Abstract
Background and aims
This research aimed to characterize social information processing abilities in a population of regular nondisordered poker players compared to controls.
Methods
Participants completed the Posner cueing paradigm task including social cues (faces) to assess attention allocation towards social stimuli, including the effect of the presentation time (subliminal vs supraliminal) and of the emotion displayed. The study included two groups of participants: 30 regular nondisordered poker players (those who played at least three times a week in Texas Hold'em poker games for at least three months) and 30 control participants (those who did not gamble or gambled less than once a month, whatever the game).
Results
The group of regular nondisordered poker players displayed an enhancement of the inhibition of return during the Posner cueing task. This means that in valid trials, they took longer to respond to the already processed localization in supraliminal conditions compared to controls. However, our results did not evidence any particular engagement or disengagement attention abilities toward specific types of emotion.
Discussion and Conclusions
These results suggest that regular nondisordered poker players displayed social information processing abilities, which may be due to the importance to efficiently process social information that can serve as tells in live poker. The observed enhancement of the inhibition of return may permit poker players to not process a localization that has already processed to save attentional resources. Further research regarding the establishment of the IOR in other forms of gambling and with non-social cues needs to be performed.
Introduction
Poker as a social game of skills
Poker is a social game involving a competition between players. Contrary to bank games, in which participants play against the bank and can expect to win in the short term but not to be beneficiaries in the long term, social games can allow long-term gains, provided that the players use the skill gap between them and their opponents to their advantage (Bjerg, 2010). Studies about the impact of skill on poker performances have emerged, but there is still a debate regarding the relative importance of chance versus skill. Hence, poker skills may be useful in the long term and are not always observable with only a few hands (Palomäki, Laakasuo, Cowley, & Lappi, 2020), suggesting a continuum between chance and skill (Fiedler & Rock, 2009). For example, the use of strategies was demonstrated to increase the probability of winning (DeDonno & Detterman, 2008), and rankings in a tournament can predict rankings in the following competitions (Croson, Fishman, & Pope, 2008). Another study demonstrated that ranks can be predicted by previous rankings, type of strategies used and experience in playing poker, and that skills would dominate chance after 1,500 hands played (Potter van Loon, van den Assem, & van Dolder, 2015). Nevertheless, other studies were more mitigated. Indeed, when giving the same cards to expert and nonexpert players in 60 hands of “Texas Hold'em” poker, the only significant difference was that the experts handled bad cards better than the controls (Meyer, von Meduna, Brosowski, & Hayer, 2013), showing the importance of chance.
Poker players' processing of social information
Among all the skills that may be involved in poker, players engaged in live poker have to efficiently process the social information displayed by their opponents (Billings, Davidson, Schaeffer, & Szafron, 2002), as it can serve as tells. Indeed, social information permits speculation on opponents' strategies and the prediction of their actions (opponent modelling, Billings et al., 2002). Some research has highlighted the importance of social information processing in poker (Palomäki et al., 2020). For example, in a simplified poker game, a study showed that the level of trustworthiness displayed by an opponent could modulate the wager (Schlicht, Shimojo, Camerer, Battaglia, & Nakayama, 2010). Gender displayed by opponents (Palomäki, Yan, Modic, & Laakasuo, 2016) and playing with a human or a computer (Carter, Bowling, Reeck, & Huettel, 2012) may also impact poker players' decisions. Additionally, research has shown a positive association between the self-reported quality of social competencies and performance in poker (Leonard & Williams, 2015; Schiavella, Pelagatti, Westin, Lepore, & Cherubini, 2018). The cognitive assessment of social information processing can include the conscious identification of social stimuli (such as identifying an emotion displayed by supraliminal vocal or facial stimuli) or the unconscious processing of social stimuli (such as discriminating an emotion displayed by a face and presented subliminally, which is invisible). In such assessments, social stimuli, such as faces, can be used as cues to capture the attention towards one of two locations on a screen before the presentation of a target, which may be presented either at the cued or noncued location. The measurement of reaction times then allows us to evaluate the facilitating or costing effects of cues (Posner, 1980). The engagement of attention towards a cue is indeed supposed to facilitate the detection of the cued target, leading to faster reaction times. Furthermore, the disengagement of attention from a cue is supposed to slow uncued target detection. Nevertheless, stimulus-onset asynchrony (SOA, i.e., the delay between cue onset and target onset) also impacts attention allocation. Indeed, when SOA is greater than approximately 200–300 ms and up to 3,000 ms (Samuel & Kat, 2003), the phenomenon of the inhibition of return (IOR) occurs, which is an adaptive mechanism that prevents an already processed localization from being processed again and allows more attentional resources to be allocated to information located elsewhere (Klein, 2000; Tang et al., 2015). This results in slowing the detection of the cued target while accelerating the detection of the noncued target.
Objectives and hypotheses
In the present study, we were interested in determining whether a regular poker practice, defined as a repeated regular exposition to poker playing, is associated with specific social information attention abilities. To this end, we compared these abilities, which were assessed using a spatial cueing paradigm, between regular nondisordered poker players and controls with no or limited poker experience. Moreover, we were also interested in determining the types of emotions that are better processed by regular nondisordered poker players compared to controls.
