You are on page 1of 23

Running Head: AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER

Arousal and cognitive load in Texas Hold em poker Psych 499C Honours Thesis January, 2012

Student: Brendan Sheehan

Faculty Advisor: Dr. Michael J. Dixon

University of Waterloo, Department of Psychology

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER Abstract Texas Hold em poker is one of the most popular gambling games today. A previous study from our lab (Lee & Dixon) investigated whether the action of bluffing could be detected overtly or covertly. Their results were unexpected and suggested cognitive load plays a crucial

role in the interpretation of arousal changes measured during strong hands, weak hands and bluff hands. We replicated this study and added a subjective rating of arousal as a new measure to assess their claims concerning cognitive load. The latter involved a baseline-mental arithmetic task which used heart rate variability (HRV). Using the baseline-mental arithmetic task we found a specific HRV variable, SDNN, was most sensitive to cognitive load. Applied to our study we found no significant differences in SDNN between the three hand types a finding which fails to support Lee and Dixons cognitive load hypothesis. Importantly, we did replicate their HR findings and showed that bluffing leads to as great an increase in arousal as a strong hand. We contend that bluffing leads to high arousal. Given the strong theoretical link between arousal, reinforcement and problem gambling, our findings may explain the popularity of this game, but also sound a cautionary warning about a key game feature that could lead to gambling problems.

Key Terms: Texas Hold em poker, Arousal, Cognitive Load, Heart Rate Variability

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER

Texas Hold em is one variant of the family of card and betting games collectively known as poker. Texas Hold em is unique in that gameplay involves both cards known only to the player as well as a set of communal cards which all players can act upon. In addition, the skill of the player has at least some effect on the outcome of the game, as opposed to betting games of pure chance such as roulette or slot machines (Shead, Hodgins, & Scharf, 2008). The game is composed of four stages of play. First, each player is dealt two cards face down. Next, three communal cards are dealt which all players can utilize. In the third and fourth round each, a single communal card is dealt. As the game progresses the player has an increasing pool of available cards from which to assemble the strongest possible combination of five cards (a hand). The strength of the players assembled hand is in relation to a predefined order of hands, standard throughout most variations of poker. After each of the four rounds of play, the player must mentally estimate the probability of their hand winning. They must also approximate the relative strength of their opponents hands. Based on these two mental estimates the player has three options at each stage: decline further play (to fold), continue playing at the current level of investment (to check) or, raise the level of investment (to raise). Each player must indicate their decision in a predefined order (e.g. clockwise around the playing surface). If a player chooses to raise, other players have the option to fold, to match the raise (to call) or to re-raise. Crucially, the intensity with which a player bets does not necessarily correlate with the objective strength of their hand. A player may decide to play a weak hand as if it were a strong one, in the hopes of deceiving his or her opponents into folding (to bluff). This strategy is made possible because of the two confidential player-dealt cards. Bluffing adds an extra interpersonal dynamic to the

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER game, as well as a considerable amount of excitement, as players must attempt to deceive their opponents into believing their hand has a high probability of winning.

Texas Hold em poker has grown immensely in popularity over the past decade. There are over 60-80 million players in the US alone, a figure that continues to rise (The Economist, 2007). Closer to home, over 20% of Ontarians play poker for money, 16% of whom play more than once a week (RGC, 2006). Yet despite this tremendous, and growing, interest in Texas Hold em little is understood about the allure of the game. There is a dearth of empirical research investigating what the attraction to Texas Hold em is or which mechanisms lie behind its addictive potentiality (Hopley & Nicki, 2010). We believe physiologic arousal may be involved in both of these aspects of the game. Arousal is a psychophysiological process linking signals from the environment with somatic and physical state alterations, which serve to increase ones ability to act upon their environment. Fowles (1980) suggests that certain stimuli form links with specific mentations, which ultimately bring about somatic and kinetic reactions (Fowles, 1980). These stimuli act upon areas of the brain, including the reticular activating system (Steriade, 1996), causing acute alterations in cardiovascular, autonomic nervous system and endocrine function within the body. The net effect of processing these external signals is ultimately an increased ability of the individual to act upon their environment in an adaptive manner (Woody & Szechtman, 2011). As a concept, arousal appears throughout a considerable portion of the gambling literature. Various mediums of gambling activity have been linked to alterations in physiological arousal including: slot machines (Brown, Rodda, & Phillips, 2004; Dixon, Harrigan, Sandhu, Collins, & Fugelsang, 2010), video poker machines (Leary & Dickerson, 1985), video lottery

