-9 months. Indeed, it appears that for group therapy interventions,

durations of 12 or more weeks up to 1 year are correlated with incrementally improved

outcome results, in general. This is distinct from the individual treatment interventions in

the meta-analysis, which did not gain consistently on outcome effect sizes with

durations extending beyond 12 weeks.

Duration of treatment is not a singularly valid way to assess treatments in terms

of time, energy, and resources, however. For example, the 1 year study by Humberson

(1998) involved approximately 96 hours of actual treatment intervention, whereas

Stubenborts (2000) theraplay intervention study had a duration of only 12 weeks, but

the equivalent of over 180 client contact hours. Actual time spent receiving/delivering

therapeutic interventions may well be a better gauge of dose effects and feasibility of

treatment. For this reason, individual client contact time is included in the summary

tables in Appendix B.

A two-tailed Pearson correlational analysis of the relationship between client

contact hours and Cohens effect size was calculated for the 33 studies in the meta-

analysis. Overall, there was no correlation between the two variables (r = -.005, p =

61
.98), indicating that greater therapy contact hours were not clearly related to greater

changes on targeted outcome measures. For studies specifically examining treatments

for sexually abused youth, the relationship between effect size and contact time was

somewhat stronger but also not significant (r = .38, p = 07); nor was it significant for

studies designed for physically abused and neglected children (r = -.05, p = .90). On the

other hand, separating out the group therapy intervention studies, the relationship

between contact hours and effectiveness of treatment was significant at the .05 level (r

= .46, p = .03). This is congruent with the findings discussed above of improved

outcomes with longer group treatment durations, and the general superiority of group

treatment interventions for sexually abused youth, in particular.

Study Characteristics: Treatment Intervention

It was originally hoped that enough studies would meet inclusion criteria to allow

for meaningful comparison of different treatment interventions for abused and neglected

children. Unfortunately, empirical research on the topic is clearly in its nascent stages,

and the present sample of 33 studies is unbalanced in terms of treatment

representation. Because a number of the studies included separate outcome data for

two or more distinct interventions, a total of 38 treatment trials were analyzed from the

33 studies. These include family therapy (N = 2), individual therapy (N = 8), play therapy

(N = 5), group therapy (N = 15), multi-modal therapy (N = 5), and miscellaneous

experimental therapies (N = 3). Statistical comparison of the group means of these six

treatment categories yielded non-significant results, F(5,32) = .44; p = .82, indicating

that they do not differ meaningfully from each other in terms of their efficacy (see Figure

8).

62
 2.5

2.0

1.5

1.0
26

.5

0.0

-.5
D

N= 2 10 5 3 5 15

family individual play misc combined group

TX

Figure 8. Effect size means for studies grouped by type of treatment.

The two family treatment interventions were components of larger controlled

studies for sexually abused youth that compared cognitive behavioral/trauma-focused

family therapy outcomes to cognitive-behavioral individual therapy outcomes (Deblinger

et al., 1996; King et al., 2000). Trauma-focused family therapy had a medium effect

size, and cognitive-behavioral family therapy had a large effect size, making for a large

combined family therapy average (d = .80; SD = .29; 95% CI -1.8-3.40).

Of the individual therapy samples, five utilized cognitive-behavioral approaches

(Cohen & Mannarino, 1996; Cohen & Mannarino, 1998; Deblinger et al., 1996; King et

al, 2000; Nasrin et al., 2004) and five utilized non-directive/supportive therapies

(Cohen & Mannarino, 1996; Cohen & Mannarino, 1998; Nolan et al., 2002; Baker, 1985;

Sullivan et al., 1992). Altogether, the average Cohens effect size of these 10 studies

63
was medium (d = .67; SD = .18; 95% CI .14-.81). The average effect size of the

cognitive-behavioral interventions was medium (d = .73) and the average effect size of

the non-directive interventions was medium (d = .54); this difference in effect sizes is

not statistically significant, t(8) = 1.19; p = .29.

The five play therapy interventions included a trained peer-dyad and adult-dyad

play comparison (Fantuzzo et al., 1988), an educational theraplay intervention

(Stubenbort, 2000), and three non-specific individual and group play therapy

interventions (Cotton-Cornelius, 2000; Perez, 1987; Zion, 1999). The average effect

size for the play therapy interventions was medium (d = .64; SD = .41; 95% CI .14-

1.15).

Combined therapies represented the lowest combined effect size average of the

different treatment types (d = .49; SD = .27; 95% CI .15-.83). These studies included a

structured play therapy and milieu therapy combination (Reams & Friedrich, 1994), a

multi-modal parent support plus therapeutic day treatment (Culp et al., 1991), a

combination of individual, family, and group therapies (Lanktree & Briere, 1995), and a

combination of group plus individual plus parent group therapy (Rust & Troupe, 1991).

Of the group therapy studies, 13 were aimed at youth with histories of sexual

abuse and are described above. The remaining two group intervention studies for

physically abused and neglected children investigated parallel cognitive-behavioral

therapy for young children and their caregivers (Stauffer & Deblinger, 1996) and

compared group treatment to a control group (McGain & McKinzey, 1995). Their large

mean effect size (d = 1.21) averaged with the other 13 group therapy studies for

64
sexually abused youth makes for an overall large group therapy effect size (d = .80; SD

= .27; 95% CI .49-1.11).

The three studies making up the miscellaneous/experimental treatment group

include a neurotherapy intervention for maltreated children by Huang-Storms et al.

(2007; d = .82), an EMDR intervention for Iranian adolescents with PTSD by Nasrin et

al. (2004; d = 1.07), and an outdoor adventure group intervention for maltreated

adolescents by Tucker (2006; d = .08). As a group, these studies have a mean medium

effect size (d = .66; SD = .51; 95% CI -.6 -1.9).

Summary

The grand total average effect size for the 33 studies included in the meta-

analysis was medium (d = .70). No particular therapeutic interventions stood out as

statistically more effective than the others, though some trends may be tentatively

speculated from the data. Cognitive-behavioral approaches appeared superior to non-

directive/client-centered approaches, overall; however, this may be an effect of the

larger number of CBT-oriented studies represented in the meta-analysis. Because the

meta-analysis also included a majority of group therapy studies, data on group

treatment can likely be interpreted with most confidence.

Overall, group therapy was associated with medium to very large effect size

changes on measures of behavior, with longer treatment durations producing stronger

effects. Specifically, it appears that for group therapy interventions, durations of 12

weeks up to 1 year are correlated with incrementally improved outcome results,

particularly for youth with histories of sexual abuse. Individual treatment interventions,

65
on the other hand, did not gain consistently on outcome effect sizes with increased

client contact hours or treatment durations.

Other variables that were hypothesized as potentially relevant to treatment

outcome were, across the board, not statistically significant contributors to the

effectiveness of treatment. Type of abuse (sexual or physical/neglect) and age group

did not appear to impact outcomes, nor did mixed vs. all-female sample group makeup.

Unfortunately, there were not enough data available to evaluate the impact of socio-

economic status or racial background. There was also only one all-male study, so

gender group comparisons were not possible. Outcome measures varied so greatly

between studies that no meaningful patterns could be discerned regarding the types of

outcome measures most responsive to client changes following treatment.

Controlled studies made up just under one-third of the studies in the meta-

analysis and their authors reported strong positive outcomes overall. Although this

limited number of studies (culled from over 30 years of literature) indicates there is a

great need for further empirical research, these studies do build a foundation of support

for the efficacy of therapies for abused and neglected children.

66
 CHAPTER 6

DISCUSSION

There were five major aims of the present two-part study: (1) to explore through a

clinical pilot study the effectiveness of neurofeedback training for a sample of children

with histories of abuse and neglect; (2) to establish through meta-analysis the average

effect size of treatment outcomes for abused and neglected children, overall; (3) to

ascertain whether treatment outcome data for children with histories of abuse and

neglect are as robust as those for the broader population of children; (4) to ascertain

whether certain treatment modalities are associated with stronger effect sizes for this

population; and (5) to compare the efficacy of neurofeedback training to other

empirically substantiated treatment modalities for abused and neglected children. All of

these goals were met by the present study. However, the scarcity of empirical research

on treatment outcomes for this population of children makes it difficult to answer with

confidence the question of the comparative effectiveness of different treatment

modalities.

Comparison of Findings to Previous Research

The medium-sized total average Cohens effect size of the 33 studies included in

the meta-analysis (d = .70) falls within the range of three broad-based meta-analyses of

general child psycho-therapy interventions. Weisz et al. (1995) looked at 150 child and

adolescent psychotherapy outcome studies and reported a mean Cohens effect size of

.54 (medium); Kadzin et al. (1990) analyzed 223 studies and reported a mean Cohens

effect size of .88 (large); and Casey and Berman (1985) reported a mean effect size of

.71 (medium) for 75 studies with children 12 years of age and younger. Reeker et al.

67
(1997) found a somewhat larger aggregate effect size in their meta-analysis of 15 group

therapy studies for sexually abused children (d = .79), and Hoag and Burlingame (1997)

found a slightly smaller aggregate effect size for their meta analysis of 56 group therapy

studies for a general population of children and adolescents (d = .61). The present

results are therefore in line with previous research; moreover they are encouraging, as

abused and neglected children have been described in the literature as less responsive

to treatment than the general pediatric population (Chanitz, 1995; Wilson, 2001). Based

on these effect size outcomes, it appears that a variety of empirically researched

treatment interventions have clear benefits for maltreated children.

Although further research is greatly needed, the present meta-analysis tentatively

indicates that various forms of psychotherapy are similar in effectiveness with this

population. This is congruent with broad-based adult psychotherapy meta-analyses,

which also suggest the Dodo verdict, i.e., everybody has won and all must have

prizes (Parloff, 1984; Smith, Glass & Miller, 1980). Although evidence has been mixed,

larger meta-analyses have suggested that behavioral methods yield larger effect sizes

for the general population of children than non-behavioral methods (Casey & Berman,

1985; Weisz et al., 1995). Although not statistically significant, effect size trends in the

present meta-analysis also suggest that cognitive-behavioral approaches are correlated

with stronger outcomes than non-directive therapy interventions for abused and

neglected children. As mentioned above, group approaches further demonstrated a

trend toward increased effectiveness with sexually abused youth, in particular.

It may be important to note which modalities of therapy have not been the subject

of empirical studies, and therefore were not included in the meta-analysis. For example,

68
a great deal of attention has been paid in recent years to holding therapy as a treatment

intervention for children with reactive attachment disorder (RAD), in particular. No

holding therapy studies meeting the minimum research criteria were found during the

literature search. Similarly, parent attunement and re-birthing therapies were nowhere to

be found in the research literature.

It has also been debated among researchers whether the age of the child

impacts psychotherapy outcomes. Some have suggested that younger children respond

more cooperatively to adult authority (Kendall, Lerner, & Craighead, 1984), while others

suggest that cognitive maturity favors psychotherapy for adolescents and teens (Shirk,

1988). The evidence on age and therapy outcomes has been mixed thus far. Casey and

Berman (1985) found no relationship between child age and study effect size, but Weisz

et al. (1987) found a significant negative relationship suggesting older youth do less well

in therapy than their younger counterparts. Trends in the current meta-analysis, though

not to statistical significance, suggest that 9-13 year olds make more positive gains on

outcomes than younger and older age groups. Research has been similarly mixed thus

far regarding gender effects on psychotherapy outcomes for youth. Unfortunately the

present meta-analysis cannot offer much clarification on the topic due to the paucity of

all-male studies; however, trends from the data suggest that all-girl study samples had

better outcomes than mixed group samples.