For trials with an SOA lower than 200 ms, it is possible to assess the engagement and disengagement of attention towards emotional cues (faces) without the establishment of the IOR. We hypothesized that regular nondisordered poker players would display overall higher social information attention abilities than controls. In this case, lower reaction times (benefit) may be observed for cued trials (i.e., facilitation of the detection of the target due to the previous engagement of attention by the cue displayed at the same location) and higher reaction times (cost) may be observed for noncued trials (i.e., the cost related to disengaging attention from the location of the cue before detecting the target at the opposite location), regardless of the facial emotion type. Moreover, we hypothesized that regular nondisordered poker players would display higher information attention abilities compared to controls for the types of emotion that may be useful is the context of poker, i.e., joy, sadness, and surprise. This would result in higher benefit effects and lower cost effects in poker players compared with controls for trials with these emotions as cues.
In the case of trials with an SOA greater than 200 ms, the IOR is supposed to appear, which can lead to reversed effects on Posner's paradigm, i.e., costs instead of benefits for cued trials and benefits instead of costs for noncued trials. Regardless of the emotion explored, we hypothesized that regular nondisordered poker players would display a higher IOR effect than controls, which would result in a lower benefit effect for cued trials and a higher cost effect for noncued trials. Similarly, we expected a higher IOR effect for regular nondisordered poker players for the types of emotion that may be useful is the context of poker, i.e., joy, sadness, and surprise. This would result in lower benefit effects and higher cost effects for these emotions in nondisordered poker players than in controls.
Methods
This study is part of the PERHAPS research program (NCT02590211), which aims to assess the cognitive functioning of poker players with or without a gambling disorder.
Participants
This study included two groups of participants. The first group comprised regular nondisordered poker players (RPPG, n = 30). A regular playing was defined as the repeated practice of Texas Hold'em poker games at least once a week for at least three months. All regular players had to play poker in the live version, so that they had the chance to accrue knowledge pertaining to social cognitive processing. They could also play online but not exclusively.
The control group (CG, n = 30) included nonpoker players and participants who gambled less than once a month (poker or other gambling activities).
All participants were recruited from February 2017 to May 2018 through media announcements and from the registry of volunteers for research established by the research team.
Inclusion and noninclusion criteria
We included only men aged between 18 and 60 years old who presented correct vision and hearing (even after correction). We did not include women because more men tend to play poker (McCormack, Shorter, & Griffiths, 2014) and we anticipated that the gender ratio may have been very unbalanced between the two groups and would have introduce a bias. The other noninclusion criteria were as follows: (i) participants who had taken part in a pharmaceutical trial during the previous month; (ii) those who presented colour blindness; (iii) those who presented severe depression (assessed with the BDI-13); (iv) those who presented a high level of anxiety (assessed with the STAI-YB); (v) those who presented heart problems and/or had electrical implants; (vi) those who were under a gambling ban procedure; (vii) those under a guardianship or curatorship; (viii) those with any history of neurological diseases, such as a history of seizures; (ix) those who had a physical condition that could disturb the assessment; (x) those who presented a nonstable current psychiatric disorder (assessed with the MINI); (xi) those who presented a gambling disorder (assessed with the NODS); (xii) those who displayed a cognitive impairment (assessed with the MMSE); and (xiii) those who had used any psychoactive substance (self-declared) in the 8 h preceding the assessment (except for nicotine).
Measures used for inclusion and confounding factors
The Mini International Neuropsychiatric Interview (MINI) (Lecrubier et al., 1997) is a structured interview investigating mood, anxiety and addictive (substance) disorders, and psychotic syndrome.
The National Opinion Research Center DSM-IV Screen for Gambling Problems (NODS) (Gerstein et al., 1999) is a structured interview investigating the DSM-IV diagnostic criteria for pathological gambling. We used a modified version to follow DSM-5 changes (American Psychiatry Association, 2013) (i.e., removing the item exploring illegal acts and using a threshold of 4 instead of 5).
The 13-item Beck Depression Inventory self-report questionnaire (BDI-13) (Beck, Steer, & Carbin, 1988) is a questionnaire assessing the level of depression. A score ≥16 indicates a severe level of depression.
The trait version of the State-Trait Anxiety Inventory self-report questionnaire (STAI Y-B) (Spielberger, 1999) assesses anxiety. A score ≥56 indicates a high level of anxiety.
The Mini Mental State Examination (MMSE) (Folstein, Folstein, & McHugh, 1975) assesses the global cognitive level. A score ≤24 indicates cognitive impairment.
Measures used to investigate social attention allocation
We used the Posner Cueing Paradigm (or Spatial Cueing Paradigm) to assess attention engagement and disengagement towards social cues. This paradigm includes a cue to allocate the participant's attention to a location that may later contain a target and manipulates the spatial validity between the location of the target related to the location of the cue (Hayward & Ristic, 2013; Posner, 1980).
For each trial, two grey boxes (5.3 cm height × 3.0 cm width) appeared symmetrically on each side of a fixation cross for 1,000 ms. Then, a cue appeared on the centre of one of the two boxes for 17 ms (subliminal condition, automatic processing) or 500 ms (supraliminal condition, strategic processing, supposed to trigger the IOR).