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER terminals (Ladouceur, Svigny, Blaszczynski, O'Connor, & Lavoie, 2003) and horse racing (Cocco, Sharpe, & Blaszczynski, 1995). In addition, arousal features prominently throughout

problem gambling research. Blaszczynski and Nower (2002) included arousal in their influential pathways model of problem gambling. They postulated that arousal acts as a reinforcer in both an operant (e.g. intermittent wins) and in a classical (e.g. gambling stimuli, such as the smell of a casino) conditioning sense (Blaszczynski & Nower, 2002). Sharpe et al. (1995) performed an extensive analysis of arousal in both problem and non-problem gambling. They found that gambling stimuli alone, even in the absence of any actual gambling activity, were sufficient to evoke increases in physiologic arousal in problem gamblers (Sharpe, Tarrier, Schotte, & Spence, 1995). Previous work in our lab (Lee & Dixon, unpublished) set out to examine a topic at the forefront of most dedicated poker players minds: is it possible to tell, either overtly or with physiologic measures, when a person is bluffing? Twenty-four healthy, non-problem gambler, volunteers were tested in triads. Although the players believed the decks were completely randomized, the games card decks were actually pre-arranged so that each participant would receive two strong hands and four weak hands for a combined eighteen hands. Before play commenced each participant was given confidential written instructions that, should they encounter a specific weak hand, they were to bluff (to play the weak hand to win). Consequently, within the eighteen pre-arranged decks there were, for each player: two strong hands, two weak hands and two weak hands for which the players were to bluff. The dependent variables selected were chosen to measure both overt and covert effects of bluffing. These included number of eye blinks, hand gestures, mean heart rate (HR) and mean skin conductance levels (SCL).

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER Although no significant findings were obtained for the eye blink or hand gesture variables, there were two significant effects for HR and SCL. During the first recording epoch, HR was significantly higher for bluff and win hands than for fold hands. During the second

epoch, SCLs were significantly higher for bluff hands over strong hands, as well as strong hands over fold hands. Lee and Dixon interpreted these data as an interplay between cognitive load and arousal. They hypothesized that the initial increases in HR for strong and bluff hands, during the first epoch, mostly reflected an increase in cognitive load. That is, participants mentally worked to acquaint themselves with game play in a novel environment (in the lab, with electrodes affixed) and more importantly, worked to decide how much to bet in order to maximize their chances of winning. Lee and Dixon hypothesized that, together, these demands amounted to a partial depletion of the participants cognitive resources during the first epoch. This would have precluded the participants from actively engaging in the deception required in bluffing and thus explained the negligible increases in arousal (as measured by SCL) during epoch one. During the second epoch, having been acclimatized to game play, the players would have had sufficient cognitive resources available to actively engage in deception. This would explain the elevated SCL during the bluff hands in the second epoch. One potential criticism of the Lee and Dixon study was the exclusive reliance on psychophysical measures to assess arousal. Their hypotheses concerning arousal, cognitive load, and how these factors interact with hand type would have been strengthened had subjective measures of arousal supplemented the objective measures. The current study has three purposes: replication of the previous study, adding subjective measures of arousal to complement psychophysical measures of arousal, and a specific

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER investigation of the effects of cognitive load on poker gameplay. To more closely examine cognitive load we will be making two additions to Lee and Dixons experimental design: a subjective rating of arousal and a baseline-mental arithmetic task utilizing heart rate variability