Clinical Implications for Neurofeedback

With a large mean effect size of .82, the neurofeedback pilot study outcomes

stand shoulder to shoulder with other more established treatment interventions for

children with abuse and neglect histories. The results indicate that neurotherapy

69
treatment was more effective than the average treatment intervention for this particular

population, as based on the aggregate meta-analysis effect size mean of d = .70. This is

particularly encouraging when client contact hours are considered an important factor in

treatment feasibility. With total contact time averaging 15 hours and treatment duration

averaging five months, the neurotherapy pilot study represents one of the more efficient

therapeutic interventions in the meta-analysis.

There were 12 studies included in the meta-analysis with equal or higher effect

size outcomes compared to the neurotherapy pilot study. Eleven of these studies were

designed specifically for youth who had experienced sexual abuse, and seven of these

were group therapy interventions. As discussed above, group therapy appears from the

present meta-analysis to be the most effective treatment for this sub-population of

abused children, potentially because of its success in normalizing the sexual abuse

experience through group sharing and psycho-education. The other four studies

included treatment of sexual abuse victims through individual cognitive behavioral

interventions (two studies), individual or group play therapy (one study), and eye

movement desensitization and reprocessing therapy (one study).

Only one study designed for the treatment of abused and neglected children

reported stronger effect size outcomes than the neurotherapy pilot study. This was the

theraplay intervention designed for maltreated preschoolers (Stubenbort, 2000; d = 95).

Compared to the neurotherapy intervention, the theraplay intervention was significantly

more extensive in terms of contact time and multi-modal support for the child clients.

Theraplay was described by the studys authors as providing developmentally

appropriate therapeutic and educational activities in a therapeutic pre-school setting, 3

70
hours per day, 5 days per week for 12 weeks (totaling approximately 180 contact

hours). Outcomes were measured in terms of improvements on broad developmental

inventories. The neurotherapy intervention, by contrast, was the equivalent of

approximately 15 contact hours and treated an older and broader age range from 6 to

15.5 years.

According to the meta-analysis by Weisz et al. (1995), social aggression is

typically least amenable to change following psychotherapy interventions for children.

Interestingly, the neurotherapy pilot study indicates that aggressive, delinquent, and

socially problematic behaviors (e.g., lying, fighting, demanding attention, etc.) were

most strongly impacted by treatment. Other behaviors associated with attentional

problems, anxiety/depression, and thought problems also were significantly reduced.

The global nature of the changes observed following neurotherapy treatment may well

be one of its distinct assets as a treatment modality. Changes observed in both

physiological and behavioral outcomes following the pilot treatment can be

characterized as anchored in improved emotional self-regulation. As neglect and abuse

represent diverse experiences with a common component of autonomic nervous system

(ANS) dysregulation (rather than a specific symptom-based diagnosis), it may be that

the self-regulating effects of neurotherapy are particularly well-suited to this population

of children.

Study Strengths and Limitations

Unlike the large meta-analyses referred to above, a strongpoint of the present

study is that it did not rely solely on published reports. Although published studies afford

meta-analyses a level of quality control, relying solely on them may introduce a

71
distortion of effect sizes due to publication bias. Ample evidence suggests that journal

editors and reviewers favor publications of statistically significant findings (e.g.,

Atkinson, Furlong, & Wampold, 1982). McLeod and Weisz (2004) suggest that,

paradoxically, large meta-analyses requiring publication for inclusion may actually skew

outcomes too positively. They further suggest that dissertations entail a high level of

peer review via the committee process, and therefore constitute a methodologically

sound alternative sample that is less vulnerable to the "file drawer effect." The present

meta-analysis included approximately one-quarter dissertations, with effect size findings

similar to those of previously published meta-analyses.

Another strength of the present meta-analysis is its heavy inclusion of clinical

studies, as opposed to majority inclusion of laboratory research studies. This is atypical

of meta-analyses, and was possible due to the leniency of the inclusion criteria in

accepting non-controlled designs. There has been criticism regarding the over-

representation of laboratory studies in psychotherapy outcome literature (e.g., Weisz et

al., 1995). Specifically, clinical studies are reported to have weaker outcomes than

laboratory research studies due to the real-life complexity of their participants in terms

of heterogeneity and co-morbidity of diagnoses. Typically, laboratory research studies

differ significantly from clinical studies in terms of their well-screened (or recruited)

samples, better retention rates, manualization of treatments, and controlled treatment

durations. It has been suggested that this may unrealistically inflate effect sizes found in

most meta-analyses. The majority of the studies included in the present meta-analysis

were based in community mental health centers or other clinical settings; thus, the

72
robust effect sizes bode well for the positive impact of therapeutic interventions in

clinical as well as research settings.

Drawbacks to the meta-analysis design have been discussed above in the meta-

analysis methods section primarily the risk of inflated effect size findings when

including single treatment group repeated measures designs and small sample sizes in

a meta-analysis (Lipsey & Wilson, 1993). However, efforts were made in accordance

with other meta-analyses to offset this potential problem by using effect size

calculations that were less vulnerable to inflation and by adjusting statistically for

sample sizes.

There are caveats to the neurofeedback pilot study as well, as described above

in the pilot study discussion section - chiefly regarding the non-experimental research

design and the inability to control for medication effects. However, these design

weaknesses are unfortunately the norm amongst clinical studies with this population, so

are not reasonable grounds for minimizing the comparative strength of the

neurofeedback outcome data. In fact, the pilot sample size was comparable to or larger

than most studies included in the meta-analysis, and the outcome measures had the

unique advantage of both a broad behaviorally-based parent report (Child Behavior

Checklist) and a physiological assessment (Test of Variables of Attention). As a first

study of this kind, the results are necessarily tentative, but very encouraging.

It is hoped that the present meta-analysis will give impetus to further outcome

research for this population, and may serve as a practical yardstick comparison to

gauge the efficacy of treatment interventions in the future. It is also hoped that the

positive outcomes of the pilot study will stimulate interest in the potential of

73
neurofeedback treatment for the many children and families struggling with the

developmental effects of abuse/neglect. As neurotherapy is generally utilized as an

adjunctive treatment, one clinical conclusion that could be drawn from the present study

is the suggestion that adding neurofeedback to an existing proven multi-modality

therapy, such as theraplay, might provide a very a powerful bio-psycho-social approach

for treating this challenging population.

74
 APPENDIX A

INTERNATIONAL 10-20 ELECTRODE PLACEMENT SYSTEM

75
(Looking from above the head)

Front of head

Back of head

76
 APPENDIX B

SUMMARY TABLES OF STUDIES INCLUDED IN META-ANALYSIS

77
 APPENDIX B TABLE OF CONTENTS

Study Page
Ashby, Gilchrist & Miramontez (1987)....79
Baker (1985)...80
Berliner & Saunders (1996)..81
Clendenon-Wallen (1991).83
Cohen & Mannarino (1996)..84
Cohen & Mannarino (1998)..85
Cotton-Cornelius (2000)86
Culp, Little, Letts & Lawrence (1991)......87
Deblinger, Lippman, J., & Steer, R. (1996)....88
De Luca, Boyes, Grayston & Romano (1995)...89
Fantuzzo et al. (1988)90
Hack, Osachuk & De Luca (1994)...91
Hall-Marley & Damon (1993)....92
Huang-Storms, Bodenhamer-Davis, Davis & Dunn (2006).....93
Humberson (1998).....94
King,Tonge et al. (2000)95
Kruczek & Vitanza (1999).....96
Lanktree & Briere (1995)...97
Mackay, Gold & Gold (1987)....98
McGain & McKinzey (1995)..99
Nasrin et al. (2004)......100
Nolan et al. (2002)101
Perez (1987).....103
Reams & Friedrich (1994)...104
Rust & Troupe (1991)..105
Stauffer & Deblinger (1996)106
Stubenbort (2000)....107
Sullivan, Scanlan, Brookhouser & Schulte (1992)......108
Thun, Sims, Adams & Webb (2002)..109
Tucker (2006)110
Verleur, Hughes & Dobkin de Rios (1986)...111
Waters (1998)...112
Zion (1999)113

78
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Ashby, Group therapy (with N = 11 Piers-Harris Childrens Self- d = 1.0 d = 1.0

Gilchrist & culturally relevant Concept Scale (Piers & (large)
Miramontez activities) Age: 12 - 17 ys. Harris, 1984)
(1987) (Child)
Sessions: Gender:
Design: 2 hours/week 100% girls
Single
group Duration: Inclusion criteria:
Repeated 10 weeks Hx of sexual
measures abuse

Treatment Ethnicity:
Setting: Native American
Community
church SES:
(Contact time: 20 hours) n/a

79
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Baker (1985) Short term group N = 39 1. Piers Harris Childrens Group CSS CSS total
therapy (unstructured) Self Concept Scale (Piers & d = 2.0 d = .70
Design: (n = 48) Age: 13 - 17 ys. Harris, 1984) Individual CSS
Randomized (M = 14.7) (Child) d = .91 ASI total
comparison OR d = .75
group Gender:
Client centered 100% Girls DSI total
individual therapy 2. Anxiety Scale Index Group ASI d = .20
Treatment (n = 32) (Institute for Personality d = .22
Setting: Inclusion criteria: and Ability Testing) Individual ASI Group
University Sessions: CPS verified Hx of (Child) d = .68 total
counseling 1 hr/week Individual sexual abuse d = .81
center 1.5 hr/week group
Ethnicity: Individual
Duration: n/a 3. Depression Scale Index Group DSI total
10 weeks Individual (Institute for Personality d = .21 d = .89
6 weeks Group SES: and Ability Testing) Individual DSI
n/a (Child) d = 1.1 Study
total
d = .85
(large)

(Contact time:
10 hours Individual
9 hours Group)

80
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Berliner & Group therapy N = 80 1. Fear Survey Schedule for FSSC-R SIT total FSSC-R
Saunders (structured) Children Revised (FSSC-R) d = .21 total
(Ollendick, 1983) (Child) FSSC-R GT total
(1996) Age: 4 - 13 ys. d = 1.19 d = .70
WITH
Design: Gender: Failure/Criticism SIT d = .22 SAFE
Randomized Stress Inoculation 88% girls GT d = 1.16 total
comparison Training (SIT) 22% boys Fear of Unknown SIT d = .08 d = .75
GT d = 1.19
group Injury/Small Animals SIT d = .25
(n = 48) Inclusion criteria: GT d = 1.11 RCMAS
CPS verified Hx of Danger/Death SIT d = .38 total
Treatment OR sexual abuse GT d = 1.10 d = .20
Setting: Medical Fears SIT d = .17
GT d = .98
Community Group therapy Ethnicity: CBCL
mental health (structured) 74% Caucasian 2. Sexual Abuse Fear SAFE SIT total total
clinic 11% African Evaluation Scales (SAFE) d = .40 d = .68
(n = 32) American (Wolfe & Wolfe, 1986) (Child) SAFE GT total
7% Latina/o d = 1.11 CDI total
Sessions: 8% Other Sex Associated Fears SIT d = .14 d = .59
8 10 GT d = 1.14
Interpersonal Discomfort SIT d = .66
SES: n/a GT d = 1.09
CSBI total
Duration: d = .48
10 weeks 3. Revised Childrens Manifest RCMAS SIT total
Anxiety Scale (RCMAS) d = .15
(Contact time: ~9 hours) (Reynolds & Richmond, 1985) RCMAS GT total d
(Child) = .24
(table
continues)

81
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Berliner & Note: Physiological Anxiety SIT d = .18 SIT total

Saunders The addition of Stress GT d = .77 d = .35
Worry/Oversensitivity SIT d = .01
(1996) Inoculation Training GT d = .58
(continued). (SIT) did not result in Concentration Anxiety SIT d = .18 GT total
significant GT d = .52 d = .78
improvements over
Group Therapy (GT) 4. Child Behavior Checklist CBCL SIT total Study
alone. (CBCL) Achenbach & d = .47 total
CBCL GT total d = .56
Edelbrock, 1983) (Parent) d = .89
(medium)
Social SIT d = .27
GT d = 1.1
Internalizing SIT d = .37
GT d = .58
Externalizing SIT d = .22
GT d = .67

