E. Fiona Bailey, Keith W. Fridel and Amber D.

Rice
J Neurophysiol 98:3284-3291, 2007. First published Oct 10, 2007; doi:10.1152/jn.00865.2007

You might find this additional information useful...

This article cites 20 articles, 13 of which you can access free at:
http://jn.physiology.org/cgi/content/full/98/6/3284#BIBL

This article has been cited by 2 other HighWire hosted articles:

Force-EMG Changes During Sustained Contractions of a Human Upper Airway Muscle
K. Schmitt, C. DelloRusso and R. F. Fregosi
J Neurophysiol, February 1, 2009; 101 (2): 558-568.
[Abstract] [Full Text] [PDF]

Genioglossus and Intrinsic Electromyographic Activities in Impeded and Unimpeded

Protrusion Tasks
L. J. Pittman and E. F. Bailey
J Neurophysiol, January 1, 2009; 101 (1): 276-282.
[Abstract] [Full Text] [PDF]

Updated information and services including high-resolution figures, can be found at:

Downloaded from jn.physiology.org on November 10, 2009

http://jn.physiology.org/cgi/content/full/98/6/3284

Additional material and information about Journal of Neurophysiology can be found at:
http://www.the-aps.org/publications/jn

This information is current as of November 10, 2009 .

Journal of Neurophysiology publishes original articles on the function of the nervous system. It is published 12 times a year
(monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2005 by the
American Physiological Society. ISSN: 0022-3077, ESSN: 1522-1598. Visit our website at http://www.the-aps.org/.
J Neurophysiol 98: 3284 –3291, 2007.
First published October 10, 2007; doi:10.1152/jn.00865.2007.
Sleep/Wake Firing Patterns of Human Genioglossus Motor Units
E. Fiona Bailey,1 Keith W. Fridel,2 and Amber D. Rice1
1
Department of Physiology, College of Medicine, and 2Department of Psychology, College of Science, The University of Arizona,
Tucson, Arizona
Submitted 3 August 2007; accepted in final form 8 October 2007
Bailey EF, Fridel KW, Rice AD. Sleep/wake firing patterns of
human genioglossus motor units. J Neurophysiol 98: 3284 –3291,
2007. First published October 10, 2007; doi:10.1152/jn.00865.2007.
Although studies of the principal tongue protrudor muscle genioglos-
sus (GG) suggest that whole muscle GG electromyographic (EMG)
activities are preserved in nonrapid eye movement (NREM) sleep, it
is unclear what influence sleep exerts on individual GG motor unit
(MU) activities. We characterized the firing patterns of human GG
MUs in wakefulness and NREM sleep with the aim of determining 1)
whether the range of MU discharge patterns evident in wakefulness is
preserved in sleep and 2) what effect the removal of the “wakeful-
ness” input has on the magnitude of the respiratory modulation of MU
activities. Microelectrodes inserted into the extrinsic tongue protrudor
muscle, the genioglossus, were used to follow the discharge of single
MUs. We categorized MU activities on the basis of the temporal
relationship between the spike train and the respiration cycle and
quantified the magnitude of the respiratory modulation of each MU
using the eta (␩2) index, in wakefulness and sleep. The majority of
MUs exhibited subtle increases or decreases in respiratory modulation
but were otherwise unaffected by NREM sleep. In contrast, 30% of
MUs exhibited marked sleep-associated changes in discharge fre-
quency and respiratory modulation. We suggest that GG MUs should
not be considered exclusively tonic or phasic; rather, the discharge
pattern appears to be a flexible feature of GG activities in healthy
young adults. Whether such flexibility is important in the response to
changes in the chemical and/or mechanical environment and whether
it is preserved as a function of aging or in individuals with obstructive
sleep apnea are critical questions for future research.
INTRODUCTION
Research in the area of obstructive sleep apnea (OSA) has
long focused on the upper airway and in particular on the
consequences of the loss of the “wakefulness” input on the
whole muscle electromyographic (EMG) activities of the pha-
ryngeal dilator muscle, genioglossus (GG). Such studies dem-
onstrate that in wakefulness the GG muscle has a clear inspira-
tory phasic (i.e., greater activity in inspiration) pattern of
activity superimposed on a background of tonic (i.e., present
throughout the respiratory cycle) activity. Importantly, sleep
does not appear to differentially affect either component of
whole muscle GG EMG, consistent with the view that both
phasic and tonic GG activities contribute to preserved airway
patency (Tangel et al. 1992).
That the motoneuronal activities constituting whole muscle
GG EMG are more complex than suggested by the phasic/tonic
dichotomy is evident from the results of a recent study of
hypoglossal motor unit (MU) activities in wakefulness (Sa-
boisky et al. 2006). Beyond the traditional categories of phasic
Address for reprint requests and other correspondence: E. F. Bailey, De-
partment of Physiology, College of Medicine, The University of Arizona,
Tucson, AZ 85721-0093 (E-mail: ebailey@u.arizona.edu).
3284 0022-3077/07 $8.00 Copyright © 2007 The American Physiological Society
and tonic activities the authors characterized six principal MU
types based on firing pattern and the relation to the respiratory
cycle. Many of these MUs (⬃44%) exhibited inspiratory pha-
sic activities, but slightly more than half the MUs recorded
were classified as “tonic.” Interestingly, the majority of these
tonic MUs also exhibited either inspiratory- and, in rare cases,
expiratory-related increases in firing frequency, indicative of
respiratory modulation.
To date, no comparable study has characterized single GG
Downloaded from jn.physiology.org on November 10, 2009
MU activities both in wakefulness and in sleep. Thus it is
unclear whether the range of discharge patterns evident in
wakefulness is preserved in sleep or whether certain “types” of
MU are more or less susceptible to the removal of the wake-
fulness input. Accordingly, we sought to characterize GG MU
activities separately for wakefulness and NREM sleep based on
the relationship between the firing pattern and the respiratory
cycle (Netick and Orem 1981; Orem and Dick 1983). In view
of evidence of sleep-associated increases in whole muscle GG
EMG activities in healthy human subjects two to three breaths
after the wake–sleep transition (Tangel et al. 1992; Worsnop et
al. 1998, 2000), we anticipated that inspiratory phasic and
background tonic motor unit activities would be robustly pre-
served in NREM sleep.
Second, in view of the recent work of Saboisky et al. (2006)
that documents respiratory modulation of tonic GG MU activ-
ities in wakefulness, we sought to quantify the magnitude of
the modulation and to ascertain the effect of NREM sleep on
this aspect of MU activities. Earlier work by Orem and col-
leagues assessed the correlation between the activities of cen-
tral respiratory neurons and the ventilatory cycle and quantified
the strength of that association using the eta (␩2) index (Orem
and Dick 1983). In this study, we adapted this technique, using
it to assess the relative strength of the respiratory modulation of
single GG MU activities in wakefulness and subsequently in