(HRV). To compare the participants subjective feelings of arousal with objective measurements of cognitive load we will use a 9-point Likert-type visual analogue scale (VAS; Morris, 1995). To analyze their theory on the interaction of cognitive load and poker gameplay we will also have the participants complete a baseline-mental arithmetic task. Specifically the serial sevens subtraction task (Pennington, 1947; Smith, 1967; Frigy, Varga, Orbn, & Incze, 2005). For this task we will utilize HRV measures as it has been suggested they are an accurate indicator of cognitive load (Gunther Moor, Crone, & van der Molen, 2010; Verkuil, Brosschot, Borkovec, & Thayer, 2009). HR differs from HRV in its composition. As opposed to measuring the mean rate at which the heart beats, HRV is a number of statistical processes performed on a series of interheartbeat intervals recorded over an epoch of interest (Clifford, 2002). Overall, we intend to measure cognitive load and arousal in volunteer participants as they play Texas Hold em poker. During gameplay, we expect cognitive load to be higher during the first epoch than during the second. We predict these maximal levels of cognitive load will abolish any arousal responses to the differing hand types, indicated by indifferent SCL responses. In the second epoch, after the participants become familiar with the game, HR will decrease for bluffing and strong hands (becoming equivalent to fold hands), but SCL will now differentiate the hand types: bluff hands will have the highest SCLs, strong hands next highest, and fold hands the least.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER Methodology Design

Our study was a two-factor, repeated-measures design. The first factor was type of hand (strong hand, weak hand, bluff hand). The second factor was epoch (the first iteration during which a participant received each of the three hands and the second iteration). In addition we recorded a baseline and a mental arithmetic task for each participant. Procedure After entering the testing facility and completing informed consent procedures participants were asked to wash their hands prior to electrode attachment. Gender-matched research assistants applied the five experimental electrodes. Three participants were seated at a mock Texas Hold em gameplay table facing a dealer (a research assistant). After a ten-minute acclimatization period a 60s baseline epoch was recorded. Following this, participants completed a mental arithmetic task for 60s. The task was the serial sevens subtraction task in which participants, beginning at the number 100, silently count backwards by sevens (Smith, 1967). Following the baseline and mental arithmetic tasks, the dealer administered specific poker instructions (rules of the game). They were also told that when they saw specific card combinations they were to bluff. To ensure players understood the instructions the dealer then dealt two face-up rounds of poker, explaining each round and suggesting the best course of action to each player. Participants were informed that each would receive a prize (chips, candy bars) commensurate with the amount of playing chips they had at the conclusion of the experiment. After verbally ensuring each player was comfortable with the rules and general

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER

gameplay the eighteen experimental rounds of poker began. These eighteen rounds were played in a randomized order, selected before the experiment commenced. Participants Research participants were 39 University of Waterloo students (24 male, Mage = 20.6 years, age range: 19-25 years), completing the experiment in return for course credit. To be eligible for the experiment participants must have been: 1) between the age of 19 and 65 (the former being the legal age for gambling in the province of Ontario); 2) not currently taking any anxiety reducing medication (as this would interfere with psychophysical measures); and 3) not currently be in treatment for problem gambling. Measures A visual analogue scale (VAS) was used to allow participants to record their subjective ratings of arousal. At the conclusion of each hand, participants confidentially indicated on a 9point Likert scale how physiologically aroused they felt during the last round of gameplay. A score of one corresponded to Not at all aroused and nine to Very physiologically aroused. The VAS scale uses images of a cartoon character to depict varying levels of arousal in the case a participant is not familiar with the concept of physiologic arousal (Morris, 1995; see Appendix 1). Skin conductance levels (SCL) were used to measure the participants objective levels of arousal during strong, weak and bluffing hands. To minimize the effect of baseline drift over time we used delta SCLs. The participants skin conductance rate (measured in microsieverts; Sv) at the beginning of the epoch of interest was subtracted from the highest maxima achieved throughout the epoch.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER Heart rate (HR) data were used to measure the participants levels of cognitive load.