5. Childrens Depression CDI SIT total

Inventory (CDI) (Kovacs, d = .66
1992) (Child) CDI GT total
d = .52

CSBI SIT total

6. Child Sexual Behavior
d = .21
Inventory (CSBI) Friedrich, CSBI GT total
Luecke, et al., 1992) d = .74
(Parent)

82
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Clendenon- Therapeutic group N = 14 The Adjective Checklist AC total d = .32 Study

Wallen support (Gough & Heilbrum, 1983) Total:
(1991) Age: 14 - 18 ys. (Child) d = .32
AND (M = 15) (small)
Design: Indicators d = .25
Non- Music therapy Gender:
randomized (n = 3) 82% Girls Contra-indicators d = .39
comparison 18% Boys
groups OR

Therapeutic group Inclusion criteria:

support only Hx of sexual Note: No
Treatment (n = 11) abuse statistically
Setting: significant
YWCA Sessions: differences
Sexual 90 minutes/week Ethnicity: on the
Assault support group n/a outcome
Program (and) measure
45 minutes/week music SES: n/a were
therapy observed
between
Duration: the two
12 weeks groups

(Contact time: 27 hours)

83
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Cohen & 1. Sexual Abuse Specific N = 67 children 1. Child Behavior Checklist CBCL SAS-CBT CBCL
Mannarino Cognitive-behavioral Therapy (CBCL) (Achenbach & total d = 1.18 total
(SAS-CBT), Manualized Age: 2.11 7.1 ys. CBCL NST d = .57
(1996) Edelbrock, 1983) total d = .18
(M = 4.68) (Parent)
i.e., Individual therapy for
CSBI
Design: child
Internal
SAS-CBT d = 1.16
AND Gender: NST d = .01 total
Randomized 58% girls d = .68
comparison Individual therapy for non- External SAS-CBT d = .96
offending parent 42% boys NST d = .29
groups (n = 39 children) Social Competence SAS-CBT d = .36 PRESS
Inclusion criteria: NST d = .42 total
Treatment OR CPS-validated Hx of d=0
Setting: sexual abuse in past 2. Child Sexual Behavior CSBI SAS-CBT
Urban 2. Non-directive Supportive 6 months Inventory (CSBI) d = .92 WBR total
Therapy (NST) CSBI NST d = .44 d = .86
outpatient (Friedrich, Grambsch, et
psychiatric Exclusion criteria: al., 1992) (Parent)
i.e., Individual therapy for
child clinic child Psychosis, mental
AND retardation, serious SAS-CBT
3. Preschool Symptom
Individual therapy for non- medical illness, PRESS SAS-CBT total d =
pervasive
Self-Report (PRESS) .82
offending parent total d = 0
Note: The (Martini et al. 1990)
(n = 28 children) developmental delay PRESS NST
SAS-CBT (Child) total d = 0 NST total
group showed Sessions: d = .28
Ethnicity:
clinically (not Twelve 45-min sessions for 54% Caucasian
statistically) both child and parent 4. Weekly Behavioral Rating Study
42% African-
more (WBR) (Cohen & Mannarino, WBR SAS-CBT
Total:
American
improvement Duration: 1993) (Parent) total d = 1.15
12 - 16 weeks 40% Other WBR NST
d = .55
on outcome
total d = .62 (medium)
measures than
SES: n/a
the NST group
(Contact time: 18 hours)

84
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Cohen & 1. Sexual Abuse Specific N = 49 1. Child Behavior Checklist CBCL CBT CBCL
Mannarino Cognitive-behavioral Therapy (CBCL) (Achenbach & total d = .37 total
(SAS-CBT), Manualized CBCL NST d = .29
(1998) Age: 7 - 15 ys. Edelbrock, 1983) total d = .20
i.e., Individual therapy for (M = 11.1) (Parent)
Design: CSBI
child CBT d = .34
Internal NST d = .17 total
Randomized AND Gender:
Individual therapy for non- CBT d = .25 d = .29
comparison 69% girls External
offending parent NST d = .15
groups (n = 30 children) 31% boys Social Competence CBT d = .41 CDI
NST d = .21 total
Treatment OR d = .48
Setting: Inclusion criteria: 2. Child Sexual Behavior CSBI CBT d = .41
Urban 2. Non-directive Supportive CPS-validated Hx Inventory (CSBI) (Friedrich, CSBI NST d = .16 STAIC
Therapy (NST) total
outpatient of sexual abuse in Grambsch, et al., 1992)
psychiatric past 6 months (Parent) d = .71
i.e., Individual therapy for
child clinic child CDI CBT
AND Ethnicity: 3. Child Depression Index total d = .90
Individual therapy for non- CDI NST CBT total
59% Caucasian (CDI) (Kovacs, 1985) d = .63
offending parent total d = .05
Note: The 37% African (Child)
(n = 19 children)
SAS-CBT American STAIC CBT NST total
group showed Sessions: 2% Latina/o 4. State-Trait Anxiety total d = .83 d = .25
significantly Twelve 45-min sessions for 2% Bi-racial Inventory for Children STAIC NST
more both child and parent (STAIC) (Spielberger, total d = .58 Study
improvement 1973) (Child) Total:
Duration:
SES: CBT d = .89
on outcome State d = .44
measures than 12 weeks Predominantly NST d = .32
middle income Trait CBT d = .76 (medium)
the NST group
(Contact time: 18 hours) NST d = .83

85
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Cotton- Individual play therapy N = 18 Child Behavior Checklist CBCL Child Study
Cornelius (Achenbach & Edelbrock, and Parent Total:
(2000) Sessions: Age: 5 - 13 ys. 1983) means: d = .2
1/week (M = 9.5) (Child, Parent) (small)
Design:
Single-group Duration: Gender: Withdrawn d = .36
Repeated 12 18 weeks 43% Girls
Somatic d = .17
measures (M = 13) 57% Boys
(Longitudinal) Anxiety/Depression d=0
Inclusion criteria:
Social d = .03
Treatment Hx of abuse
Setting: and/or neglect, Thought d = .34
Catholic Foster care
Attention d = .39
Charities placement
community Delinquent d = .13
mental health Ethnicity:
Aggressive d = .3
32% Caucasian
10.7% Latina/o Sex Problems d = .006
57% African
Total Internalizing Bx d = .08
American
1% Other Total Externalizing Bx d = .31

SES: n/a
(Contact time: 13 hours)

86
Study Intervention(s) Participants Outcome Measure(s) Measure Mean Effect
(Response Perspective) Effect Size(s) Size(s)

Culp, Little, Therapeutic day N = 34 (children) 1. Perceived Competence PCSAS total d = .56
Letts & treatment program, and Social Awareness d = .45 (medium)
Lawrence (group milieu focus Age: 3.9 5.9 ys. Scale (Harter & Pike, 1984)
(1991) with speech, physical, (M = 4.9) (Child)
and play therapies)
Design: Gender: Cognitive Competence d = .67 Note:
Waitlist AND 44% girls The
control 66% boys Peer Acceptance d = .46 treatment
group Parent counseling, group
group therapy, crisis Inclusion criteria: Physical Competence d = .16 showed
intervention, and Hx of physical significantly
Treatment parenting education abuse and/or Maternal Acceptance d = .50 more
Setting: services neglect improvement
In-patient (n = 17) on outcome
psychiatric Ethnicity: 2. Early Intervention EIDP total measures
hospital OR 44% Caucasian Developmental Profile d = .67 than the
56% African (Bricker, 1982) control
No treatment American (Teacher) group
(n = 17)
SES: low income Cognitive d = .90
Sessions:
6 hours x 5 days/wk Perception/Fine Motor d = .46

Duration: Social Emotional d = .62

m = 8.9 months
Gross Motor d = .75
(Contact time: 1,068 hr)

87
Study Intervention(s) Participants Outcome Measure(s) Measure Mean Effect
(Response Perspective) Effect Size(s)
Size(s)
Deblinger, 1. Individual cognitive- N = 100 (children) 1. The Schedule for Individual Individual
Lippman, J., behavioral therapy, child Affective Disorders and child child CBT:
& Steer, R. only (n = 25) Age: 7 - 13 ys. Schizophrenia - School Age CBT: d = .84 (large)
(1996) OR (M = 9.89) Version d = 1.75
2. Family cognitive- (K-SADS-E) PTSD section Family Family CBT:
behavioral therapy, mother
and child (n = 25)
Gender: (Orvaschel et al., 1982) CBT: d = 1.0 (large)
Design: OR 83% girls (Child, Parent) d = 2.1
Randomized 3. Individual cognitive- 17% boys Study Total:
comparison behavioral therapy, mother d = .93 (large)
groups only (n = 25) Inclusion criteria: 2. Child Depression Individual
OR DYFS referred for Inventory (CDI) child
4. Community comparison sexual abuse, (Kovacs, 1985) CBT: Note:
Treatment group (n = 25) >2 PTSD (Child) d = .51 Treatment
Setting: symptoms Family groups did not
Community Sessions: CBT: differ
45 min/week (child significantly
mental health Ethnicity: d = .66
individual, mother overall from
clinic individual)
70% Caucasian
each other, but
OR 21% African were
90 min/week (mother and American 3. Child Behavior Checklist Individual significantly
child dyad) 7% Latina/o (CBCL) Externalization child stronger on
2% other Scale CBT: outcome
Duration: 12 weeks (Parent) d = .27 measures as
SES: n/a Family compared to
(Contact time: CBT: the community
Dyad 18 hours d = .30 comparison
Individual 9 hours) group

88
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

De Luca, 1. Structured group N = 70 1. Self Esteem Inventory d = 1.27 Study

Boyes, therapy (SEI) (Coopersmith, 1981) Total:
Grayston & Age: 9 - 12 ys. (Child) d = 1.07
Romano (n = 35) (large)
(1995) Gender:
OR 100% Girls 2. Revised Childrens d = .80
Design: Manifest Anxiety
Non- 2. No treatment Inclusion criteria: Scale/What I Think and
randomized Hx of intra-familial Feel Questionnaire (CMAS)
community (n = 35) sexual abuse (Reynolds & Richmond,
control group 1978)
Ethnicity: (Child)
n/a
Treatment Sessions:
Setting: 1/week SES: n/a 3. Child Behavior Checklist d = 1.16
University (CBCL) (Achenbach &
psychology Duration: Edelbrock, 1983)
clinic 9 - 12 weeks (Parent)

(Contact time: ~10.5

hours)

89
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Fantuzzo, 1. Play sessions with a N = 36 1. The Preschool Behavior Peer play: Peer play:
Jurecic, trained socially initiating Questionnaire (PBQ) d =.17 d = .44
Stovali, peer Age: 3.1 5.3 ys. (Behar & Stringfiled, 1974) Adult play: (medium)
Hightower, (n = 12) (M = 4.3) (Teacher) d =1.6
Goins, & Adult Play:
Schachtel OR Gender: d =1.10
(1988) 28% girls (large)
2. Play sessions with a 72% boys
trained socially initiating 2. The Brigance Diagnostic Peer play: Study
Design: adult Inclusion criteria: Inventory of Early d = .72 Total:
Randomized (n = 12) Dept. of Social Development (Brigance, Adult play: d = .76
comparison Services referred 1978) d = .59 (large)
and control OR for abuse (24%), (Teacher)
groups neglect (46%), Note:
3. No treatment and high risk Both
(n = 12) (30%) treatment
Treatment groups
Setting: Ethnicity: showed
Community Sessions: 46% Caucasian significantly
mental health 8 54% African better gains
clinic American on outcome
Duration: measures
3 - 4 weeks SES: below than the
poverty line control
(Contact time: 8 hours) group