NREM sleep. This enabled us to test a second hypothesis that
the strength of respiratory modulation is a stable characteristic
of single GG MU activities.
METHODS
We performed 25 experiments in 17 healthy human volunteers (10
women and 7 men, ages 22–55 yr (averages: age: 30 ⫾ 8.8 yr; height:
168 ⫾ 4.1 cm; weight: 67 ⫾ 12.1 kg; body mass index: 23.3 ⫾ 2.4).
All subjects were in good health, free of skeletal abnormality, without
history of major surgery or injury involving the respiratory system,
without neural or respiratory disease or disorder, without history of
sleep disorders, and free of any medication that could affect nervous
system function. The Human Subjects Committee at The University of
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
www.jn.org
 STATE-DEPENDENT MODULATION OF LINGUAL MOTOR UNIT ACTIVITIES
Arizona approved all experimental procedures. Subjects gave their
informed consent before participation in the study.
Electromyographic (EMG) recordings
Single motor unit action potentials (SMUAPs) recorded from GG
were obtained using tungsten microelectrodes (1- to 5-␮m tip diam-
eter, 250-␮m shaft diameter, 10 M⍀ at 1 kHz; FHC, Bowdoinham,
ME). A surface electrode (4-mm diameter Ag–AgCl) attached to the
skin overlying the mastoid process served as an indifferent electrode,
and both were referenced to a ground strap placed around the upper
arm. Motor unit potentials were amplified (⫻1,000), band-pass fil-
tered (0.3–3 kHz; Grass Instruments, West Warwick, RI), and dis-
played on a storage oscilloscope to monitor the size and shape of the
MU impulses and recorded on the Spike2 data acquisition and anal-
ysis program [Cambridge Electronic Design (CED), Cambridge, UK].
Whole muscle GG EMG activities were obtained in the same manner
but for these recordings the terminal 10 mm of the microelectrode tip
was bared of insulation. SMUAPs were recorded both ipsilateral and
contralateral to recordings of EMG from the whole muscle.
To optimize electrode placement, the musculature of the mouth
floor was initially visualized by ultrasonography (Pro Sound 3500;
Aloka, Tokyo, Japan) and the distance from the submental surface to
the inferior border of the GG muscle was determined using an
electronic caliper (Eastwood et al. 2003). To preserve motor unit
recordings in quiet wakefulness and during sleep necessitated a
specialized arrangement of the experimental setup as follows. First,
although tungsten microelectrodes render recordings of remarkable
fidelity and permit in situ manipulation of the electrode, their use
places considerable constraints on subject position and movement.
Thus all the experiments in this study were conducted with subjects in
the supine position in a dental chair. Second, to preserve electrode
position across wakefulness and sleep, we inserted electrodes more
posteriorly and to depths ⬎20 mm. These coordinates typically
rendered the most productive and stable MU recording sites relative to
more superficial and/or anterior locations (see Fig. 1).
Respiratory movements of the chest were monitored by a respira-
tory effort transducer (Biopac Systems, Goleta, CA) positioned at the
FIG. 1. Schematic lateral view of the mandible and genioglossus (GG)
muscle showing posterior (B) (adopted in the current protocol) and more
anterior (A) insertion sites, approximate angle of electrode, and the correspond-
ing recording regions (shown in white) within the GG muscle.
J Neurophysiol • VOL
3285
midsternal level and at the level of the umbilicus. Output from these
sensors reflected changes in the anteroposterior dimensions of the rib
cage and abdomen. These signals were used to estimate thoracic
expansion and to define the inspiratory and expiratory portions of the
respiratory cycle.
Polysomnographic recordings
Electrodes were placed according to the international 10 –20 system
using a referential montage with central (C4/A1) and occipital (O1/A2)
electroencephalogram (EEG) derivations, and right and left outer
canthus to obtain electrooculogram (EOG) recordings. Signals were
amplified (sensitivity 7.0 ␮V mm⫺1), band-pass filtered (0.3–100 Hz;
Grass Instruments), and relayed to the Spike2 data acquisition unit.
Protocol
Subjects were instructed to refrain from consuming any caffeinated
products in the 12 h leading up to the study. To facilitate subjects’
sleeping during the experiment, subjects were asked to retire about 2 h
Downloaded from jn.physiology.org on November 10, 2009
later than the usual bed time the night before. All experiments were
conducted in the early afternoon, commencing around 1:30 pm.
Subjects lay supine, head supported, in a dental chair. Each trial
consisted of periods of quiet wakefulness and NREM sleep. Subjects
were provided with a manual clicker and instructed to rest quietly with
their eyes closed and to depress the clicker until instructed otherwise.
This enabled subjects to maintain quiet wakefulness and obviated the
need for verbal communication that might otherwise displace the
needle electrode. After 2–3 min of stable wakefulness, the “clicker”
was removed from the subject’s hand and the subject was instructed
to rest quietly with eyes closed and the subject was allowed to fall
asleep. After recording in quiet wakefulness and sleep, the subject was
awakened and another MU was sought. If a MU with significantly
different firing pattern and shape (as seen on the oscilloscope) was
found it was considered to be a distinct MU and the protocol was
repeated. Lights remained off throughout the recording period and
subjects were exposed to a broadband noise (⬃85 dbA) that masked
their own breath sounds and external noise sources. Subjects were not
required to perform any maneuvers that might activate GG motor
units nor were they provided with visual or auditory feedback of
respiratory-related movements or motor unit activities.
Data analysis
All data were acquired using Spike2 software (CED). Subsequent
off-line analysis of EMG activities was performed using customized
computer software (Spike2).
SLEEP SCORING. Periods of sustained wakefulness and uninterrupted
sleep were identified using the following criteria. For each subject,
EEG activity initially was assessed as being predominantly alpha
(8 –12 Hz) or theta (4 –7.75 Hz) wave activity on the basis of
sequential 30-s epochs using the criteria of Rechtschaffen and Kales
(1968). Sleep onset was defined as two consecutive epochs of Stage 1
or one epoch of Stage 2, free from overt EEG arousals (⬎3.0 s)
(ASDA 1992; Loredo et al. 1999). Awakening was defined as an
instance of ⬎1 min of wakefulness occurring after sleep onset. In
addition, power spectral analyses were performed on the C4/A1
derivation on six consecutive 5-s epochs using commercial software
(Spike2, CED). In this manner, multiple nonoverlapping windows
were averaged to yield 30-s epoch (0.5-Hz resolution). Power spectral
distributions were computed for the following bandwidths: slow-wave
activity (0.75– 4.5 Hz), theta (4.0 –7.75 Hz), alpha (8.0 –12.0 Hz),
sigma (12.25–15.0 Hz), and beta (15.25–31.0 Hz) for 10 –15 consec-
utive breaths in wakefulness and 10 –15 consecutive breaths in NREM
sleep.
98 • DECEMBER 2007 • www.jn.org
3286 E. F. BAILEY, K. W. FRIDEL, AND A. D. RICE