10

Mean beats per minute (bpm) were calculated from an average of interbeat intervals, producing a mean instantaneous heart rate (HRi) for each epoch of interest. Heart rate variability (HRV) data were used to assess Lee and Dixons contention that increased levels of cognitive load lead to their unexpected results (increased HR in epoch one for strong and bluff hands, without any effect on SCL). HRV data consist of various statistical operations performed on a series of interbeat intervals, recorded over an epoch of specific length (Task Force of the European Society of Cardiologists, 1996). Given the ultra-short duration of our recording epochs (60s) we limited our HRV analysis to time-domain measures. The current consensus of the literature is that frequency-domain analysis of epochs of less than five minutes duration is ill-advised as the data are too susceptible to artifact. We used three redundant measures, as the current recommendations are to use two or more when performing HRV analysis within a psychophysiological study (Clifford, 2002). We selected the three most popular time-domain measures: SDNN, NN50 and RMSSD. As opposed to increasing in response to increased cognitive load (as HR does), HRV decreases. This signals a reduction in the variability of the interbeat intervals. Heartbeats deemed to be artifact in nature were not included in these analyses. SDNN is the standard deviation, of the intervals between each pair of adjacent normal heartbeats (all non-artifact, non-ectopic R-R intervals). NN50 is the number of successive, normal, interbeat intervals which differ by more than 50 ms. This range is somewhat arbitrarily selected and is used as another indicator of the variability of the heart rate over an epoch of interest.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER

11

RMSSD is a rather complicated statistical operation: the root, of the mean, of the square, of the standard deviations, of all of the intervals between normal heartbeats over an epoch of interest. Apparatus SCL, HR and HRV information were acquired through an eight-channel AD Instruments Powerlab (model 8/30; Powerlab; Colorado Springs, CO, USA). SCL data were collected via two velcro-electrodes placed at the distal phalanges of the index and ring fingers of the participants non-dominant hand. HR and HRV data were collected from three self-adhesive electrodes placed on the participants skin in a modified Mason-Likar arrangement (Mason & Likar, 1966; see Appendix 2). This arrangement places two electrodes in the infraclavicular fossae, 2cm medial to the deltoid border. A third electrode, acting as an earth ground, is placed on the abdomen in the left anterior axillary line, 3-4 cm superior to the iliac crest. The Mason-Likar arrangement considerably reduces movement artifact from skeletal muscle allowing the participants to move more freely during gameplay, an otherwise considerable challenge to the ecological validity of the study. Data files were analyzed in LabChart Version 7.2.3 on a MacBook laptop. Exclusions were set with a macro to analyze only the twenty 60s epochs of interest (baseline, mental arithmetic and eighteen poker hands) for each participant. As HRV analysis cannot be conducted on epochs of unequal lengths (Clifford, 2002) any quickly played poker hands resulting in an epoch of less than 60s were excluded. A digital bandpass filter was applied to the data signal with frequencies individually selected to optimize the signal-to-noise ratio of R-waves to other cardiac waveforms. All variables were outlier-corrected to within 2.5 standard deviations of their mean.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER Results Data Attrition Unfortunately, given the complexity of the experiment, a considerable portion of

12

experimental epochs were lost. As players engaged in poker gameplay (reaching across the table to pick up cards or place their bets) movement artifact rendered some epochs completely unanalyzable. Compounding these losses is the aforementioned criteria that all HRV epochs to be compared must be of equal duration. Taken together these factors caused the sample size to be as low as 20 cases in some of the dependent variables (detailed specifically in the ensuing tables). Arousal and Cognitive Load As summarized in Table 1. Table 1 Arousal and Cognitive Load
VAS (n=33) Type of Hand Mean SD SCL (n=24) Mean First Epoch Strong Weak Bluff 6.58 3.36 5.82 1.275 1.558 1.811 2.72 1.483 2.603 Second Epoch Strong Weak Bluff 5.73 3.45 5.00 1.275 2.181 2.107 2.41 1.548 1.960 1.768 1.152 1.587 83.81 80.31 81.55 10.821 11.478 9.011 2.226 1.532 1.187 82.75 80.52 85.60 10.835 11.274 10.789 SD HR (n=34) Mean SD

Note: VAS was based on a 9-point Likert scale, SCL measured in Sv, HR measured in mean bpm.