The cue was a picture of a face displaying an emotion (surprise, disgust, joy, anger, fear, sadness or neutral), which was the same size as the grey boxes and extracted from the Karolinska Directed Emotional Faces (KDEF) corpus (Lundqvist, Flykt, & Öhman, 1998). In order to select which pictures to use, we performed a preliminary validation study on 15 volunteers using a procedure adapted from Bradley, Mogg, White, Groom, and De Bono (1999). Each picture from the KDEF corpus was rated by ten volunteers. They had to rate how much the picture elicited each of the following six emotions: surprise, disgust, joy, anger, fear, sadness, on a scale from 1 (not at all) to 6 (extremely). Emotional pictures were retained only if the mean rating of the corresponding emotion was higher than 3.5 and if the mean ratings of the other emotions were at least one point below the mean rating of the corresponding emotion. Neutral pictures were retained only if the mean ratings of all emotions were under 2. Pictures with too much luminosity were removed. Finally, we selected only the four best pictures for each emotion (surprise, disgust, joy, anger, fear, sadness or neutral), i.e. those which had the higher mean rating of the corresponding emotion for emotional pictures and those which had the lowest mean ratings of all emotions for neutral pictures.
A mask without any stimulus (the two grey boxes) was then displayed for 17 ms in the subliminal condition and 50 ms in the supraliminal condition. Afterwards, a target (a black circle with a 0.3 cm diameter) appeared on the lower half of one of the grey boxes for 2,000 ms or until the participant responded (see Fig. 1). This shifted down presentation of the cue, and the target was chosen to avoid a masking forward effect (i.e., when the face cue masks the following target) (Fox, Russo, & Dutton, 2002).
Flow diagram of the Posner cueing task. Fig. 1 (a) Flow diagram of the Posner cueing task - Valid subliminal trial (b) Flow diagram of the Posner cueing task - Non valid supraliminal presentation
Citation: Journal of Behavioral Addictions 12, 1; 10.1556/2006.2022.00082
Participants were instructed to localize the target as quickly as possible by pressing the extreme left or right button on a 7-button response pad. When the target appeared at the cued location, the trial was considered valid; otherwise, the trial was considered invalid. For noncued trials, rectangles were also presented in both conditions (sub- and supraliminal) but without pictures, followed by the mask (also for 17 ms or 50 ms). This type of trial was a control measure of reaction times without cueing (without the previous allocation of attention). To prevent automatic responses, the intertrial intervals randomly (without replacement) varied for 500, 750, 1,000 or 1,500 ms.
Each participant underwent 224 trials. Half of the trials (112) were valid, one quarter (56) were nonvalid, and one-quarter were noncued. Half of each type of trial (valid, nonvalid and noncued) was supraliminal, and half was subliminal. For each type of trial (valid, nonvalid and noncued) and each presentation (supraliminal and subliminal), each of the 7 emotions was displayed the same number of times on the left and on the right.
The presentation order of the type of trials, presentation times and type of emotions was randomized. At the beginning of the task, participants practised for eight trials (4 valid, 2 nonvalid and 2 noncued) with pictures that were not used for the rest of the task. A break was added at the end of the 112th trial to avoid fatigue. Anticipative answers (i.e., in which participants answered in less than 250 ms) were excluded before data processing.
The outcomes recorded were reaction times and errors (omissions, i.e., the participant did not answer, or commissions, i.e., the participant chose the wrong side). Regarding the outcomes calculated, the benefit index (reaction time of noncued trials – reaction time of valid trials) reflects the decrease in reaction times due to the previous engagement of attention induced by the cue in a valid subliminal trial compared to a noncued subliminal trial. In the supraliminal condition, the benefit index was expected to decrease because of the higher reaction times in valid trials due to the IOR onset. Moreover, the cost index (reaction time of nonvalid trials – reaction time of noncued trials) reflects the increase in reaction times due to the necessary disengagement of attention from the location of the cue in a nonvalid subliminal trial compared to a noncued subliminal trial. In the supraliminal condition, the cost index was expected to decrease because of the lower reaction times in nonvalid trials due to IOR onset.
Procedure
After screening for eligibility, an appointment was proposed in the research centre to proceed with the inclusion and carry out the research visit. Sociodemographic data (age and educational level), concomitant treatments and measures for inclusion were first collected.
Then, the participants completed a subset of 5 cognitive tasks, including the Posner Cueing Task, with a total duration of 2 h. The task order was randomized for each participant, but the Posner Cueing Task was always performed first (to avoid alterations in attention). The Posner Cueing Task was computer-administered.
All participants were seated in a quiet room and positioned 60 cm from the screen. Computerized tasks were programmed using Superlab 5 (Cedrus Corporation, San Pedro, CA, USA), and the 7-button response pad RB-730 (Cedrus Corporation) was used to record the participants' responses.
Participants received 30€ at the end of the appointment in compensation for their participation.
Statistical analysis
A descriptive analysis was performed to determine the mean and standard deviation of all variables.