90
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Hack, Semi-structured group N=6 1. Child Behavior Checklist CBCL total d = .15
Osachuk & therapy (CBCL) (Achenbach & d = 0.0 (small)
De Luca Age: 8 11 ys. Edelbrock, 1983)
(1994) Sessions: (Parent)
90 minutes x 1/week Gender:
Design: 100% boys Externalizing- d = 0.0
Single group Duration:
Repeated 12 weeks Inclusion criteria: Internalizing d = 0.0
measures Hx of sexual
abuse
Treatment 2. Child Depression d = .74
Setting: Ethnicity: Inventory (CDI) (Kovacs,
Community n/a 1986)
mental health (Child)
clinic, SES:
Canada n/a 3. Childrens Manifest d = 0.0
Anxiety Scale (CMAS)
(Reynolds & Richmond,
1978)
(Child)

4. Self-Esteem Inventory d = 0.0

(Contact time: 18 hours) (Coopersmith, 1984)
(Child)

91
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Hall-Marley Parallel group therapy N = 13 (children) 1. Child Behavior Checklist CBCL d = .92
& Damon, for children and non- (CBCL) (Achenbach & d = .92 (large)
(1993) offending parent Age: 4 - 7 ys. Edelbrock, 1983)
(2 separate groups) (Parent)
Design: Gender:
Single group Sessions: 54% girls
Repeated 90 minutes x 1/week 46% boys 2. The Child Sexual CSBI
measures (each group) Behavior Inventory (CSBI) d = .92
Inclusion criteria: (Friedrich et al., 1991)
Duration: Hx of sexual (Parent)
Treatment 13 weeks abuse
Setting:
Child Ethnicity:
guidance n/a
clinic
SES: middle
(Contact time: 39 hours) income

92
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Huang- Neurofeedback training N = 20 1. Child Behavior Checklist Total CBCL d = .82

Storms, (CBCL) (Achenbach & d = .94 (large)
Bodenhamer- Age: 6 15.5 ys. Edelbrock, 1983) (Parent)
Davis, Davis Sessions: (M = 10.4)
& Dunn .5 hr x 30 Externalizing d = .78
(2006) Gender:
Internalizing d = .59
Duration: 45% girls
Design: 2 8 months (M = 5) 55% boys Social problems d = 1.26
Single group
Repeated Inclusion criteria:
measures Hx of abuse
and/or neglect, 2. Test of Variables of Total TOVA
CPS removal from Attention (TOVA.) d = .70
Treatment biological home, (Greenberg, 1987)
Setting: QEEG for protocol (Clinician)
Private development
practice and Omission d = .77
University Ethnicity:
Commission d = .96
clinic 75% Caucasian
20% African- Variability d = .78
American
Speed d = .27
5% Latina/o

(Contact time: 15 hours) SES: n/a

93
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Humberson Group therapy - N = 40 (children) 1. Child Behavior Checklist d = 2.4 Study

(1998) Parallel groups for (Achenbach & Edelbrock, total
children and non- Age: 4-8 ys 1983) d = 1.95
Design: perpetrating caregivers (Parent) (very
Single group (using a psycho- Gender: large)
Repeated educational curriculum) 42% Boys
measures 58% Girls 2. Teacher Report Form d = 1.5
Sessions: (Achenbach & Edelbrock,
~ 48 (weekly sessions) Inclusion criteria: 1983)
Treatment Hx of sexual (Teacher)
Setting: Duration: abuse
Child 1 year
Guidance Ethnicity:
Clinic 62% Caucasian
38% Latina/o

(Contact time: SES: n/a

~ 96 hours)

94
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

King,Tonge, 1. Trauma focused N = 36 1. Fear Thermometer for Individual CBT: Individual

Mullen, cognitive-behavioral Sexually Abused Children d =.80 CBT:
Myerson, individual child therapy Age: 5.2 -17.4 ys. (Child) Family CBT: d = .63
Heyne, (n = 12) (M = 11.5) d =0.0 (medium)
Rollings, OR
Martin, & Gender: 2. Coping Questionnaire for Individual CBT:
2. Trauma focused
Ollendick cognitive-behavioral family 69% girls Sexually Abused Children d = .46 Family
(2000) therapy (n = 12) 31% boys (King et al., 2000) Family CBT: CBT:
(Child) d = .59 d =.59
Design: OR Inclusion criteria: (medium)
Randomized Sexual abuse Hx 3. Revised Childrens Individual CBT:
3. No treatment (n = 12)
comparison PTSD Manifest Anxiety Scale (R- d = .07
and wait-list Sessions: CMAS) (Reynolds & FamilyCBT: Study
control Twenty 50-min sessions Ethnicity: n/a Richman, 1978) (Child) d = .29 Total:
groups SES: n/a d = .61
Duration: 20 weeks 4. Childrens Depression Individual CBT: (medium)
Treatment Inventory (CDI) (Kovacs, d = .39
Setting: Note: The two treatment 1981) (Child) Family CBT:
Community groups did not differ d = .47
significantly from each other
mental health on outcome measures, but
clinic both were significantly higher 5. Anxiety Disorders Individual CBT:
on outcome measures than Interview for DSM-IV d = 1.46
the control group. (ADIS) (Silverman & Family CBT:
Albano, 1996) d = 1.6
(Contact time: ~ 17 hours)
(Clinician)

95
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Kruczek & Structured Solution- N = 41 1. The Solution Focused d = .73 d = .46

Vitanza Focused group therapy Recovery Scale for (medium)
(1999) Age: 13 - 17 ys. Survivors of Sexual Abuse
(M = 14.21) (Dolan, 1991)
Design: Sessions: (Child)
Single group 7 Gender:
Repeated 48% girls
measures Duration: 52% boys 2. The Skill Mastery Test d = 0.0
~ 2 weeks (SMT) (Kruczek, 1995)
Inclusion criteria: (Child)
Treatment Hx of sexual
Setting: abuse
In-patient
psychiatric Ethnicity:
hospital 61% Caucasian
32% African
American
2% Latina/o

(Contact time: 7 hours) SES: low income

96
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Lanktree & 1. Abuse-focused N = 105 1. Child Depression d = .43 Study

Briere (1995) individual therapy Inventory (CDI) Total:
Age: 8 - 15 ys. (Kovacs, 1983) d = .49
Design: AND (M = 11.6) (Child) (medium)
Single group
(mixed 2. Family therapy Gender:
treatment) 85% Girls 2. Trauma Symptom
AND/OR
Repeated 15% Boys Checklist for Children
measures 3. Group therapy (TSCC) (Briere, 1995)
Inclusion criteria:
AND/OR
Hx of sexual Anxiety d = .74
Treatment 4. Parental abuse
Setting: conjoint/group therapy Depression d = .81
Hospital Ethnicity:
outpatient Sessions: 43% Caucasian Post-traumatic Stress d = .52
treatment 1/week 31% Latina/o
center 26% Other Sexual Concerns d = .26
Duration:
~ 12 weeks SES: n/a Dissociation d = .45

Note: Effect size statistics Anger d = .47

could not be calculated for
6, 9, and 12 months of Total d = .54
treatment

(Contact time: ~36 hr)

97
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Mackay, Structured group drama N=5 1. Beck Depression Inventory BDI d = .85 d = .55
Gold & Gold therapy (BDI) (Beck, 1978) (Child) (medium)
(1987) Age: 12 18 ys. 2. Symptom Checklist-90 (SCL- SCL-90 total
90) (Derogatis, Lipman, & Covi, d = .63
Design: Sessions: Gender: 1973) (Child)
Hostility d = .47
Single group 4 - 5 hours x 1/week 100% girls d = 1.32
Repeated Depression
Psychoticism d = .79
measures Duration: Inclusion criteria: Anxiety d = .36
8 weeks Hx of sexual Interpersonal Sensitivity d = .58
abuse Paranoid Ideation d = .25
Obsessive-compulsive d = .25
Treatment d = .43
Setting: Ethnicity: Phobic Anxiety
Somatization d = .50
University 60% Caucasian d = .42
campus 40% African 3. The Texas Social Behavioral
drama studio, American Inventory-Short Form (TSBI) TSBI d = .44
Montreal, (Helmreivh & Stapp, 1974)(Child)
Canada SES: low income
4. The Attributional Style
Questionnaire (ASQ) (Petersen ASQ d = .78
et al,, 1982) (Child)

5. The Social Support SSQ d = .02

Questionnaire (SSQ) (Sarason et
al.,1983) (Child)

6. The Marlowe-Crowne Social MCSDS

(Contact time: 36 hours) Desirability Scale (Crowne & d = .27
Marlowe, 1960) (Child)

98
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

McGain & Group therapy N = 30 1. Quay Revised Behavioral RBPC total d = 1.9
McKinzey (n = 15) Problem Checklist (RBPC) d = 1.8 (very large)
(1995) Age: 9 - 12 ys (Quay & Peterson, 1987)
OR (M = 10.5) (Child)
Design: Conduct Disorder d = 1.4
Matched No treatment Gender:
control (n = 15) 44% girls Socialized Aggression d = 1.5
group 66% boys
Attention problems/ d = 3.3
Sessions: Inclusion criteria: Immaturity Note:
Treatment 1/week Hx of physical The
Setting: abuse and/or Anxiety Withdrawal d = 1.9 treatment
In-patient Duration: neglect within past group
psychiatric 6 9 months year Psychotic Behavior d = .66 made
hospital statistically
Ethnicity: Motor Excess d = 2.03 significantly
n/a greater
2. Eyberg Child Behavior ECBI total gains on
SES: n/a Inventory(ECBI) d = 2.15 outcome
(Eyberg, 1980) measures
(Parent) compared
to the
Intensity Behavior d = 3.0 control
(Contact time: ~30 group
hours) Problem Behavior d = 1.3

99
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Nasrin, Cognitive-behavioral N = 14 1. Child Report of Post- EMDR EMDR

Greenwald, individual therapy traumatic Symptoms d = 1.9 d = 1.07
Rubin, Zand (Manualized) Age: 12-13 ys (CROP) (Fletcher, 1993) CBT (large)
& (n = 7) (Child) d = .92
Dolatabadi Gender:
(2004) OR 100% Girls CBT
2. Parent Report of Post- EMDR d = .71
Design: Eye Movement Inclusion criteria: Traumatic Stress d = 1.6 (medium)
Randomized Desensitization and Hx of sexual Symptoms (CROPS) CBT
comparison Reprocessing therapy abuse (Fletcher, 1993) d = 1.1
groups (Manualized) (Parent) Study
(n = 7) Ethnicity: total
Iranian EMDR d = .89
Treatment Sessions: 3. Rutter Teacher Scale d = .78 (large)
Setting: Twelve 45 minute SES: n/a (Rutter, 1967) CBT
Urban school sessions (Teacher) d = .85
district, Iran Note: CBT
Duration: and EMDR
Approximately 12 wks EMDR groups
4. Subjective Units of d=0 were not
Distress Scale (SUDS) CBT statistically
(Shaprio, 1995) d=0 different on
(Contact time: 9 hours) (Child) outcome
measures