MOTOR UNIT DISCRIMINATION. Motor units were discriminated us-
ing a template-matching algorithm based on waveform shape and
amplitude. Subsequently, each waveform was checked by visual
inspection against the template unit waveform. Only those units that
were active in quiet wakefulness and persisted in NREM sleep or were
reactivated on awakening from sleep were included in the analysis.
Initially, we characterized motor unit discharge pattern on the basis
of the temporal relationship between the spike train and the respiration
cycle based on the summed ribcage and abdomen volume excursions.
Briefly, we used a zero and peak crossing algorithm to divide each
breath into an inspiratory portion and expiratory portion and then each
of these partitions into two equal parts corresponding to early and late
inspiration (P1 and P2) and early and late expiration (P3 and P4).
Motor units were considered tonic if their activities persisted through-
out each of the four phases of the respiratory cycle for 10 –15
consecutive breath cycles. Conversely, motor units were considered
phasic if discharge ceased in one or more of the phases of the
respiratory cycle for 10 –15 consecutive breaths. The categorization of
MUs on the basis of discharge pattern was conducted separately for
wakefulness and NREM sleep.
We assessed the strength of the relationship between MU activities
typically recorded at depths 22–28 mm from the skin surface at
electrode angles of about 45° to the horizontal (see Fig. 1,
location B).
We recorded the activities of 81 GG motor units. The
number of motor units studied per subject ranged from 3 to 7.
Our analysis revealed consistently low ISI variability (⬃6 –
11.0%) but considerable heterogeneity in average discharge
frequency (range 8.0 –28.3 Hz). We were successful in record-
ing 64 single GG MUs in quiet wakefulness and in NREM
sleep. An additional 17 motor units that were inactivated at
sleep onset, but recruited on awakening, were also recorded.
Many other motor units were detected in quiet wakefulness but
fell silent in NREM sleep and were not recruited on awaken-
ing. The latter MUs are not included in the present analysis. In
all cases in NREM sleep, the peak power fell in the ␪ (4.0 –7.75
Hz) or slow-wave (0.75– 4.0 Hz) bandwidths.
Figure 2 depicts frequency histograms of all GG MUs as a
function of the eta (␩2) value, in wakefulness and NREM sleep.