To consider possible confounding factors that may have affected attentional performance, age, education level, MMSE score, BDI score, STAI score, concomitant treatment and current mood, and anxiety, psychotic and alcohol or substance use disorders were compared between the two groups using non-parametric tests (Mann-Whitney or Fisher's exact test). Given the sample size, we used the threshold of P < 0.10 to include the corresponding potential confounding factor as a covariate in the subsequent statistical analyses.
We first tested the interaction term
Statistical analyses were performed using Stata 17.0 software. We used the mixed command to fit the linear mixed models with maximum likelihood estimation, and used the margins command to compute the adjusted predictions (estimated means) from the final model and to estimate the corresponding contrasts. All model assumptions (normality and homoscedasticity of the residuals and the BLUPs (Best Linear Unbiaised Predictions) of the random effects) were tested and satisfied.
Ethics
The PERHAPS protocol received approval from the French Research Ethics Committee (CPP) on September 12, 2016. All participants were informed about the study and provided their written informed consent.
Results
Sociodemographic, clinical, and gambling characteristics
The two groups differed for four potential confounding factors at the threshold of P < 0.10, namely, age (P = 0.021), level of education (P = 0.028), score on the MMSE (P = 0.030) and score on the BDI (P = 0.079) (see Table 1). Consequently, these four variables were included in all comparative analyses as covariates. Regarding the gambling profile, only 5 participants from the CG declared that they had played poker during their lifetime, of which only 2 declared that they played less than once a month in the past year. Regarding the other gambling activities, only 2 participants from the RPPG bet on sports more than once a week. Otherwise, no gambling activity was practised at this frequency.
Sociodemographic, clinical, and gambling characteristics
| Variables | Control Group (CG) n = 30 | Recreational Regular Poker Player Group (RNDPPG) n = 30 | P |
| Mean (sd) or n | |||
| Sociodemographic characteristics | |||
| Age | 29.13 (10.2) | 33.23 (7.8) | 0.021 |
| Level of education (years) | 14.7 (1.95) | 13.33 (2.59) | 0.028 |
| Clinical characteristics | |||
| MMSE score | 29.41 (0.9) | 28.87 (1.1) | 0.030 |
| BDI score | 1.23 (2.5) | 1.83 (2.1) | 0.079 |
| STAI score | 31.8 (8.6) | 33.5 (7.2) | 0.382 |
| Presence of concomitant pharmacological treatments | 2 | 5 | 0.143 |
| MINI diagnosis (number of subjects with a positive diagnosis) | |||
| Actual mood disorder | 1 | 1 | 1.000 |
| Actual anxious disorder | 1 | 1 | 1.000 |
| Actual addictive disorder | 2 | 4 | 0.286 |
| Actual psychotic disorder | 1 | 0 | 1.000 |
| Gambling characteristics | |||
| Number of games currently played | Number of participants per number of games | ||
| 0 | 17 | 0 | |
| 1 | 9 | 3 | |
| 2 | 3 | 9 | |
| 3 | 1 | 9 | |
| 4 | 0 | 4 | |
| 5 | 0 | 5 | |
| Frequency of games played | |||
| Lottery (number of players who had already played in their lifetime) | 22/23 | 29/30 | |
| Current frequency (past 12 months) | |||
| More than once a week | 0 | 1 | |
| Once a week | 0 | 4 | |
| Once or more a month | 0 | 2 | |
| Less than once a month | 9 | 11 | |
| Did not play in the actual period | 13 | 11 | |
| Slot machines (number of players who had already played in their lifetime) | 12/23 | 22/30 | |
| Current frequency (past 12 months) | |||
| More than once a week | 0 | 0 | |
| Once a week | 0 | 0 | |
| Once or more a month | 0 | 1 | |
| Less than once a month | 2 | 7 | |
| Did not play in the actual period | 10 | 14 | |
| Blackjack (number of players who had already played in their lifetime) | 2/23 | 14/30 | |
| Current frequency (past 12 months) | |||
| More than once a week | 0 | 0 | |
| Once a week | 0 | 0 | |
| Once or more a month | 0 | 2 | |
| Less than once a month | 0 | 3 | |
| Did not play in the actual period | 2 | 9 | |
| Horse race betting (number of players who had already played in their lifetime) | 9/23 | 15/30 | |
| Current frequency (past 12 months) | |||
| More than once a week | 0 | 0 | |
| Once a week | 0 | 0 | |
| Once or more a month | 0 | 1 | |
| Less than once a month | 1 | 6 | |
| Did not play in the actual period | 8 | 8 | |
| Sport betting (number of players who had already played in their lifetime) | 11/23 | 26/30 | |
| Current frequency (past 12 months) | |||
| More than once a week | 0 | 2 | |
| Once a week | 0 | 1 | |
| Once or more a month | 0 | 3 | |
| Less than once a month | 4 | 15 | |
| Did not play in the actual period | 6 | 5 | |
| Not known | 1 | ||
| Poker (number of players who had already played in their lifetime) | 5/23 | 30/30 | |
| Current frequency (past 12 months) | |||
| More than once a week | 0 | 26 | |
| Once a week | 0 | 4 | |
| Once or more a month | 0 | 0 | |
| Less than once a month | 2 | 0 | |
| Did not play in the actual period | 3 | 0 | |
Bold P-values are those under the threshold of 0.10, and the corresponding variables were including as covariates in the subsequent analyses.