100
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Nolan, Carr, 1. Individual therapy N = 38 1. Child Depression CDI Total (I) Total CDI
Fitzpatrick, (n = 20) Inventory (CDI) d = .28 d = .41
OFlaherty, Age: 6 - 17 ys. (Kovacs, 1983) CDI Total (C)
Keary, OR (M = 12.6) (Child) d = .55
Turner, Total
OShea, 2. Combined individual Gender: Negative Mood d = .20 (I) TSCC
Smythe & and group therapy (non- 93% Girls d = .29 (C) d = .39
Tobin (2002) manualized) 7% Boys Interpersonal Problems d = .33 (I)
d = .43 (C)
(n = 18)
Ineffectiveness d = .07 (I)
Design: Inclusion criteria: d = .69 (C)
Total
Non- Sessions: Hx of sexual Anhedonia d = .50 (I) CBCL
randomized Individual m = 18 abuse d = .46 (C) d = .76
comparison Combined m = 20 Negative Self-esteem d = .14 (I)
groups Ethnicity: d = .23 (C)
Duration: n/a Total YSR
6 months 2. Trauma Symptom TSCC Total (I) d = 0.0
Treatment SES: Checklist for Children d = .49
Setting: Predominantly (TSCC) (Briere, 1995) TSCC Total (C)
8 separate middle income d = .29 Study
Anxiety d = .52 (I)
mental health (Contact time: Total:
d = .56 (C)
agencies, Individual 18 hours d = .39
Depression d = .91 (I)
Ireland Combined 20 hours) d = .26 (C) (small)
Anger d = .23 (I)
Note: d = .10 (C)
The Individual and Post-traumatic Stress d = .66 (I)
Combined groups were d = .22 (C)
not statistically different Total Dissociation d = .27 (I)
on outcome measures d = .30 (C) (table
Total Sexual Problems d = .34 (I) continues)
d = .31 (C)

101
Nolan, Carr,
Fitzpatrick, 3. Child Behavior Checklist CBCL Total (I)
OFlaherty, (CBCL) (Achenbach & d = .72
Keary, Edelbrock, 1983) CBCL Total (C)
(Parent) d = .81
Turner,
OShea,
Smythe & Internalizing T-score d = 1.54 (I)
Tobin (2002) d = 1.47 (C)
(continued). Externalizing T-score d = .60 (I)
d = 1.25 (C)
Withdrawn d = .67 (I)
d = 2.5 (C)
Somatic Complaints d = .49 (I)
d = .65 (C)
Anxious/Depressed d = .66 (I)
d = 5.7 (C)
Social problems d = .39 (I)
d = .62 (C)
Thought Problems d = .75 (I)
d = .17 (C)
Attention Problems d = .72 (I)
d = 2.7 (C)
Delinquent Problems d = .25 (I)
d = 2.5 (C)
Aggressive Behaviors d = .39 (I)
d = .89 (C)

4. Youth Self Report Form d = .00 (C)

(YSR) (Achenbach & d = .00 (I)
Edelbrock, 1983)

102
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Effect
Size(s) Size(s)

Perez (1987) 1. Individual play N = 55 1. Primary Self Concept d = .64 Study

therapy (n = 18) Inventory (PSCI) Total:
Design: Age: 4 - 9 ys. (Muller & Leonetti, 1974) d = 1.09
Randomized OR (Child) (large)
comparison Gender:
and control 2. Group play therapy 63% Girls
group (n = 21) 37% Boys 2. Locus of Control Scale d = 1.55 Note:
(LCS) (Johnson, 1976) Group and
OR Inclusion criteria: (Child) Individual
Treatment Hx of sexual play
Setting: 3. No treatment (n = 16) abuse confirmed therapy
Community by police or social groups
mental health service agency in were not
agency Sessions: past 6 months statistically
60 minutes x 1/week different
Ethnicity: from each
Duration: n/a other, but
12 weeks were
SES: n/a statistically
stronger on
outcome
measures
than the
control
(Contact time: 12 hours) group

103
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Reams & Individual play therapy N = 36 1. Peabody Picture PPVT-R Study

Friedrich (structured) Vocabulary Test-Revised d = .35 Total:
(1994) Age: 3.5 - 5 ys. (Dunn & Dunn, 1981) d = .14
AND (Clinician) (small)
Design: Gender:
Randomized Mileu therapy 17% Girls 2. Draw a Person (DAP) DAP d = 0.0
comparison (n = 16) 83% Boys (McCarthy, 1972)
groups (Clinician)
OR
Inclusion criteria: 3. Eyberg Child Behavior ECBI total Note: No
Treatment Milieu therapy only Hx of abuse Inventory (ECBI) d = .10 statistically
Setting: (n = 15) and/or neglect (Robinson, Eyberg, & Ross, significant
Therapeutic 1980) (Parent) differences
nursery for on
maltreated Ethnicity: Problems d = .15 outcome
children Sessions: 28% Caucasian measures
50 minutes x 1/week 42% African Intensity d = .05 were
American observed
Duration: 30% Other/Mixed 4. Preschool Behavior PBQ d = 0.11 between
15 weeks Questionnaire (PBQ) the play
SES: n/a (Behar & Stringfield, 1974) therapy
and the
(Contact time: 12.5 Aggression d = 0.13 comparison
hours) group
Hyperactivity d = 0.09

104
Study Intervention(s) Participants Outcome Measure(s) Measure Mean Effect
(Response Perspective) Effect Size(s)
Size(s)
Rust & Group therapy N = 50 1. Stanford Achievement SAT total d = .89
Troupe Test (SAT) Form F, d = .73 (large)
(1991) AND Age: 9 18 ys. Intermediate level
(M = 12.5) (Gardner et al., 1982)
Individual therapy
Design: (Clinician)
(n = 25)
Matched Gender:
control 100% girls Math d = .68
AND
group
Separate perpetrator Reading d = .56
Note: and parent therapy Inclusion criteria: Note:
Comparison groups Hx of sexual The
group had abuse 2. Piers-Harris Self- PHSCS treatment
not OR Concept Scale (Piers & d = 1.06 group had
experienced Ethnicity: Harris, 1984) statistically
No treatment
abuse 96% Caucasian (Child) significantly
(n = 25)
4% African higher
American scores on
Treatment outcome
Sessions:
Setting: SES: measures as
1/week
In-patient n/a compared to
psychiatric the
Duration:
hospital comparison
6 months
group
(Contact time: > 24
hours)

105
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Stauffer Parallel cognitive- N = 19 children 1. Child Behavior Checklist d = .30 Study

& behavioral group and 19 caregivers (CBCL) (Achenbach & Total:
Deblinger therapy for children and Edelbrock, 1983) d = .52
(1996) their non-offending Childrens ages: (Parent) (medium)
caretakers 2 - 6 ys.
Design: (M = 4.21)
Single Sessions: 2. The Child Sexual d = .73
group 2 hours/week Gender: Behavior Inventory (CSBI)
Repeated (for both parent and 74% Girls (Friedrich et al., 1991)
measures child groups) 26% Boys (Parent)

Duration:
Treatment 11 weeks Inclusion criteria:
Setting: Hx of abuse
University and/or neglect
teaching
hospital
outpatient Ethnicity:
clinic 84% Caucasian
16% African
American

(Contact time: 44 hours) SES: n/a

106
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Stubenbort Theraplay model N = 39 1. Battelle Developmental d = .73 d = .95

(2000) (developmentally Inventory (Newborg, Stock, (large)
appropriate therapeutic Age: 2.8 - 5.6 ys. Wnek, Guidubaldi &
Design: and educational (M = 3.8) Svinicki, 1984)
Quasi- activities) (Teacher)
experimental Gender:
Repeated Sessions: 48% girls
measures 3 hr x 5 days/wk 52% boys
(1 treatment, 2. American Guidance d = 1.17
2 groups) Duration: Inclusion criteria: Service Early Screening
12 wks Hx of Profile (ESP) (Harrison,
maltreatment 1990)
(Teacher) Note:
Treatment Ethnicity: Both
Setting: 60% African treatment
Community American groups
mental health 34% Caucasian were
center 6% Bi-racial combined
(Therapeutic for the
Parent and SES: n/a present
Childrens (Contact time: > 180 analysis
Center) hours)

107
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Sullivan, 1. Individual N = 72 Child Behavior Checklist Study

Scanlan, psychotherapy (CBCL) (Achenbach & Total:
Brookhouser (in sign language) Age: 5.2 -17.4 ys. Edelbrock, 1983) d = .89
& Schulte (n = 35) (M = 11.5) (Houseparent) (large)
(1992)
OR Gender: Internal d = 1.54
Design: 29% girls
External d = .94
Randomized 2. No treatment 71% boys
control group (n = 37) Anxiety d = .47
Inclusion criteria:
Depressed d = .50
Sexual abuse Hx
Treatment Sessions: by staff or older Somatic d = 1.0
Setting: 2 x 2 hr/week students while at
Schizoid d = 1.0
Residential residential school
school for the Duration: 36 weeks Uncommunicative d = .75
deaf
Immature d = .94
Ethnicity: n/a Obsessive d = .72
Hostile d = 1.6
SES: n/a
Delinquent d = .75
Aggressive d = .93
Cruel d = .65
(Contact time: 144
Hyperactive d = .72
hours)

108
Study Intervention(s) Participants Outcome Measure(s) Measure Mean Effect
(Response Perspective) Effect Size(s) Size(s)

Thun, Sims, Group therapy N = 11 The Offer Self-Image Total study

Adams & (structured) Questionnaire, Revised d = .23
Webb (2002) (n = 4) Age: 16 18 ys. (OSIQ-R) ( offer, Ostrov, (small)
Howard, & Dolan, 1992)
Design: OR Gender: (Child)
Randomized 100% girls
control group No treatment Impulse Control d = .56
(n = 7) Inclusion criteria:
Hx of sexual Self Confidence d = .12
Treatment abuse, Note: There
Setting: Sessions: Hx of high school Self Reliance d = .02 were no
Military- 1.5 hr/week drop-out meaningful
based Body Image d = .17 statistical
residential Duration: difference
treatment 12 weeks Ethnicity: observed
program n/a between the
control and
SES: experimental
(Contact time: 18 hours) n/a groups

109
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Tucker Outdoor Adventure N = 103 1. Social Skills Rating Study

(2006) groups System (Gresham & Elliot, Total:
(n = ~8) Age: 9-15 ys. 1990) d = .08
(M = 11.71) (Child) (negligible)
Design:
Single group Sessions: Gender: Total d = .006
Repeated ~ 18 outdoor after- 33% Girls
measures school day trips 67% Boys Cooperation d = .03

Duration: Inclusion criteria: Assertion d = .15

Treatment 9 weeks Massachusettes
Setting: Dept. of Social Empathy d = .09
Dept. of Services referral
Social for maltreatment, Self-Control d = .06
Services Removal from
affiliated home
outdoor
adventure Ethnicity: 2. Norwicki-Strickland d = .16
program 85% Caucasian Locus of Control Scale for
3% African Children (N-SLCS)
American (Norwicki & Strickland,
12% Latina/o 1973)
Asian (n = 1) (Child)
(Contact time: ~ 80
hours) SES: n/a

110
Study Intervention(s) Participants Outcome Measure(s) Measure Mean Effect
(Response Perspective) Effect Size(s) Size(s)

Verleur, Group therapy N = 30 1. Coopersmith Self- CSI d = 1.54 d = 1.24

Hughes & WITH Esteem Inventory (CSI) (very large)
Dobkin sex education Age: 13 - 17 ys. (Coopersmith, 1981)
de Rios (n = 15) (Child)
(1986) Gender:
OR 100% girls
Design: 2. Anatomy/Physiology APSAS total
Matched No treatment Inclusion criteria: Sexual Awareness Scale d = .98
control (n = 15) Hx of intrafamilial (APSAS) (Verleur, 1986) Note:
group sexual abuse, (Child) Experimental
Sessions: CPS/court referral group scored
1/week Anatomy/physiology d = .82 statistically
Treatment Ethnicity: significantly
Setting: Duration: n/a Venereal Disease d = 1.04 higher than
24-hour 6 months control group
residential SES: n/a Birth Control d = .95 on outcome
treatment measures
center Anatomy Education d = 1.10
(Contact time: 24 hours)