VAS had a main effect of type of hand, F(2,64) = 39.761, MSE = 3.340, p<.000, n2 = 0.554. In addition, there was a main effect of epoch order, F(1, 32) = 6.013, MSE = 2.271, p<.02,

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER n2 = 0.158. The type of hand by epoch order interaction was not significant, F(2, 64) = 1.866, MSE = 2.52, p<n.s.. To further understand the type of hand main effect we conducted Bonferroni-corrected pairwise comparisons. Participants subjective ratings of their arousal during strong hands (M=6.342) were significantly higher than during both bluff hands (M=5.434), t(75) = 3.543, SEM = 0.256, p<.001, and weak hands (M=3.533), t(76) = 10.694, SEM = .266, p<.000. As well, bluff hands were significantly higher than weak hands, t(75) = 6.182, SEM = .302, p<.000. SCL had a main effect of type of hand, F(2,46) = 6.775, MSE = 2.078, p<.003, n2 =

13

0.228. Neither the epoch order, F(1, 23) = 1.175, MSE = 2.692, p<.n.s., nor the type of hand by epoch order interaction, F(2, 46) = 0.841, MSE = 1.797, p<n.s., were significant. Bonferronicorrected pairwise comparisons conducted on type of hand revealed participants objective arousal levels during strong hands (M=2.562) were significantly higher than during weak hands (M=1.515) t(47) = 20.907, SEM = .343, p<.017. Bluff hands (M=2.281) were significantly higher than weak hands as well, t(47) = 30.008, SEM = .175, p<.001. Bluff and strong hands were not significantly different, t(47) = 5.747, SEM = 0.334, p<n.s.. HR had a main effect of type of hand F(2,66) = 8.175, MSE = 25.327, p<.001, n2 = 0.199. In addition, there was a main effect of the type of hand by epoch order interaction, F(2, 18) = 6.091, MSE = 19.725, p<.004, n2 = 0.156. Epoch order was not significant, F(1, 33) = 1.951, MSE = 29.592, p<.n.s. Bonferroni-corrected pairwise comparisons of the main effect of type of hand indicated that participants heart rates during strong hands (M=83.278) were significantly higher than during weak hands (M=80.413), t(67) = 24.764, SEM = 0.947, p<.014. Bluff hands (M=83.572) were significantly higher than weak hands as well, t(67) = 32.403, SEM

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER = 0.798, p<.001. Strong and bluff hands were not significantly different, t(67) = 2.872, SEM = 0.838, p<.n.s.. To unpack the type of hand by epoch order interaction, we conducted simple main effects comparisons to analyze the two epochs separately. Bonferroni-corrected pairwise comparisons indicated that, in epoch one, bluff hands (M=85.594) lead to significantly higher heart rates than strong hands (M=82.741), t(33) = 2.61, SEM = 1.092, p<.014, and than weak

14

hands (M=80.518), t(33) = 3.99, SEM = 1.271, p<.000. In epoch two, strong hands (M=83.809) lead to higher heart rates than weak hands (M=80.309), t(33) = 2.7, SEM = 1.294, p<.011. Bluff hands (M=81.549) were not significantly higher than weak hands in the second epoch t(33) = 1.27, SEM = 0.977, p<n.s.. Baseline-Mental Arithmetic task As shown in Table 2, skin conductance levels were not significantly different between the resting baseline condition and the mental arithmetic task. As previously noted, HRV has been used to track cognitive load. Of the three HRV variables only SDNN significantly responded to the increases in cognitive load brought on by the mental arithmetic task, t(37) = 2.948, SEM = 2.403, p<.006. As such this measure was used to address whether there was a preferential increase in cognitive load during the first epoch of the poker task. Table 2 Four variable analysis of the Baseline-Mental Arithmetic task
SCL Mean Baseline Mental Arithmetic Difference 0.759 1.055 0.296 SD 1.167 1.853 SDNN Mean 52.986 45.902 -7.084* SD 16.035 14.077 NN50 Mean 13.3 11.5 -1.8 SD 9.17 9.61 RMSSD Mean 47.476 42.561 -4.915 SD 24.593 26.511

Note: SCL measured in Sv, SDNN and RMSSD measured in ms, NN50 is simple occurrence count; SCL N=39, SDNN, NN50 and RMSSD N=38; * p<.006; SCL, NN50 & RMSSD p< n.s.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER As shown in Table 3, SDNN showed neither a main effect of type of hand, F(2, 38) =