Downloaded from jn.physiology.org on November 10, 2009

and the respiratory cycle by one-way ANOVA (Netick and Orem In wakefulness, the distribution is bimodal with the majority
1981) and quantified the magnitude of the association between a (n ⫽ 55) of MUs exhibiting moderate to weak respiratory-
motor unit’s activities and the respiratory cycle using the eta (␩2) related modulation (␩2 ⬍0.60). The remaining MUs exhibited
value (Netick and Orem 1981; Orem and Dick 1983; Orem et al. strong respiratory modulation (␩2 ⬎0.6). Whereas the total
2002). In this study, the ␩2 value indicates the proportion of the total number of active MUs declined (n ⫽ 64) in NREM sleep, there
variance (SStotal) of a MU’s activities over a series of breaths that is
was no significant difference in the magnitude of respiratory
made up by variance between phases of the respiratory cycle and is
calculated as follows: ␩2 ⫽ SSbetween/(SSbetween ⫹ SSwithin), where modulation of the population of MUs in wakefulness versus
SSbetween represents the variability between groups (i.e., the range in
impulse/count means across the four phases of a respiratory cycle) and
SSwithin represents the variance within groups (i.e., the variability in
impulse/count across 10 –15 breaths within individual phases of the
respiratory cycle). Using this approach we calculated ␩2 for individual
motor units based on 10 –15 breaths in quiet wakefulness and 10 –15
in uninterrupted NREM sleep. We subsequently averaged eta values
obtained from individual MUs to derive an eta value for the group in
wakefulness and sleep. Paired t-tests were used to determine statistical
significance of the change in ␩2 values separately between wakeful-
ness and sleep within subjects and to assess the overall effect of state
on the magnitude of respiratory modulation of the population.
Theoretically, ␩2 values can range from 0.00 to 1.0, indicating that
none or all of the MU activity is related to respiration. Throughout the
text, we refer to the magnitude of the association between motor unit
activities and respiration as the strength of respiratory-related modu-
lation with the following ranges considered: strong (␩2 ⬎0.60),
moderate (␩2 ⫽ 0.5– 0.3), and weak (␩2 ⬍0.3). Based on this
categorization scheme we completed a Pearson chi-square test of
association to assess the effect of state on MU type.
We also determined the average impulse count and rate of MU
firing, calculated as the reciprocal of the average interspike interval
(ISI), in each of the four phases of the respiratory cycle for 10 –15
consecutive breaths in quiet wakefulness and in stable NREM sleep.
Differences in average discharge rate in each of the four phases of the
breath cycle in wakefulness and NREM sleep were tested by two-way
repeated-measures ANOVA (Phase ⫻ Condition). Post hoc pairwise
comparisons were evaluated with the Student–Neuman–Keuls proce-
dure. Statistical significance was set at P ⬍ 0.05.

RESULTS

Single motor unit (SMU) recordings were obtained at loca-

tions about 1.0 –1.5 cm on either side of the midline and at a
distance about 2.0 –3.5 cm from the inferior margin of the FIG. 2. Frequency histograms of the population of GG motor units (MUs)
mandible (Fig. 1). The average depth to the inferior border of as a function of the consistency and strength of their respiratory activity as
the GG muscle was 12–14 mm and SMU activities were estimated by ␩2, in wakefulness (n ⫽ 81) and NREM sleep (n ⫽ 64).