Regarding the RPPG, poker players all had poker as their preferred type of gambling among all the games they experimented, even if the large majority of them also played to other gambling activities. They all played both live and online poker, but more than half of them (53.3%) preferred gambling live poker. They started gambling on average at 15 years old (m = 14.9, sd = 3.9), and were mainly initiated through scratch cards (66.7%) or poker (13.3%). They had on average a regular practice of gambling for close to 10 years (m = 9.6, sd = 6.1).
Posner cueing task
Benefit index
The estimated coefficients of the linear mixed model for the benefit index are presented in Table 2, and effects are plotted in Fig. 2. The variance explained by the fixed effects was 5.7% and the variance explained by both the fixed and random effects was 16.5%.
Final model from the linear mixed model analysis on the benefit index
| Variables | Coefficient | Standard error | Confidence interval | P value |
| Group (ref. CG) | −5.69 | 4.90 | [−15.29; 3.92] | 0.246 |
| Presentation (ref. subliminal) | −1.25 | 3.48 | [−8.07; 5.58] | 0.720 |
| Group × Presentation | −10.54 | 4.88 | [−20.11; −0.97] | 0.031 |
| Covariates | ||||
| Age | −0.29 | 0.23 | [−0.16; 0.74] | 0.213 |
| Level of education (years) | 0.29 | 1.00 | [−1.67; 2.25] | 0.769 |
| MMSE score | −2.79 | 2.25 | [−7.20; 1.63] | 0.216 |
| BDI score | 1.61 | 0.90 | [−0.15; 3.37] | 0.074 |
Italic lines represent the covariates integrated in the model as potential confounding factors.
Adjusted predictions and 95% confidence intervals of the benefit index from the final model
Note: Vertical bars represent the 95% confidence intervals of the adjusted predictions. CG: Control Group; RPPG: Regular Poker Player Group
Citation: Journal of Behavioral Addictions 12, 1; 10.1556/2006.2022.00082
There was no significant effect of the type of emotion on the benefit index.
With the supraliminal presentation, the benefit index in the RPPG group was 16.2 points [95% CI: −25.8; −6.6] lower than in the CG (P < 0.001). With the subliminal presentation, this difference was not significant (−5.7 [−15.3; −3.9], P = 0.246).
Moreover, there was no significant effect of the presentation time in the CG (−1.2 [−8.1; 5.6]. However, for the RPPG, the benefit index with the supraliminal presentation was 11.8 [−18.5; −5.1] points lower than with the subliminal presentation. This seems to reflect the establishment of an IOR in supraliminal trials, that was stronger for the RPPG compared to the CG.
Cost index
The estimated coefficients of the linear mixed model for the cost index are presented in Table 3, and effects are plotted in Fig. 3. The variance explained by the fixed effects was 5.4% and the variance explained by both the fixed and random effects was 18.9%.
Final model from the linear mixed model analysis on the cost index
| Variables | Coefficient | Standard error | Confidence interval | P value |
| Presentation (ref. subliminal) | −13.82 | 2.85 | [−19.40; −8.23] | <0.001 |
| Covariates | ||||
| Age | 0.51 | 0.23 | [0.05; 0.96] | 0.030 |
| Level of education (years) | 2.25 | 1.02 | [0.26; 4.24] | 0.027 |
| MMSE score | 1.52 | 2.28 | [−2.95; 5.99] | 0.505 |
| BDI score | −1.42 | 0.92 | [−3.22; 0.38] | 0.122 |
Italic lines represent the covariates integrated in the model as potential confounding factors.
Adjusted predictions and 95% confidence intervals of the cost index from the final model Note: Vertical bars represent the 95% confidence intervals of the adjusted predictions.
Citation: Journal of Behavioral Addictions 12, 1; 10.1556/2006.2022.00082
There was no significant difference on the cost index between CG and RPPG. The cost index with the supraliminal presentation was 13.8 [−19.4; −8.2] points lower than with the subliminal presentation.
Discussion
This study assessed the social allocation of attention in poker players with repeated regular exposition to poker playing compared to controls with no or limited poker experience, using a Posner paradigm task. Regardless of the group, this study showed that the disengagement of attention was facilitated in nonvalid supraliminal trials compared to nonvalid subliminal trials. When comparing the two groups, this study showed that the IOR was higher for the poker player group than the CG.
Indeed, the analysis of nonvalid trials showed that the cost index diminished significantly in supraliminal trials. These results suggest that the nonvalidity slowing is especially linked to lower SOA and that higher SOA already permits disengaging from nonvalid cues. This may also be linked with the establishment of the IOR. Indeed, the further the target is from the cue in the supraliminal condition, the faster the reaction time (Bennett & Pratt, 2001). The IOR may facilitate re-engagement towards another localization.