111
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Waters 1. Child group therapy N = 29 (children) 1. Child Behavior Checklist CBCL

(1998) (CBCL) (Achenbach & total
AND Edelbrock, 1983) d = .34
Design: Age: 6 18 ys. (Parent)
Single group 2. Parent group therapy CRTIR
Repeated Internal d = .59 total
measures Gender: d = .31
29% girls External d = .12
Treatment Sessions: 71% boys
Setting: n/a 2. Child Report of
Community Treatment Issue Resolution Study
mental health Duration: Inclusion criteria: (CRTIR) (Nelson-Gardell, Total:
center ~ 20 weeks Sexual abuse Hx 1995) d = .33
(Child) (small)

Ethnicity: n/a Social Buffering d = .02

Self-Protection d = .24
SES: n/a
Self-Blame d = .70

Stigma/Shame/Fear d = .31

(Contact time:
~ 40 hours)

112
Study Intervention(s) Participants Outcome Measure(s) Measure Mean
(Response Perspective) Effect Size(s) Effect
Size(s)

Zion (1999) Individual client- N = 26 1. Abuse Behavior d = .17 Study

centered play therapy Checklist (ABC) Total:
Design: Age: 3 - 9 ys. (Chaffin & Wherry, 1993) d = .24
Single group Sessions: (Clinician) (small)
Repeated 1/week Gender:
measures 73% Girls 2. Joseph Preschool and
Duration: 27% Boys Primary Self-Concept
Treatment 12 weeks Screening Test (JPPSCS)
Setting: Inclusion criteria: (Joseph, 1979)
Catholic Hx of sexual (Clinician)
Charities abuse in past
community year, Global Scale d = .25
mental health Referral from
Division of Child Competence Scale d = .11
and Family
Services
3. Behavior Assessment d = .43
Ethnicity: System for Children
All Caucasian with Parents Rating Scale
1 Native American (BASC-PRS) Behavioral
Index Score
SES: n/a (Reynolds & Kamphaus,
1992)
(Parents)
(Contact time: 12 hours)

113
 REFERENCES

Achenbach, T. M. (1991). Integrative guide to the 1991 CBCL/4-18, YSR, and TRF

profiles. Burlington, VT: University of Vermont.

Adamec, R. E., & Stark-Adamec, C. (1989). Behavioral inhibition and anxiety:

Dispositional, developmental, and neural aspects of the anxious personality of

domestic cat. In J. S. Reznick (Ed.), Perspectives on behavioral inhibition.

Chicago: University of Chicago Press.

Allin, H., Wather, C.N. & MacMillan, H. (2005). Treatment of child neglect: A systematic

review. Canadian Journal of Psychiatry, 50(8), 497-504.

Ainsworth, M. D., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment.

Hillsdale, NJ: Erlbaum.

Ashby, M.R., Gilchrist, L.D., & Miramontez, A. (1987). Group treatment for sexually

abused American Indian adolescents. Social Work with Groups, 10(4), 21-32.

Atkinson, D.R., Furlong, M.J., & Wampold, B.E. (1982). Statistical significance, reviewer

evaluations, and the scientific process: Is there a statistically significant

relationship? Journal of Consulting Psychology, 29, 189-194.

Ayers, M.E. (1995). A controlled study of EEG neurofeedback and physical therapy with

pediatric stroke, age seven months to age fifteen, occurring prior to birth.

Biofeedback and Self-Regulation, 20(3), 318.

Baker, C.R. (1985). A comparison of individual and group therapy as treatment of

sexually abused adolescent females. Unpublished doctoral dissertation.

University of Maryland, College Park, MD.

114
Barach, P.M. (1991). Multiple personality disorder as an attachment disorder.

Dissociation, 4, 117-123.

Bateson, G., Jackson, D. D., Haley, J., & Weakland, J.H. (1968). Toward a theory of

schizophrenia. In D.D. Jackson (Ed.), Communication, family, and marriage:

Human communication. (Vol.1). Palo Alto, CA: Science and Behavior Books.

Becker, L.A. (2000). Effect size calculators. Retrieved January 12, 2007, from

http://web.uccs.edu/becker/psych590/essalc3.htm

Beitchman, J., Zucker, K., Hood, J., DaCosta, G., Ackman, D., & Cassavia, E. (1992). A

review of the long-term effects of child sexual abuse. Child Abuse and Neglect,

16, 101-118.

Berliner, L., & Saunders, B.E. (1995) Treating fear and anxiety in sexually abused

children: Results of a controlled 2-year follow up study. Child maltreatment, 1(4),

294-309.

Black, L. (2005). Effects of childhood sexual abuse on adult brain plasticity as measured

by quantitative electroencephalogram, neuropsychological, and psychological

tests. Unpublished doctoral dissertation, University of North Texas, Denton.

Black, L., Hudspeth, W., Townsend, A., & Bodenhamer-Davis, E. (2002, October).

Effects of childhood sexual abuse on adult brain plasticity as measured by

quantitative electroencephalogram. Presentation at the Society for Neuronal

Regulation annual conference, Scottsdale, AZ.

Bradshaw, G.A., Schore, A., Brown, J.L., Poole, J.H., & Moss, C.J. Elephant

breakdown. Nature, 433, 807.

115
Boney-McCoy, S., & Finkelhor, D. (1995). Psychosocial sequelae of violent victimization

in a national youth sample. Journal of Consulting and Clinical Psychology, 63,

726-736.

Borgman, R. (1984). Problems of sexually abused girls and their treatment. Social

Casework: Journal of Contemporary Social Work, 5, 182-186.

Bowen, M. (1976). Theory in the practice of psychotherapy. In P.J. Guerin, Jr. (Ed.),

Family therapy: Theory and practice. New York: Gardner.

Bowlby, J. (1969). Attachment and loss (Vol. 1). New York: Basic Books.

Bradley, S. J. (2000). Affect regulation and the development of psychopathology. New

York: Guilford Press.

Brown, J., Cohen, P., Johnson, J.G., & Smails, E.M. (1999). Childhood abuse and

neglect: Specificity of effects on adolescent and young adult depression and

suicidality. Journal of the American Academy of Child and Adolescent Psychiatry,

38, 1490-1496.

Carozza, P.M., & Heirsteiner, C.L. (1982). Young female incest victims in treatment:

Stages of growth seen in an art therapy model. Clinical Social Work Journal, 10,

165-175.

Carrion, V., Weens, C., Eliez, S., Patwardhan, A, Brown, W., Ray, R., et al. (2001).

Attenuation of forntal asymmetry in pediatric post-traumatic stress disorder.

Biological Psychiatry, 50(12), 943-951.

Casey, R.J. & Berman, J.S. (1985). The outcome of psychotherapy with children.

Psychological Bulletin, 98, 388-400.

116
Cassidy, J. (1994). Emotion regulation: Influences of attachment relationships. In N.A.

Fox (Ed.), The development of emotion regulation: Biological and behavioral

considerations. Monographs of the Society for Research in Child Development,

59(2-3), 228-248.

Chanitz, S. P. (1995). Interventions with children with developmental disabilities and

attachment disorders. Developmental and Behavioral Pediatrics, 16(3), 17-20.

Choi, D. W. (1992). Excitotoxic cell death. Journal of Neurobiology, 23, 1261-1276.

Chugani, H. T., Behen, M. E., Muzik, O., Juhasz, C., Nagy, F., & Chugani, D. C. (2001).

Local brain functional activity following early deprivation: A study of post-

institutionalized Romanian orphans, Neuroimage, 14, 12901301.

Cicchetti, D., & Tucker, D. (1994). Development and self-regulatory structures of the

mind. Development and Psychopathology, 6, 533-549.

Clendenon-Wallen, J. (1991). The use of music therapy to influence the self-confidence

and self-esteem of adolescents who are sexually abused. Music Therapy

Perspectives, 9, 73-81.

Clutton-Brock, T.H. (1991). The evolution of parental care. Princeton, NJ: Princeton

University Press.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (Revised ed.).

Hillsdale, NJ: Lawrence Earlbaum Associates, Inc.

Cohen, J.A. & Mannarino, A.P. (1996). A treatment outcome study for sexually abused

preschool children: Initial findings. Journal of the American Academy of Child and

Adolescent Psychiatry, 35(1), 43-50.

117
Cohen, J.A. & Mannarino, A.P. (1998). Interventions for sexually abused children: Initial

treatment findings. Child Maltreatment, 3, 17-26.

Cotton-Cornelius, D. (2003). Change processes and the behavior of foster children as a

function of an increased number of counseling sessions and the therapeutic

alliance: A 12-18 month longitudinal analysis. Unpublished doctoral dissertation.

The Chicago School of Professional Psychology, Chicago, IL.

Crittenden, P. M. (1995). Attachment and psychopathology. In S. Goldberg, R. Muir & J.

Kerr (Eds.), Attachment theory: Social, developmental and clinical perspectives.

Hillsdale, NJ: Analytic Press.

Culp, R.E., Little, V., Letts, D., & Lawrence, H. (1991). Maltreated childrens self-

concept: Effects of a comprehensive treatment program. American Journal of

Orthopsychiatry, 61(1), 115-121.

Damon, L. & Waterman, J. (1986). Parallel group treatment of children and their

mothers, In K. MacFarlane, J. Waterman, S. Conerly, L. Damon, M. Durgee, & S.

Long (Eds.), Sexual abuse of young children: Evaluation and treatment. NY:

Guilford Press.

Dario, M., Wallen, K., & Carroll, K. A. (1997). Infant abuse runs in families of group-

living pigtail macaques. Child Abuse and Neglect, 21(5), 465-471.

Davenport, R. K. & Rogers, C. J. (1970). Differential rearing of the chimpanzee: A

project survey. In G.H. Bourne (Ed.), The Chimpanzee. Basel: Karger.

De Bellis, M., Keshavan, M., Clark, D., Casey, B., Giedd, J., Boring, A., et al. (1999).

Developmental traumatology part II: Brain development. Biological Psychiatry,

45, 1271-1284.

118
Deblinger, E. & Heflin, A.H. (1996). Treating sexually abused children and their non-

offending parents: A cognitive-behavioral approach. Thousand Oaks, CA: Sage.

DeGangi, G. (2000). Pediatric disorders of regulation in affect and behavior: A

therapists guide to assessment and treatment. San Diego, CA: Academic Press.

De Luca, R.V., Boyes, D.A., Grayson, A.D. & Romano, E. (1995). Sexual abuse: Effects

of group therapy on pre-adolescent girls. Child Abuse Review, 4, 263-277.

De Luca, R.V., Hazen, A., & Cutler, J. (1993). Evaluation of a group counseling program

for preadolescent female victims of incest. Elementary School Guidance and

Counseling, 28, 104-114.

Denenberg, V. H. (1983). Lateralization of function in rats. American Journal of

Physiology, 245, 505-509.

De Shazer, S. & Molnar, A. (1984). Four useful interventions in brief family therapy.

Journal of Marital and Family Therapy, 10, 297-304.

Devilly, G.J. (2004). Assessment devices. Retrieved January 12, 2007, from

Swineburne University, clinical and forensic psychology website:

http://www.swin.edu.au/victims/resources/assessment/assessment.html

Dube, S.R., Anda, R.F., Felitti, V.J., Chapman, D.P., Williamson, D.F., & Giles, W.H.

(2001). Childhood abuse, household dysfunction, and the risk of attempted

suicide throughout the lifespan. Journal of the American Medical Association,

286, 3089-3096.

Dunlop, W. P., Cortina, J. M., Vaslow, J. B., & Burke, M. J. (1996). Meta-analysis of

experiments with matched groups or repeated measures designs. Psychological

Methods, 1, 170-177.