J Neurophysiol • VOL 98 • DECEMBER 2007 • www.jn.org

 STATE-DEPENDENT MODULATION OF LINGUAL MOTOR UNIT ACTIVITIES 3287

TABLE 1. Categorization of motor unit activities based on sleep, the tonic discharge pattern persisted but the magnitude
traditional classification scheme of the respiratory modulation was diminished (␩2 ⬍0.30)
relative to quiet wakefulness.
Discharge Pattern Wakefulness NREM Sleep A second group of MUs discharged phasically in wakeful-
Wake 3 NREM Sleep ␩2 (n) (n)
ness and converted to tonic discharge in NREM sleep (n ⫽ 10).
Tonic 3 tonic Weak 13 22 These MUs, designated phasic–tonic, were strongly modulated
Moderate 25 16 (␩2 ⬎0.6) in wakefulness (see Fig. 4) attaining peak discharge
Strong 0 0
Phasic 3 tonic Weak 0 6
in early inspiration and falling silent during expiration. In
Moderate 0 4 NREM sleep, the strength of the respiratory modulation de-
Strong 10 0 clined (␩2 ⫽ 0.3– 0.6) and the discharge pattern changed such
Tonic 3 phasic Weak 4 4 that activities persisted through all phases of the respiratory
Moderate 9 0
Strong 0 9 cycle.
Phasic 3 phasic Weak 0 0 Representative recordings of the third MU pattern (n ⫽ 13),
Moderate 0 0 designated tonic–phasic, are depicted in Fig. 5. These MUs
Strong 3 3 discharged tonically in wakefulness converting to a phasic
Tonic 3 extinction Weak 3 0
Moderate 0 0 discharge pattern in NREM sleep. The majority of these MUs
Strong 0 0 (n ⫽ 9) exhibited moderate to weak respiratory modulation in
Phasic 3 extinction Weak 0 0 wakefulness (␩2 ⬍0.6) that increased in NREM sleep (␩2
Moderate 0 0 ⬎0.60). However, a smaller subset (n ⫽ 4) appeared distinct

Downloaded from jn.physiology.org on November 10, 2009

Strong 14 0
Leftmost column shows motor units (MUs) grouped in accordance with
traditional classification scheme; i.e., phasic or tonic based on discharge
pattern evident in wakefulness and in stable NREM sleep. Rightmost columns
show the corresponding number of MUs in each eta category: weak (␩2 ⬍0.3),
moderate (␩2 ⫽ 0.3– 0.6), and strong (␩2 ⬎0.6) respiratory modulation in
wakefulness and in NREM sleep.
sleep (P ⬍ 0.064), although the clear bimodality of the distri-
bution was lost.
We subsequently categorized MU activities using traditional
designations, i.e., tonic or phasic based on the pattern of
discharge in wakefulness and in NREM sleep and on the basis
of the corresponding ␩2 category: weak, moderate, or strong in
each state (see Table 1). There was no significant effect of state
on MU type (P ⬍ 0.258).
Representative raw recordings of the prevailing MU types
are shown in Figs. 3– 6. Of the total number of MUs recorded
in wakefulness and NREM sleep (n ⫽ 64), the majority (n ⫽
38) exhibited activities comparable to those depicted in Fig. 3.
These MUs typically discharged tonically, i.e., throughout
inspiration and expiration in quiet wakefulness with moderate
respiratory-related modulation (␩2 ⫽ 0.3– 0.5). In NREM
from all other MUs in that tonic discharge in wakefulness
converted to episodic bursting activity in NREM sleep (see
Fig. 6) with negligible respiratory modulation (␩2 ⬍0.3).
Figure 7 shows corresponding impulse histograms for the
MUs depicted in Figs. 3–5. Each histogram depicts the impulse
count/bin (0.2 s) for inspiratory (I) and expiratory (E) portions
of two breath cycles in wakefulness and in NREM sleep and
the associated ␩2 value for that motor unit. The topmost
histograms (Fig. 7A) are representative of motor units with
tonic–tonic discharge patterns with weak respiratory mod-
ulation (␩2 ⬍0.3) in wakefulness and sleep. By comparison,
phasic–tonic MU types (Fig. 7B) were characterized by
strongly respiratory modulation in wakefulness (␩2 ⬎0.6)
but exhibited weak to moderate respiratory modulation in
NREM sleep (␩2 ⫽ 0.3– 0.5) (P ⬍ 0.05). In contrast, motor
units with tonic–phasic discharge patterns (Fig. 7C) exhib-
ited the reverse pattern of activity, i.e., moderate respiratory
modulation (␩2 ⫽ 0.3– 0.5) in quiet wakefulness that con-
verted to strong respiratory-related modulation (␩2 ⬎0.6) in
NREM sleep (P ⬍ 0.05).
The effects of sleep on discharge frequency also varied.
Figure 8 depicts group average discharge frequencies for the

FIG. 3. Representative raw recordings of

a MU with a tonic–tonic discharge pattern.
These MUs discharged continuously through
the respiratory cycle in wakefulness and in
nonrapid eye movement (NREM) sleep. Top
trace: instantaneous firing rate (dots). Mid-
dle trace: discriminated motor unit poten-
tials, and insets show overlay of 10 consec-
utive potentials of the identified unit from
each trial. Bottom traces: intramuscular
whole muscle GG electromyogram (EMG),
central electroencephalogram (EEG) record-
ing site (C4/A1), right outer canthus (ROC)
electrooculogram (EOG), and lung volume
(sum of ribcage and abdomen volumes) sig-
nals.