Our study also showed that poker players displayed a higher decrease of the benefit index in the supraliminal condition compared to controls. These results suggest that the IOR of poker players is enhanced compared to that of controls. This type of result has already been observed in a population of cannabis users (Vivas, Estevez, Moreno, Panagis, & Flores, 2012) (see (Colzato & Hommel, 2009) for opposite results) and cocaine users after taking a dose of dextroamphetamine (Fillmore, Rush, & Abroms, 2005). The enhanced IOR found in these populations suggests that the IOR can be improved by the ingestion of drugs, such as cannabis. Nevertheless, the mechanisms are still not understood (Vivas et al., 2012). Other studies, gathered in a systematic review, have assessed the effect of alcohol, ecstasy, and hallucinogens. Nonetheless, only the absence of a difference or a lower IOR was evidenced (Olthuis & Klein, 2012). As for drugs use, repeated poker practice may also impact the IOR. Conversely, an optimal IOR may be helpful to win a poker game, and those with a good IOR may tend to continue playing poker because of this advantage.
Interestingly, a trend towards an enhanced IOR has also been found in a group of patients with Asperger syndrome (Rinehart, Bradshaw, Moss, Brereton, & Tonge, 2008) but not in patients with autism spectrum disorders (Antezana, Mosner, Troiani, & Yerys, 2016; Lin, Miao, & Zhang, 2020). In patients with Asperger syndrome, this result was linked to the fact that this population displays improved attentional visual search efficiency. Thus, it would be interesting to see if poker players also display a better visual search efficiency or if an enhanced IOR is only present when social cues are presented.
As the IOR seems to be helpful not to process a localization that has already been assessed (Klein, 2000; Satel, Wilson, & Klein, 2019), this enhancement may be a sign of a better attentional system in poker players. Indeed, diminishing the refreshment rate of information permits increasing the amount of information that can be processed, which may be very helpful in the practice of poker. Thus, this enhanced IOR may reflect attention particularities linked with the practice of live poker, in which social information needs to be processed efficiently.
Moreover, this research did not evidence any particular engagement or disengagement attention abilities toward specific emotions. This is contrary to our hypothesis that regular nondisordered poker players would display higher information attention abilities compared to controls for the types of emotion that may be useful is the context of poker, i.e., joy, sadness, and surprise. Indeed, the processing efficiency hypothesis (Calvo, Avero, & Lundqvist, 2006; Gordillo León, Mestas Hernández, Pérez Nieto, & Arana Martínez, 2021) postulates that negative stimuli (especially fear) is processed more efficiently than neutral or positive stimuli to react more quickly. For example, in studies regarding spatial attention, fear is the main emotion that is known to automatically capture attention (Bannerman, Milders, de Gelder, & Sahraie, 2009). This hypothesis would explain why lower attentional resources are needed to identify fear and therefore increase reaction speed (Calvo et al., 2006). Anger can be included in this theory as well (Pinkham, Griffin, Baron, Sasson, & Gur, 2010). Nevertheless, our study did not show a significant advantage of any kind of facial cues compared to neutral facial cues.
Finally, beyond social information processing assed by neuropsychological tasks, higher self-reported interpersonal relationships quality has been cited as a predictor of good poker abilities (Schiavella et al., 2018). Interestingly, lower level of interpersonal relationships quality was linked with higher level of addiction in poker players (Schiavella et al., 2018), and an alteration of conscious social information processing was also evidenced in a group of patients with gambling disorder compared to controls (Kornreich et al., 2016). Therefore, higher social cognition abilities may be a marker of good skills in poker while lower social cognition abilities may be a feature of addictive disorders. It is thus important to confront results obtained with samples of poker players with and without a gambling disorder to identify what cognitive processes are linked to a repeated but non pathological poker practice to those related to the addictive process. This may provide useful knowledge to clinicians and researchers specialised in gambling and gambling disorder, both for improving theoretical models of addictive behaviours at various stages of the addictive process, such as in the I-PACE (Interaction of Person-Affect-Cognition-Execution) model (Brand et al., 2019), and for a use in therapy with poker players with a gambling disorder. Indeed, with reference to the present study, it could be possible to explain to patients with a poker-related gambling disorder that the repeated practice of poker may not lead to the development of specific emotion identification strategies for example.
Nevertheless, we did not ask to poker players participants whether they actively tried to practice various skill elements (including reading facial expressions from opponents for example) in the game. Poker-related social skills improvement can also be reached with the help of other players (social learning) (Talberg, 2019). Finally, several social-related personality traits may have been pre-existent in certain participants, independently of the practice of poker, and have influenced the social allocation of attention examined in this study. As a consequence, the effects highlighted in this study may come from either implicit learning (via mere exposure to poker), deliberate learning (via active practice of social skills in the game) or pre-existing social abilities (e.g. trait sociability, extraversion). Future studies on this topic should include information on mere exposure versus deliberate practice, as well as social-related personality traits, in data collected to be able to distinguish such effects one from each other.
Strengths and limitations
The main strength of our research is that we investigated an understudied population with an attentional task designed to assess engagement and disengagement biases towards social-emotional information. One of the limitations of this study is that we did not follow our group over time to observe if their cognitive functioning impacted the onset of addiction or if their attentional particularities changed over time and practice. Additionally, results obtained in this study only concern live poker players. Indeed, the observation of facial expressions and nonverbal information is lacking in online poker. Moreover, there was no women included which do not permit to generalize our results to both genders. Finally, we did not control for nicotine use before the experiment, despite its dopaminergic effect.