119
Egeland, B., & Sroufe, L. A. (1981). Attachment and early maltreatment. Child

Development, 52, 44-52.

English, D.J. (1998). The extent and consequences of child maltreatment. Future of

Children, 8(1), 39-53.

Field, T. (1994). The effects of mothers physical and emotional unavailability on

emotion regulation. In N.A. Fox (Ed.), The development of emotion regulation:

Biological and behavioral considerations. Monographs of the Society for

Research in Child Development, 59, 208-227.

Finkelhor, D. & Berliner, L. (1995). Research on the treatment of sexually abused

children: A review and recommendations. Journal of the American Academy of

Child and Adolescent Psychiatry, 34, 1408-1423.

Fisher, B. S, Turber, J., & Gunkelman, J. (2005, September). Healing the attachment

disorder before and after behavioral and neurological functioning. Workshop

presented at the annual meeting of the International Society for Neuronal

Regulation, Denver, CO.

Fisher, S. F. (2002). Neurofeedback: A treatment for reactive attachment disorder. EEG

Spectrum .[On-line]. Available: http://www. www.eegspectrum.com/

Friederich, W.N. (1995). Psychotherapy with sexually abused boys. Thousand Oaks,

CA: Sage.

Fride, E., Dan, Y., Feldon, J., Halevy, G., & Weinstock, M. (1986). Effects of prenatal

stress on vulnerability to stress in prepubertal and adult rats. Physiology and

Behavior, 37(5), 681-688.

120
Fuchs, T., Birbaumer, N., Lutenberger, W., Gruzelier, J. H., & Kaiser, J. (2003).

Neurofeedbacktreatment for attention-deficit/hyperactivity disorder in children: A

comparison with Methylphenidate. Applied Psychophysiology and Biofeedback,

28(1), 1-12.

Galaburda, A. M. (1984). Anatomical asymmetries in the human brain. In N. Geschwind

& A.M. Galaburda (Eds.), Biological foundations of cerebral dominance.

Cambridge, MA: Harvard University Press.

George, C., & Main, M. (1979). Social interactions of young abused children: Approach,

avoidance and aggression. Child Development, 50, 306-318.

Gil, E. (1998). Essentials of play therapy with abused children. New York: Guilford

Press.

Goldenberg, I. & Goldenberg, H. (1996). Family therapy: An overview. Pacific Grove,

CA: Brooks/Cole Publishing Company.

Goldberg, S., MacKay-Soroka, S., & Rochester, M. (1994). Affect, attachment and

maternal responsiveness. Infant Behavior and Development, 17, 335-339.

Gorman, J. M., Liebowitz, M. R., Fryer, A. J., & Stein, J. (1989). A neuroanatomical

hypothesis for panic disorder. American Journal of Psychiatry, 146, 148-161.

Grammer, H., & Shannon, J. (1992). Survivors group: Clinical interventions for the

sexually abused child in treatment foster care. Community Alternatives:

International Journal of Family Care, 4, 19-32.

Green A., Voeller, K., & Gaines, R. (1981). Neurological impairment in maltreated

children. Child Abuse and Neglect, 5, 121-134.

121
Greenberg, L. (1987). An objective measure of methylphenidate response: Clinical use

of the MCA. Psychopharmacology Bulletin, 23, 279-282.

Grotstein, J. S. (1986). The psychology of powerlessness: Disorders of self-regulation

and interactional regulation as a new paradigm for psychopathology.

Psychoanalytic Inquiry, 6, 93-118.

Hack, T.F., Osachuk, T.A.G., & De Luca, R.V. (1994, April). Group treatment for

sexually abused preadolescent boys. Families in Society: Journal of

Contemporary Human Services, 217-228

Hagans, K.B., & Case, J. (1988). When your child has been molested: A parents guide

to healing and recovery. Lexington, MA: Lexington Books.

Hall, S. E. K., & Geher, G. (2003). Behavioral and personality characteristics of children

with reactive attachment disorder. Journal of Psychology, 137(2), 145-162.

Hall-Marley, S.E., & Damon, L. (1992). Impact of structured group therapy on young

victims of sexual abuse. Journal of Child and Adolescent Group Therapy, 3(1),

41-48.

Harlow, H. F., & Harlow, M. K. (1965). The affectational systems. In A. M. Schrier, H. F.

Harlow, & F. Stollnitz (Eds.), Behavior of nonhuman primates, Volume II. New

York: Academic Press.

Hedges, L.V., & Olkin, I. (1985). Statistical methods for meta-analysis. San Diego, CA:

Academic Press.

Hirshberg, L. M., Chiu, S., & Frazier, J. A. (2005). Emerging brain-based interventions

for children and adolescents: Overview and clinical perspective. Child and

Adolescent Psychiatric Clinics of North America, 14, 1-19.

122
Hoag, M.J. & Burlingame, G.M. (1997). Evaluating the effectiveness of child and

adolescent group treatment: A meta-analytic review. Journal of Clinical Child

Psychology, 26(3), 234-246.

Huang-Storms, L., Bodenhamer-Davis, G., Davis, R., & Dunn, J. (2007). QEEG-guided

neurofeedback for children with histories of abuse and neglect:

Neurodevelopmental rationale and pilot study. Journal of Neurotherapy, 10(4),

3-16.

Humberson, M. (1998). A preliminary investigation of the impact of a parallel group

model of treatment on behavior problems in young victims of sexual abuse.

Unpublished doctoral dissertation. California School of Professional Psychology,

Los Angeles, CA.

Hunter, J. & Schmidt, F. (1990). Methods of meta-analysis: Correcting error and bias in

research findings. Newberry Park, CA: Sage Publication, Inc.

Ito, Y., Teicher, M., Glod, C., & Ackerman, E. (1998). Preliminary evidence for aberrant

cortical development in abused children: A quantitative EEG study. Journal of

Neuropsychiatry and Clinical Neurosciences, 10, 298-307.

Iwaniec, D. (1995). The emotionally abused and neglected child: Identification,

assessment, and intervention. West Sussex, England: John Wiley & Sons, Ltd.

Jacobson, R.S., & Straker, G. (1982). Peer group interaction of physically abused

children. Child Abuse and Neglect, 6, 321-327.

Jarusiewics, B. (2002). Efficacy of neurofeedback for children in the autistic spectrum.

Journal of Neurotherapy, 6(4), 39-49.

123
Jasper, H. (1958). The ten-twenty electrode system of the International Federation.

EEG and Clinical Neurophysiology, 10, 371-375.

John, E. R., Prichep, L. S., & Easton, P. (1987). Normative data banks and

neurometrics: Basic concepts, methods and results of norm constructions. In A.

S. Gevins & A Remond (Eds.), Handbook of electroencephalography and clinical

neurophysiology. (Vol.1). Amsterdam: Elsevier Publishing Co.

Kadzin, A.E., Bass, D., Ayers, W.A. & Rogers, A. (1990). Empirical and clinical focus of

child and adolescent psychotherapy research. Journal of Consulting and Clinical

Psychology, 58, 729-740.

Kaplan, S.J., Pelcovitz, D., & Labruna, V. (1999). Child and adolescent abuse and

neglect research: A review of the past 10 years. Part I: Physical and emotional

abuse and neglect. Journal of the American Academy of Child and Adolescent

Psychiatry, 38, 1214-1222.

Kaufman, J., & Zigler, E. (1989). The intergenerational transmission of child abuse. In

D. Cicchetti & V. Calson (Eds.), Child maltreatment: Theory and research on the

causes and consequences of child abuse and neglect. New York: Cambridge

University Press.

Kendall, P.C., Lerner, R.M., & Craighead, W.E. (1984). Human development and

intervention in child psychopathology. Child Development, 55, 71-82.

Kopp, C. B. (1989). Regulation of distress and negative emotions: A developmental

view. Developmental Psychology, 25, 343-354.

Kruczek, T., & Vitanza, S. (1999). Treatment effects with an adolescent abuse survivors

group. Child Abuse and Neglect, 5, 477-485.

124
Landreth, G. Play therapy: The art of the relationship. (1991). Accelerated

Development: Levittown, PA.

Lanktree, C.B., & Briere, J. (1995). Outcome of therapy for sexually abused children: a

repeated measures study. Child Abuse and Neglect, 19(9), 1145-1155.

LeDoux, J. (1996). The emotional brain. New York: Simon & Schuster.

Linden, M., Habib, T., & Radojevic, V. (1996). A controlled study of the effect of EEG

biofeedback on cognition and behavior of children with attention deficit disorder

and learning disabilities. Biofeedback and Self-Regulation, 21(1), 35-49.

Lipsey, M. & Wilson, D. (1993). The efficacy of psychological, educational, and

behavioral treatment: Confirmation from meta-analysis. American Psychologist,

48, 1181-1209.

Lubar, J. F. (1997). Neocortical dynamics: Implications for understanding the role of

neurofeedback and related techniques for the enhancement of attention. Applied

Psychophysiology and Biofeedback, 22(2), 111-126.

Lubar, J. F., Smartwood, M. O., Smartwood, J. N., & ODonnell, P. H. (1995). Evaluation

of the effectiveness of EEG neurofeedback training for ADHD in a clinical setting

as measured by changes in T.O.V.A. scores, behavioral ratings, and WISC-R

performance. Biofeedback and Self-Regulation, 20(1), 83-99.

Mackay, B., Gold, M., & Gold, E. (1987). A pilot study in drama therapy with adolescent

girls who have been sexually abused. Arts in Psychotherapy, 14, 77-84.

MacMillan, H.L. (2000). Child maltreatment: What we know in the year 2000. Canadian

Journal of Psychiatry, 45(8), 702-800.

125
Malamuth, N.M., Sockloskie, R.J., Koss, M.P., & Tanaka, J.S. (1991). Characteristics of

aggressors against women: Testing a model using a national sample of college

students. Journal of Consulting and Clinical Psychology, 59, 670-681.

Manassis, K., & Bradley, S. J. (1994). The development of childhood anxiety disorders:

Toward and integrated model. Journal of Applied Developmental Psychology, 15,

345-356.

McGain, B., & McKinzey, R.K. (1995). The efficacy of group treatment in sexually

abused girls. Child Abuse and Neglect, 19(9), 1157-1169.

Meaney, M.J., Aitken, D.H., van Berkel, C., Bhatnagar, S., & Sapolsky, R.M. (1988).

Effect of neonatal handling of age-related impairments associated with the

hippocampus. Science, 239, 766-768.

Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University

Press.

Moghaddam, B., Bolinao, M. L., Syein-Behrens, B., & Sapolsky, R. (1994).

Gluccocorticoids mediate the stress-induced extracellular accumulation of

glutamate. Brain Research, 655, 251-254.

Monastra, V. J., Monastra, D. M., & George, S. (2002). The effects of stimulant therapy,

EEG biofeedback, and parenting style on the primary symptoms of attention-

deficit /hyperactivity disorder. Applied Psychophysiology and Biofeedback, 27(4),

231-249.

Nasrin, J., Greenwald, R., Rubin, A., Zand, S.O., & Dolatabadi, S. (2004). A

Comparison of CBT and EMDR for sexually abused Iranian girls. Clinical

Psychology and Psychotherapy, 11, 358-368.

126
National Clearinghouse on Child Abuse and Neglect Information (October, 2005).

National child abuse and neglect data system (NCANDS): Summary of key

findings from calendar year 2000. DHHS Publication [On-line]. Available:

http://www.childwelfare.gov/

Nelki, J.S., & Watters, J. (1987). A group for sexually abused young children:

Unravelling the web. Child Abuse and Neglect, 13, 369-377.