J Neurophysiol • VOL 98 • DECEMBER 2007 • www.jn.org

3288 E. F. BAILEY, K. W. FRIDEL, AND A. D. RICE
principal MU categories in each of the four phases of the breath
cycle in wakefulness (left) and NREM sleep (right). Consistent
with the raw recordings, tonic–tonic MUs showed no signifi-
cant state-related change in average discharge frequency (av-
erage 19.8 –22.5 Hz), although the absolute distribution of
discharge frequencies broadened in NREM sleep. In contrast,
phasic–tonic and tonic–phasic MUs exhibited significant alter-
ations in discharge frequency as a function of state. Phasic–
tonic MUs attained peak frequencies (⬃19.6 Hz) in early (i.e.,
Phase 1) and, in some cases, late inspiration (i.e., Phase 2) that
declined to zero in late expiration. In NREM sleep, discharge
persisted throughout the respiratory cycle with comparable
discharge frequencies (⬃17.3–20.5 Hz) in each of the four
phases of the respiratory cycle. The tonic–phasic MUs exhib-
ited the reverse pattern, discharging at slightly higher average
frequencies (⬃22.5–28.4 Hz) throughout the respiratory cycle
in wakefulness and converting to inspiratory-only bursts in
NREM sleep. Despite a sleep-associated decline in average
discharge frequency, this decline failed to attain significance
J Neurophysiol • VOL
FIG. 4. Representative raw recordings of
the phasic–tonic discharge pattern (n ⫽ 10).
In wakefulness, these MUs discharged max-
imally in inspiration falling silent in late
expiration. In NREM sleep discharge per-
sisted throughout the respiratory cycle. De-
tails as for Fig. 3.
Downloaded from jn.physiology.org on November 10, 2009
(P ⬍ 0.07). Thus when considered as a population, the absolute
discharge frequency was preserved within a given phase of the
breath across wakefulness and sleep states.
DISCUSSION
Although studies of the whole muscle GG EMG indicate that
GG activities are preserved in NREM sleep, to date it has been
unclear what influence sleep exerts on individual motor unit
activities. The present recordings confirm a rich pattern of
tongue muscle activation in wakefulness attributable to con-
siderable heterogeneity in discharge rate and timing of GG MU
discharge with respect to phases of the respiratory cycle. We
show that NREM sleep is associated with changes in the
activities of GG MUs in terms of both the strength of
respiratory modulation and the discharge frequency. Al-
though the activities of many GG MUs that exhibit strong
respiratory-related modulation in wakefulness are extin-
guished at sleep onset, the activities of the majority of MUs
FIG. 5. Representative raw recordings of
the tonic–phasic discharge pattern (n ⫽
9/13). In wakefulness, these MUs discharged
continuously throughout the respiratory cy-
cle converting to a phasic pattern in NREM
sleep. Details as for Fig. 3.
98 • DECEMBER 2007 • www.jn.org
 STATE-DEPENDENT MODULATION OF LINGUAL MOTOR UNIT ACTIVITIES
with weak to moderate respiratory modulation persist and,
in some cases, the magnitude of the respiratory modulation
of these MUs increased. Accordingly, we report robust GG
MU activities in NREM sleep with a slight but nonsignifi-
cant increase in the overall strength of the respiratory-
related modulation (see Fig. 2).
FIG. 7. Impulse count histograms (bin duration ⫽ 0.2 s) for inspiratory (I)
and expiratory (E) portions of the breath cycle and the associated ␩2 values in
wakefulness and in NREM sleep. Impulse count histograms are derived from
MUs with tonic–tonic, phasic–tonic, and tonic–phasic discharge patterns
displayed in Figs. 3–5.
J Neurophysiol • VOL
3289
FIG. 6. Raw recordings of a second tonic–
phasic discharge pattern exhibited by a sub-
set (n ⫽ 4/13) of GG MUs. These motor
units discharged continuously through the
respiratory cycle in wakefulness, converting
to episodic bursting in NREM sleep. Details
as for Fig. 3.
Downloaded from jn.physiology.org on November 10, 2009
Critique of method
To determine whether the strength of respiratory modulation
of motor unit activities varies as a function of state we adapted
the method previously used by Orem and Dick (1983) to
characterize the proportion of a cell’s activity that is attribut-
able to a respiratory component. There are two important
distinctions between the approach outlined by Orem and Dick
(1983) and the current protocol. First, in the original method
the ␩2 value was determined on the basis of 50 breath cycles,
with each breath cycle divided into 20 equal parts (Orem and
Dick 1983). In the present study, statistical analysis of the data
was completed on a smaller numbers of breaths (10 –15