Conclusions
In conclusion, our study showed particularities in social information processing associated with the regular practice of poker. We found a higher IOR in the poker player group compared to the CG. Further research regarding the establishment of the IOR in various forms of gambling, specifically what can cause its enhancement, needs to be performed. Indeed, it would be interesting to know if this enhancement of the IOR also appears in nonsocial cues. Nevertheless, more studies are needed to explore the directionality of the link. Indeed, this study cannot conclude if poker practice led to these changes or if these particularities were displayed by the individuals before practising poker and contributed to the experimentation and maintenance of poker practice, given the advantage drawn from this specificity. Interestingly, poker players may practice their social information attention abilities on other people in everyday life, but also specifically on other poker players, who may be bluffing and/or hiding their own social information. On the contrary, the non-players would practice these specific abilities only on everyday life situations, when people are not attempting to conceal the expression of their emotions most of the time. This specific feature of poker is an important particularity that have to be taken into account when assessing and discussing poker players' social information processing abilities. Moreover, it would be interesting to explore whether individuals with gambling disorder show a different response pattern that could be a marker between controlled and excessive practice.
Funding sources
The PERHAPS study was funded by a grant from CHU Nantes (Internal call for tenders; RC14_0036). The funder had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Authors' contribution
EH: collected data; detected any cause for invalid responses in all cognitive tasks; analysis and interpretation of the data; and wrote the first draft of the manuscript. MGB: obtained funding; study supervision; and inclusion of participants. ET and JL: collected data and detected any cause for invalid responses in all cognitive tasks. ML and BP: performed statistical analysis. GCB: study concept and design; obtained funding; analysis and interpretation of the data; and study supervision. All authors had full access to all data and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors provided feedback on the first draft of the manuscript and approved the final manuscript.
Conflicts of interest
MGB, JL, ET, EH and GCB declare that CHU Nantes has received funding from the gambling industry (FDJ and PMU) in the form of a philanthropic sponsorship (donations that do not assign a purpose of use). This funding had no influence on the present work, and scientific independence towards gambling industry operators is warranted. There were no publishing constraints. ML and BP declare that they have no conflicts of interest.
Acknowledgements
This research was conducted based on the initiative of and coordinated by the UIC Psychiatrie et Santé Mentale of CHU Nantes, which sponsored this study.
References
American Psychiatry Association (2013). DSM-5: Diagnostic and statistical manual of mental disorders. Washington, DC: American Psychiatric Publishing.
Antezana, L., Mosner, M. G., Troiani, V., & Yerys, B. E. (2016). Social-emotional inhibition of return in children with autism spectrum disorder versus typical development. Journal of Autism and Developmental Disorders, 46(4), 1236–1246. http://doi.org/10.1007/s10803-015-2661-9.
Bannerman, R. L., Milders, M., de Gelder, B., & Sahraie, A. (2009). Orienting to threat: Faster localization of fearful facial expressions and body postures revealed by saccadic eye movements. Proceedings of the Royal Society B: Biological Sciences, 276(1662), 1635–1641. http://doi.org/10.1098/rspb.2008.1744.
Beck, A. T., Steer, R. A., & Carbin, M. G. (1988). Psychometric properties of the beck depression inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8(1), 77–100. https://doi.org/10.1016/0272-7358(88)90050-5.
Bennett, P. J., & Pratt, J. (2001). The spatial distribution of inhibition of return. Psychological Science, 12(1), 76–80. http://doi.org/10.1111/1467-9280.00313.
Billings, D., Davidson, A., Schaeffer, J., & Szafron, D. (2002). The challenge of poker. Artificial Intelligence, 134(1), 201–240. http://doi.org/10.1016/S0004-3702(01)00130-8.
Bjerg, O. (2010). Problem gambling in poker: Money, rationality and control in a skill-based social game. International Gambling Studies, 10(3), 239–254. https://doi.org/10.1080/14459795.2010.520330.
Bradley, B. P., Mogg, K., White, J., Groom, C., & De Bono, J. (1999). Attentional bias for emotional faces in generalized anxiety disorder. British Journal of Clinical Psychology, 38(3), 267–278. http://doi.org/10.1348/014466599162845.
Brand, M., Wegmann, E., Stark, R., Müller, A., Wölfling, K., Robbins, T. W., & Potenza, M. N. (2019). The interaction of person-affect-cognition-execution (I-PACE) model for addictive behaviors: Update, generalization to addictive behaviors beyond internet-use disorders, and specification of the process character of addictive behaviors. Neuroscience and Biobehavioral Reviews, 104, 1–10. http://doi.org/10.1016/j.neubiorev.2019.06.032.
Calvo, M. G., Avero, P., & Lundqvist, D. (2006). Facilitated detection of angry faces: Initial orienting and processing efficiency. Cognition & Emotion, 20(6), 785–811. http://doi.org/10.1080/02699930500465224.
Carter, R. McK., Bowling, D. L., Reeck, C., & Huettel, S. A. (2012). A distinct role of the temporal-parietal junction in predicting socially guided decisions. Science (New York, N.Y.), 337(6090), 109–111. http://doi.org/10.1126/science.1219681.