15

0.218, MSE = 100.549, p<.n.s., epoch order, F(1, 19) = 1.414, MSE = 54.719, p<.n.s., or a type of hand by epoch order interaction, F(2, 38) = 0.399, MSE = 53.217, p<.n.s. Table 3 Arousal and Cognitive Load with HRV
VAS (n=33) Type of Hand Mean SD SCL (n=24) Mean First Epoch Strong Weak Bluff 6.58 3.36 5.82 1.275 1.558 1.811 2.72 1.483 2.603 Second Epoch Strong Weak Bluff 5.73 3.45 5.00 1.275 2.181 2.107 2.41 1.548 1.960 1.768 1.152 1.587 48.040 50.731 48.583 11.872 12.547 9.695 2.226 1.532 1.187 51.034 50.819 50.320 12.840 13.841 11.903 SD SDNN (n=20) Mean SD

Note: VAS was based on a 9-point Likert scale, SCL measured in Sv, SDNN measured in ms.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER Discussion

16

The current study had three purposes: replication of the previous study, adding subjective measures of arousal to complement psychophysical measures of arousal and a specific investigation of the effects of cognitive load on poker gameplay. Replication For SCL measures, Lee and Dixon found that bluffing led to higher SCL levels than either strong or weak hands but this effect only occurred during epoch two. In our study, during epoch one, strong and bluff hands led to significantly higher SCL levels than during weak hands. For HR measures, Lee and Dixon found significantly higher heart rates for bluffing and strong hands in epoch one but not in epoch two. In our study, bluffing led to higher HR, than either strong or weak hands, during epoch one. Yet during epoch two, only strong hands lead to higher heart rates than weak hands. Subjective measures of arousal Our study added a subjective measure of arousal using VAS. We found that participants subjective ratings of arousal were highest for strong hands when compared with both bluff and weak hands. As well their subjective ratings of arousal were higher for bluff hands than for weak hands. Measuring cognitive load using HRV When baseline resting HRV was contrasted with an epoch during which participants engaged in a task known to increase cognitive load, all three measures of HRV showed nominal reductions in the predicted direction. However, only SDNN showed a significant response to cognitive load. As such, with these participants, SDNN was able to strongly track increases in

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER

17

cognitive load. After determining that SDNN was best suited for this analysis we examined the type of hand and epoch order factors alongside VAS, delta SCL and mean HR. Although SDNN reacted sharply to the mental arithmetic task there were no significant differences between strong, weak or bluff hands. If Lee and Dixon were correct in their contention that cognitive load increases elevated HR during bluffing and strong hands (and somehow masked SCL effects), then we should have shown significant reductions in SDNN during bluffing and strong hands in epoch one. However, no such effects were noted. These data suggest that the large increases in HR during the bluff hands were not due to increased cognitive load. By contrast, it appears that the deception and/or risk involved in bluffing served to elevate HR. When a player bluffs, they are holding a weak hand yet proceed as if they are playing from a position of strength. To win, the player must successfully masquerade as if they were holding a strong hand whilst under the direct scrutiny of the other players. In addition, there is considerable risk involved in bluffing, as players continue to bet with a hand that may be substantially inferior to their opponents. Outside of the gambling literature there exists strong support for the links between deception and arousal (DeTurck & Miller, 1985; Gdert, Rill, & Vossel, 2001) as well as between risk and arousal (Critchley, Mathias, & Dolan, 2001). Unlike the study of Lee and Dixon, we found converging effects for HR and SCL measures. Most importantly for both HR and SCL, significantly larger effects were noted when players attempted to bluff with weak hands compared to when they were simply going to fold a weak hand. We propose that this elevated state of arousal for bluffing may be one of the key features which makes this form of poker particularly arousing.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER

18

In our study the subjective ratings of arousal were highest for strong hands, followed by bluffing hands, and weakest for fold hands. The literature on arousal and risk led us to predict that bluffing would lead to more subjective arousal than strong hands because of the increased risk. In our experiment, players were forced to bluff on specific hands. By removing the choice of when and when not to bluff, we may have reduced arousal in this condition. Importantly for both subjective and psychophysiological measures of arousal, bluffing on a weak hand was significantly more arousing than folding on a weak hand. As afore mentioned, little is currently understood about the mechanisms underlying the allure and addictive potentially of Texas Hold em poker. There is much evidence to suggest that the mechanisms involved in the attraction and addiction of other gambling mediums is arousal. Blaszczynski and Nower (2002) cite arousal as the reinforcement process behind problem gambling (Blaszczynski & Nower, 2002). Our data suggest that arousal features prominently throughout poker. Beyond the excitement one would intuitively associate with playing a strong hand to win there is the thrill involved with bluffing. These are weak hands which, without the concept of bluffing, a player would simply toss aside with minimal arousal. Bluffing provides a mechanism whereby even weak hands can be highly arousing. When one bluffs and wins, such arousal could be highly reinforcing. A recurring challenge within gambling research is ecological validity. Both the previous study (Lee and Dixon) as well as the current study went to great pains to preserve what authenticity is available within an experiment conducted on undergraduates within the confines of a psychology research lab. We purchased a Texas Hold em mat screened with similar markings to those found in casinos and on televised poker tournaments. We utilized actual

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER gameplay chips and rewarded the participants, beyond their course credit, commensurate with

19

their performance. Despite these measures we would be interested to study arousal and cognitive load in Texas Hold em under even more realistic conditions. Although our participants were unaware that the decks were pre-arranged, or that there were multiple decks at all (the dealer concealed the swap of pre-sorted decks while shuffling), this is still far from the completely randomized nature of normal poker gameplay. Complete randomization might be possible if one could devise an apparatus to have each participant confidentially indicate whether their intentions with the current hand were to play to win from a position of strength, to fold a weak hand or to play a weak hand to win (to bluff). As well, although we endeavored to ensure at least a baseline competency in gameplay before the experimental rounds began, it would be interesting to quantify the participants skill in poker and analyze statistically whether this had a significant effect on physiologic spectra. If much of the revenue and addiction of the gambling world are predicated on generating physiologic arousal then Texas Hold em poker is a prime candidate. The proportion of time a players arousal levels are elevated during gameplay is exaggerated thanks to bluffing. As well, given the simple nature and design of the game, there is relatively little downtime in Texas Hold em. This is especially the case in on-line versions of the game, where a player can play at multiple tables at once. This means the mind and body of a poker player is subjected to constant stimuli and cues which increase arousal over the course of their play. As our understanding of the game continues to improve it is entirely possible that policy change will be required to protect future generations of poker players developing gambling problems to this potentially highly addictive game.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER References

20

Blaszczynski, A., & Nower, L. (2002). A pathways model of problem and pathological gambling. Addiction, 97(5), 487499. Brown, S. L., Rodda, S., & Phillips, J. G. (2004). Differences between problem and nonproblem gamblers in subjective arousal and affective valence amongst electronic gaming machine players. Addictive Behaviors, 29(9), 18631867. doi:10.1016/j.addbeh. 2004.03.030 Clifford, G. D. (2002). Signal Processing Methods for Heart Rate Variability. (L. Tarassenko, Ed.) (pp. 1244). Cocco, N., Sharpe, L., & Blaszczynski, A. P. (1995). Differences in preferred level of arousal in two sub-groups of problem gamblers: A preliminary report. Journal of Gambling Studies, 11(2), 221229. Critchley, H. D., Mathias, C. J., & Dolan, R. J. (2001). Neural activity in the human brain relating to uncertainty and arousal during anticipation. Neuron, 29(2), 537545. Elsevier. DeTurck, M. A., & Miller, G. R. (1985). Deception and arousal: Isolating the behavioral correlates of deception Human Communication Research. Oxford Univ Press. Dixon, M., Harrigan, K., Sandhu, R., Collins, K., & Fugelsang, J. (2010). Losses disguised as wins in modern multi-line video slot machines. Addiction, 105(10), 18191824. Economist. (2007, December 17). Poker: A big deal. The Economist. London. Retrieved March 9, 2011, Fowles, D. (1980). The Three Arousal Model: Implications of Gray's Two-Factor Learning Theory for Heart Rate, Electrodermal Activity, and Psychopathy. Psychophysiology. Fowles, D., Christie, M., Edelberg, R., GRINGS, W., Lykken, D., & Venables, P. (1981). Publication recommendations for electrodermal measurements. Psychophysiology, 18(3), 232239. Frigy, A., Varga, L., Orbn, S., & Incze, A. (2005). Effect of mental arithmetic on heart rate and qtc interval in young, healthy individuals. Folia Cardiologica, 12(supplement D), 239240.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER

21

Gdert, H., Rill, H., & Vossel, G. (2001). Psychophysiological differentiation of deception: the effects of electrodermal lability and mode of responding on skin conductance and heart rate. International Journal of Psychophysiology, 40(1), 6175. Gunther Moor, B., Crone, E. A., & van der Molen, M. W. (2010). The Heartbrake of Social Rejection: Heart Rate Deceleration in Response to Unexpected Peer Rejection. Psychological Science, 21(9), 13261333. doi:10.1177/0956797610379236 Hopley, A. A. B., & Nicki, R. M. (2010). Predictive Factors of Excessive Online Poker Playing. Cyberpsychology, 13(4), 379385. Ladouceur, R., Svigny, S., Blaszczynski, A., O'Connor, K., & Lavoie, M. (2003). Video lottery: winning expectancies and arousal. Addiction, 98(6), 733738. Leary, K., & Dickerson, M. (1985). Levels of arousal in high-and low-frequency gamblers. Behaviour Research and Therapy. Malmivuo, J., & Plonsey, R. (1995). Bioelectromagnetism: Principles and Applications of Bioelectric and Biomagnetic Fields. New York: Oxford University Press. Mason, R. E., & Likar, I. (1966). A new system of multiple-lead exercise electrocardiography. American heart journal, 71(2), 196205. Elsevier. Morris, J. (1995). Observations: SAM: The self-assessment manikin. Journal of Advertising Research, 35(6), 6368. Cambridge University Press. Pennington, L. (1947). The Serial sevens test as a psychometric instrument. American Journal of Orthopsychiatry, 17(3), 488499. Wiley Online Library. RGC. (2006, March 9). The Poker Boom: Harmless Pastime or Gateway to Gambling Problems? Retrieved March 9, 2011, Sharpe, L., Tarrier, N., Schotte, D., & Spence, S. H. (1995). The role of autonomic arousal in problem gambling. Addiction, 90, 15291540. Shead, N. W., Hodgins, D. C., & Scharf, D. (2008). Differences between Poker Players and NonPoker-Playing Gamblers. International Gambling Studies, 8(2), 167178. doi: 10.1080/14459790802139991 Smith, A. (1967). The serial sevens subtraction test. Archives of neurology, 17(1), 78. Am Med Assoc.

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER Steriade, M. (1996). Arousal-Revisiting the Reticular Activating System. Science.

22

Task Force of the European Society of Cardiologists (1996). Heart Rate Variability: Standards of measurement, physiological interpretation, and clinical use. European heart journal, 17, 354381. Verkuil, B., Brosschot, J. F., Borkovec, T. D., & Thayer, J. F. (2009). Acute autonomic effects of experimental worry and cognitive problem solving: Why worry about worry International Journal of Clinical and Health Psychology, 9(3), 439453. Woody, E. Z., & Szechtman, H. (2011). Adaptation to potential threat: The evolution, neurobiology, and psychopathology of the security motivation system. Neuroscience & Biobehavioral Reviews, 35(4), 10191033. doi:10.1016/ j.neubiorev.2010.08.003

AROUSAL AND COGNITIVE LOAD IN TEXAS HOLD EM POKER Appendices Appendix 1: VAS

23

Adapted from: Morris, J. (1995). Observations: SAM: The self-assessment manikin.

Appendix 2: Mason-Likar electrode placement

Adapted from: Malmivuo, J., & Plonsey, R. (1995). Bioelectromagnetism

You might also like