Nolan, M., Carr, A., Fitzpatrick, C., OFlherty, A., Keary, K., Turner, R., OShea, D.,

Smyth, P., & Tobin, G. (2002). A comparison of two programmes for victims of

child sexual abuse: A treatment outcome study. Child Abuse Review, 11, 103-

123.

Othmer, S., Othmer, S. F., & Kaiser, D. A. (1999). EEG biofeedback: An emerging

model for its global efficacy. In J.R. Evans & A. Abarbanel (Eds.), Introduction to

quantitative EEG and neurofeedback. San Diego, CA: Academic Press.

Parloff, M.B. 91984). Psychotherapy research and its incredible credibility crisis. Clinical

Psychology Review, 4, 95-109.

Perez, C.L. (1987). A comparison of group play therapy and individual play therapy for

sexually abused children. Unpublished doctoral dissertation. University of

Northern Colorado.

Perry, B. D. (1994). Neurobiological sequelae of childhood trauma: PTSD in children. In

M. Murberg (Ed.), Catecholamines in post-traumatic stress disorder. Washington,

DC: American Psychiatric Press.

127
Perry, B. D., & Pate, J. (1994). Neurodevelopment and the psychobiological roots of

post-traumatic stress disorder. In L. Koziol & C. Stout (Eds.), The

neuropsychology of mental illness: A practical guide. Springfield, IL: Charles C.

Thomas.

Perry, B., & Pollard, R. (1998). Homeostasis, stress, trauma, and adaptation: A

neurodevelopmental view of childhood trauma. Child and Adolescent Psychiatric

Clinics of North America, 7 (3), 3-51.

Phillips, R.D. (1985). Whistling in the dark?: A review of play therapy research.

Psychotherapy, 33, 752-760.

Porter, F.S., Blick, L.C., & Sgroi, S.M. (1983). Treatment of the sexually abused child. In

S.M. Sgroi (Ed.), Handbook of clinical intervention in child sexual abuse.

Lexington, MA: Lexington Books.

Post, R. M., Rubinow, D. R., & Ballenger, J. C. (1984). Conditioning, sensitization and

kindling: Implications for the course of affective illness. In R. M. Post & J. C.

Ballenger (Eds.), Neurobiology of mood disorders. Baltimore, MD: Williams &

Wilkins.

Post, R. M., Weiss, R. B., & Leverich, G. S. (1994). Recurrent affective disorder: Roots

in developmental neurobiology and illness progression based on changes in

gene expression. Development and Psychopathology, 6, 781-813.

Punzo, F., & Ludwig, L. (2002). Contact with maternal parent and siblings affects

hunting behavior, learning, and central nervous system development in

spiderlings of Hogna Carolinensis Araeneae Lycosidae. Animal Cognition, 5,

63-70.

128
Read, J., Perry, B. D., Moskowitz, A., & Connolly, J. (2001). The contribution of early

traumatic events to schizpophrenia in some patients: A truamagenic

neurodevelopmental model. Psychiatry, 64(4), 319-345.

Rust, J.O., & Troupe, P.A. (1991). Relationship of treatment of child sexual abuse with

school achievement and self-concept. Journal of Early Adolescence, 11(4), 420-

429.

Reams, R. & Friedrich, W. (1994). The efficacy of time-limited play therapy with

maltreated pre-schoolers. Journal of Clinical Psychology, 50(6), 889-899.

Reeker, J., Ensing, D. & Elliott, R. (1997). A meta-analytic investigation of group

treatment outcomes for sexually abused children. Child Abuse and Neglect, 21

(7), 669-680.

Ronen, T. (1998). Linking developmental and emotional elements into child and family

cognitive-behavioral therapy. In P. Graham (Ed.), Cognitive-behavior therapy for

children and families. Cambridge, England: Cambridge University Press.

Rosenthal, R. & Rosnow, R.L. (1991) Essential of behavioral research: Methods and

data analysis (2nd ed.). New York: McGraw Hill.

Rossiter, T. (2004). The effectiveness of neurofeedback and stimulant drugs in treating

AD/HD: Part II. Replication. Applied Psychophysiology and Biofeedback, 29(4),

233-243.

Rossiter, T. P. & La Vaque, T. J. (1995). A comparison of EEG biofeedback and

psychostimulants in treating attention deficit/hyperactivity disorders. Journal of

Neurotherapy, 1(1), 48-59.

129
Rutter, M., & English and Romanian adoptees study team (1998). Developmental catch-

up, and deficit following adoption after severe global early privation, Journal of

Child Psychology and Psychiatry 39, 465-476.

Sapolsky, R. M. (1993). Potential behavioral modification of glucocorticoid damage to

the hippocampus. Behavioral Brain Research, 57, 175-182.

Satir, V. (1972). Peoplemaking. Palo Alto, CA: Science and Behavior Books.

Scharff, D.E., & Scharff, J.S. (1987). Object relations family therapy. Northvale, NJ:

Aronson.

Scher, C.D., Forde, D.R., McQuaid, J.R. & Stein, M.B. (2004). Prevalence and

demographic correlates of childhood maltreatment in an adult community

sample. Child Abuse and Neglect, 28, 167-180.

Schore, A. N. (1994). Affect regulation and the origin of self: The neurobiology of

emotional development. Hillsdale, NJ: Lawrence Erlbaum Associates.

Schore, A. N. (2001). The effects of relational trauma on right brain development, affect

regulation, and infant mental health. Infant Mental Health Journal, 22, 201-269.

Sedlack, A.J. & Broadhurst, D.D. (1996). Executive summary of the third National

Incidence Study of Child Abuse and Neglect. DHHS publication [On-line].

Available: http://nccanch.acf.hhs.gov/index.cfm

Sgoifo, A., Koolhaas, J., & De Boor, S. (1999). Social stress, autonomic neural

activation, and cardiac activity in rats. Neuroscience and Biobehavioral Reviews,

23, 915-923.

Shirk, S. (1988). Cognitive development and child psychotherapy. New York: Plenum

Press.

130
Sichel, A. G., Fehmi, L. G., & Goldstein, D. M. (1995). Positive outcome with

neurofeedback treatment of a case of mild autism. Journal of Neurotherapy, 1

(1), 60-64.

Seibert, C. (2006, October 8). An elephant crackup? New York Times. Retrieved

October 9, 2006, from

http://www.nytimes.com/2006/10/08/magazine/08elephant.html

Sharf, R.S. (2000). Theories of psychotherapy and counseling: Concepts and cases.

(2nd ed.). Pacific Grove, CA: Brooks/Cole.

Siniatchkin, M., Hierundar, A., Kropp, P., Kuhnert, R., Gerber, W. D., & Stephani, U.

(2000). Self-regulation of slow cortical potentials in children with migraine: An

exploratory study. Applied Psychophysiology and Biofeedback, 25(1), 13-32.

Smith, M.L., Glass, G.V., & Miller, T.L. (1980). The benefits of psychotherapy.

Baltimore: John Hopkins University Press.

Stauffer, L.B., & Deblinger, E. (1996). Cognitive behavioral groups for nonoffending

mothers and their young sexually abused children: A preliminary treatment

outcome study. Child Maltreatment, 1(1), 65-76.

Sterman, B. (1996). Physiological origins and functional correlates of EEG rhythmic

activities: Implications for self-regulation. Biofeedback and Self-Regulation, 21, 3-

33.

Stubenbort, K.F. (2003). The effectiveness of intervention for maltreated preschoolers:

An attachment theory perspective. Unpublished doctoral dissertation. University

of Pittsburgh.

131
Sullivan, P.M., Scanlan, J.M., Brookhouser, P.E., & Schulte, L.E. (1992). The effects of

psychotherapy on behavior problems of sexually abused deaf children. Child

Abuse and Neglect, 16, 297-307.

Tansey, M. A. (1992). EEG sensorimotor biofeedback training and the treatment of a

six-year old asthmatic child. American Journal of Clinical Biofeedback, 5(2), 145-

149.

Teicher, M.H., Ito, Y., Glod, C. A., Anderson, S. L., Dumont, N., & Ackerman, E. (1997).

Preliminary evidence for abnormal cortical development in physically and

sexually abused children using EEG coherence and MRI. Annals of the New

York Academy of Sciences, 821, 160-75.

ten Bensel, R.W., Rheinberger, M.M., & Radbill, S.X. (1997). Children in a world of

violence: The roots of child maltreatment. In M.E. Helfer, R.S. Kempe, & R.D.

Krugman (Eds.), The battered child (5th ed.). Chicago: University of Chicago

Press.

Terr, L.C. (1981). Forbidden games: Post traumatic stress childrens play. Journal of the

American Academy of Child Psychiatry, 20, 741-760.

Thalheimer, W., & Cook, S. (2002, August). How to calculate effect sizes from published

research articles: A simplified methodology. Retrieved January 9, 2007, from

http:/work-learning.com/effect_sizes.htm

Thatcher, R. W. (1998). Normative EEG databases and EEG biofeedback. Journal of

Neurotherapy, 2(4), 8-39.

132
Thompson, M., & Thompson, L. (2003). The neurofeedback book: An introduction to

basic concepts in applied psychophysiology. Wheat Ridge, CO: Association for

Applied Psychophysiology and Biofeedback.

Thun, D., Sims, P., Adams, M., & Webb, T. (2002). Journal of Child Sexual Abuse, 11

(4), 1-16.

Tucker, A.R. (2006). The Impact of therapeutic adventure groups on the social

functioning of early adolescents. Unpublished doctoral dissertation. Boston

College, Boston, MA.

Van Bloem, L. L. (2000). Quantitative encephalograms (QEEG) in children with reactive

attachment disorder. Journal of Neurotherapy, 4(4), 106-107.

van der Kolk, B., & Greenberg, M. S. (1987). The psychobiology of the trauma

response: Hyperarousal, constriction, and addiction to traumatic reexposure. In

van der Kolk (Ed.), Psychological trauma. Washington DC: American Psychiatric

Press.

Wagner, W.G. (2003). Counseling, psychology, and children: A multidimensional

approach to intervention. Columbus, OH: Merrill Prentice Hall.

Watanabe, Y., Gould, E., & McEwen, B.S. (1992). Stress induces atrophy of apical

dendrites of hippocampal CA3 pyramidal neurons. Brain Research, 588, 341-

345.

Waters, K.T. (1998). Program evaluation of group treatment of sexually abused

children. Unpublished doctoral dissertation. University of Cincinnati, OH.

133
Weisz, J, Donenberg, G.R., Han, S. & Weiss, B. (1995). Bridging the gap between

laboratory and clinic in child and adolescent psychotherapy. Journal of

Consulting and Clinical Psychology, 63(5), 688-701.

Weisz, J., Weiss, B., Han, S., Granger, D.A. & Morton, T. (1995). Effects of

psychotherapy with children and adolescents revisited: A meta-analysis of

treatment outcome studies. Psychological Bulletin, 117(3), 450-468.

Verleur, D., Hughes, R.E., & Dobkin de Rios, M. (1986). Enhancement of self-esteem

among female adolescent incest victims: A controlled comparison. Adolescence,

21(84), 843-854.

White, J., & Allers, C. (2001). Play therapy with abused children: A review of the

literature. Journal of Counseling and Development, 72, 390-394.

Widom, C.S. (1989). Posttraumatic stress disorder in abused and neglected children

grown up. American Journal of Psychiatry, 156, 1223-1229.

Wilens, T. E. (2004). Straight talk about psychiatric medications for kids. New York:

Guilford Press.

Wilson, S. L. (2001). Attachment disorders: Review and current status. Journal of

Psychology, 135(1), 37-51.

Zion, T.A. (1999). Effects of individual client-centered play therapy on sexually abused

childrens mood, self-concept, and social competence. Unpublished doctoral

dissertation. Brigham Young University, Provo, UT.

134