breaths) and with fewer divisions per breath cycle (4 parts).
Whereas this technique allowed us to consider activities in
terms of early inspiratory or expiratory, late inspiratory or
expiratory, the approach rendered significantly higher values of
␩2 than those reported in dorsal and ventral respiratory groups
(Orem and Dick 1983).
Respiratory modulation of MU discharge
The focus in the current study was the relative stability of the
␩2 value across wakefulness and NREM sleep. We found that
about 30% of GG MUs changed discharge pattern with sleep
and for the majority of these MUs the ␩2 value also changed.
Thus, whereas the absolute magnitude of respiratory modula-
tion of the population of MUs did not change significantly from
wakefulness to sleep (␩2 ⫽ 0.38 vs. 0.40), the effects of sleep
on individual MU activities varied considerably with about
16% of MUs exhibiting significant decreases in respiratory
modulation and about 14% exhibiting significant increases in
the strength of respiratory modulation. Thus, the strength of
respiratory modulation, although not a stable feature of indi-
vidual GG motor unit activities, may be a stable feature of the
population.
Motor unit discharge pattern
WAKEFULNESS. Only one other study has systematically quan-
tified GG single MU activities in awake human subjects (Sab-
98 • DECEMBER 2007 • www.jn.org
3290 E. F. BAILEY, K. W. FRIDEL, AND A. D. RICE
FIG. 8. Individual and average (10 –15 breath cycles) discharge frequencies
for tonic–tonic (n ⫽ 38), phasic–tonic (n ⫽ 10), and tonic–phasic (n ⫽ 9/13)
MU types in each of the 4 phases of the breath cycle (P1–P4) in wakefulness
and NREM sleep.
oisky et al. 2006). The range of average firing rates observed in
the present study, in quiet wakefulness (8 –28 Hz), concur with
this previous report for tonic and phasic GG MU activities.
Moreover, our average onset and peak firing frequencies for
tonic–tonic, phasic–tonic, and tonic–phasic units correspond
well with those reported by Saboisky et al. for MUs with
similar discharge patterns. Discrepancies in the proportion and
type of MU activities between the two studies are attributed to
differences in recording site (see Fig. 1) (Sauerland and Harper
1976).
SLEEP. For the majority of MUs (n ⫽ 41/64), the discharge
pattern evident in wakefulness was conserved in NREM sleep.
Of this majority, some 90% of the MUs (n ⫽ 38/41) discharged
throughout the breath exhibiting subtle sleep-associated in-
creases/decreases in discharge frequency suggestive of dy-
namic modulation. We were surprised to find that the diverse
range of discharge patterns evident in wakefulness was pre-
served in NREM sleep. Indeed, in a few subjects, new dis-
charge patterns emerged (see Fig. 6). Although the breadth of
J Neurophysiol • VOL
discharge patterns was preserved, many MUs (n ⫽ 23/64)
switched discharge pattern with sleep; i.e., MUs with a phasic
discharge pattern in wakefulness switched to a tonic pattern in
sleep or vice versa. Thus GG MUs cannot be considered
exclusively tonic background or phasic-respiratory; rather, the
discharge pattern appears to be an inherently flexible feature of
GG activities in healthy young adults.
Physiologic implications
Despite a large proportion (⬃30%) of MUs that exhibited
significant state-related changes in respiratory modulation and
discharge frequency, these effects were not significant when
the results were considered for the population of MUs as a
whole. Thus in healthy adults, discharge frequency and
strength of respiratory modulation are well preserved in NREM
sleep, consistent with previous studies of whole muscle GG
EMG activities under comparable conditions (Fogel et al.
2003; Worsnop et al. 1998, 2000).
Downloaded from jn.physiology.org on November 10, 2009
The significance of variability in MU firing behavior re-
mains unclear. Previous studies of MU activities in mammals
indicate that MU firing pattern varies substantially as a func-
tion of the background level of activation. For example, in-
creasing chemical drive (John et al. 2005) or altering vestibular
input (Tsuiki et al. 2000) can change a phasic GG discharge
pattern to a tonic pattern. Importantly, because sleep is asso-
ciated with both increased PaCO2 and upper airway resistance
(Fogel et al. 2003; Horner et al. 1994; Wheatley et al. 1993) it
seems likely that the changes in MU activities reported here
occurred in response to perturbations in the chemical and/or
mechanical environment that were not controlled under the
current protocol. Additional studies of this type designed to
assess the effects of blood gases (Phillipson 1978), cardiopul-
monary variables such as lung volume (Morrell et al. 2000;
Rowley et al. 2001), airway resistance (Orem et al. 1977;
Sauerland and Harper 1976; Skatrud and Dempsey 1985;
Worsnop et al. 2000), and the effects of aging and/or obesity on
GG MU activities, are therefore of critical importance and
interest.
ACKNOWLEDGMENTS
We thank Dr. Mark Borgstrom for statistical advice and data analysis, Drs.
Ralph Fregosi and Andy Fuglevand for technical advice and helpful comments
on the manuscript, and M. Connolly and L. Pittman for assisting with motor
unit discrimination.
GRANTS
This work was supported by National Institute on Deafness and Other
Communication Disorders (NIDCD) Grant K23-DC-007597 to E. F. Bailey.
The content is solely the responsibility of the authors and does not necessarily
represent the official views of the NIDCD or the National Institutes of Health.
REFERENCES
American Sleep Disorders Association (ASDA) Atlas Task Force. EEG
arousals: scoring rules and examples. Sleep 15: 174 –184, 1992.
Eastwood PR, Allison GT, Shepherd KL, Szollosi I, Hillman DR. Hetero-
geneous activity of the human genioglossus muscle assessed by multiple
bipolar fine-wire electrodes. J Appl Physiol 94: 1849 –1858, 2003.
Fogel RB, White DP, Pierce RJ, Malhotra A, Edwards JK, Dunai J,
Kleverlaan D, Trinder J. Control of upper airway muscle activity in
younger versus older men during sleep onset. J Physiol 553: 533–544, 2003.
Horner RL, Innes JA, Morrell MJ, Shea SA, Guz A. The effect of sleep on
reflex genioglossus muscle activation by stimuli of negative airway pressure
in humans. J Physiol 476: 141–151, 1994.
98 • DECEMBER 2007 • www.jn.org
 STATE-DEPENDENT MODULATION OF LINGUAL MOTOR UNIT ACTIVITIES
John J, Bailey EF, Fregosi RF. Respiratory-related discharge of genioglossus
muscle motor units. Am J Respir Crit Care Med 172: 1331–1337, 2005.
Loredo JS, Clausen JL, Ancoli-Israel S, Dimsdale JE. Night-to-night
arousal variability and interscorer reliability of arousal measurements. Sleep
22: 916 –920, 1999.
Morrell MJ, Browne HA, Adams L. The respiratory response to inspiratory
resistive loading during rapid eye movement sleep in humans. J Physiol 526:
195–202, 2000.
Netick A, Orem J. Erroneous classification of neuronal activity by the
respiratory modulation index. Neurosci Lett 21: 301–306, 1981.
Orem J, Dick T. Consistency and signal strength of respiratory neuronal
activity. J Neurophysiol 50: 1098 –1107, 1983.
Orem J, Lovering AT, Dunin-Barkowski W, Vidruk EH. Tonic activity in
the respiratory system in wakefulness, NREM and REM sleep. Sleep 25:
488 – 496, 2002.
Orem J, Netick A, Dement WC. Increased upper airway resistance to
breathing during sleep in the cat. Electroencephalogr Clin Neurophysiol 43:
14 –22, 1977.
Phillipson EA. Control of breathing during sleep. Am Rev Respir Dis 118:
909 –939, 1978.
Rechtschaffen A, Kales A. A Manual of Standardized Terminology, Tech-
niques and Scoring System for Sleep Stages of Human Subjects. Washing-
ton, DC: National Institute of Public Health, 1968.
J Neurophysiol • VOL 98 • DECEMBER 2007 •
3291
Rowley JA, Sanders CS, Zahn BR, Badr MS. Effect of REM sleep on
retroglossal cross-sectional area and compliance in normal subjects. J Appl
Physiol 91: 239 –248, 2001.
Saboisky JP, Butler JE, Fogel RB, Taylor JL, Trinder JA, White DP,
Gandevia SC. Tonic and phasic respiratory drives to human genio-
glossus motoneurons during breathing. J Neurophysiol 95: 2213–2221,
2006.
Sauerland EK, Harper RM. The human tongue during sleep: electromyo-
graphic activity of the genioglossus muscle. Exp Neurol 51: 160 –170,
1976.
Skatrud JB, Dempsey JA. Airway resistance and respiratory muscle function
in snorers during NREM sleep. J Appl Physiol 59: 328 –335, 1985.
Tsuiki S, Ono T, Ishiwata Y, Kuroda T. Functional divergence of human
genioglossus motor units with respiratory-related activity. Eur Respir J 15:
906 –910, 2000.
Wheatley JR, Tangel DJ, Mezzanotte WS, White DP. Influence of sleep on
response to negative airway pressure of tensor palatini muscle and retro-
palatal airway. J Appl Physiol 75: 2117–2124, 1993.
Worsnop C, Kay A, Kim Y, Trinder J, Pierce R. Effect of age on sleep
onset-related changes in respiratory pump and upper airway muscle func-
tion. J Appl Physiol 88: 1831–1839, 2000.
Worsnop C, Kay A, Pierce R, Kim Y, Trinder J. Activity of respiratory
pump and upper airway muscles during sleep onset. J Appl Physiol 85:
Downloaded from jn.physiology.org on November 10, 2009
908 –920, 1998.
www.jn.org