Colzato, L. S., & Hommel, B. (2009). Recreational use of cocaine eliminates inhibition of return. Neuropsychology, 23(1), 125–129. http://doi.org/10.1037/a0013821.
Croson, R., Fishman, P., & Pope, D. G. (2008). Poker superstars: Skill or luck? Chance, 21(4), 25–28. http://doi.org/10.1007/s00144-008-0036-0.
DeDonno, M. A., & Detterman, D. K. (2008). Poker is a skill. Gaming Law Review, 12(1), 31–36. http://doi.org/10.1089/glr.2008.12105.
Fiedler, I. C., & Rock, J.-P. (2009). Quantifying skill in games—Theory and empirical evidence for poker. Gaming Law Review and Economics, 13(1), 50–57. http://doi.org/10.1089/glre.2008.13106.
Fillmore, M. T., Rush, C. R., & Abroms, B. D. (2005). D-Amphetamine-Induced enhancement of inhibitory mechanisms involved in visual search. Experimental and Clinical Psychopharmacology, 13(3), 200–208. http://doi.org/10.1037/1064-1297.13.3.200.
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). ‘Mini-Mental state’. A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12(3), 189–198. https://doi.org/10.1016/0022-3956(75)90026-6.
Fox, E., Russo, R., & Dutton, K. (2002). Attentional bias for threat: Evidence for delayed disengagement from emotional faces. Cognition & Emotion, 16(3), 355–379. http://doi.org/10.1080/02699930143000527.
Gerstein, D., Hoffmann, J., Larison, C., Murphy, S., Palmer, A., Chuchro, L., … Sinclair, S. (1999). Gambling impact and behavior study. Report to the national gambling impact study commission. National Gambling Impact Study Commission (NORC).
Gordillo León, F., Mestas Hernández, L., Pérez Nieto, M. Á., & Arana Martínez, J. M. (2021). Detecting emotion faces in a Posner’s spatial cueing task: The adaptive value of surprise. Journal of Cognitive Psychology, 33(1), 38–48. http://doi.org/10.1080/20445911.2020.1862854.
Hayward, D. A., & Ristic, J. (2013). Measuring attention using the posner cuing paradigm: The role of across and within trial target probabilities. Frontiers in Human Neuroscience, 7, 205. http://doi.org/10.3389/fnhum.2013.00205.
Klein, N. (2000). Inhibition of return. Trends in Cognitive Sciences, 4(4), 138–147. http://doi.org/10.1016/s1364-6613(00)01452-2.
Kornreich, C., Saeremans, M., Delwarte, J., Noël, X., Campanella, S., Verbanck, P., … Brevers, D. (2016). Impaired non-verbal emotion processing in pathological gamblers. Psychiatry Research, 236, 125–129. http://doi.org/10.1016/j.psychres.2015.12.020.
Lecrubier, Y., Sheehan, D., Weiller, E., Amorim, P., Bonora, I., Sheehan, K. H., … Dunbar, G. (1997). The Mini international neuropsychiatric interview (MINI). A short diagnostic structured interview: Reliability and validity according to the CIDI. European Psychiatry, 12(5), 224–231. http://doi.org/10.1016/S0924-9338(97)83296-8.
Leonard, C. A., & Williams, R. J. (2015). Characteristics of good poker players. Journal of Gambling Issues, 31(0), 45–68. http://doi.org/10.4309/jgi.2015.31.5.
Lin, Z., Miao, C., & Zhang, Y. (2020). Human electrophysiology reveals delayed but enhanced selection in inhibition of return. Cognition, 205, 104462. http://doi.org/10.1016/j.cognition.2020.104462.
Lundqvist, D., Flykt, A., & Öhman, A. (1998). The Karolinska directed emotional faces – KDEF, CD ROM from department of clinical neuroscience. Karolinska Institutet.
McCormack, A., Shorter, G. W., & Griffiths, M. D. (2014). An empirical study of gender differences in online gambling. Journal of Gambling Studies, 30(1), 71–88. http://doi.org/10.1007/s10899-012-9341-x.
Meyer, G., von Meduna, M., Brosowski, T., & Hayer, T. (2013). Is poker a game of skill or chance? A quasi-experimental study. Journal of Gambling Studies, 29(3), 535–550. http://doi.org/10.1007/s10899-012-9327-8.
Olthuis, J. V., & Klein, R. M. (2012). On the measurement of the effects of alcohol and illicit substances on inhibition of return. Psychopharmacology, 221(4), 541–550. http://doi.org/10.1007/s00213-012-2725-x.
Palomäki, J., Laakasuo, M., Cowley, B. U., & Lappi, O. (2020). Poker as a domain of expertise. Journal of Expertise, 3(2), 22.
Palomäki, J., Yan, J., Modic, D., & Laakasuo, M. (2016). ‘To bluff like a man or fold like a girl?’ – Gender biased deceptive behavior in online poker. Plos One, 11(7), e0157838. http://doi.org/10.1371/journal.pone.0157838.
Pinkham, A. E., Griffin, M., Baron, R., Sasson, N. J., & Gur, R. C. (2010). The face in the crowd effect: Anger superiority when using real faces and multiple identities. Emotion, 10(1), 141–146. http://doi.org/10.1037/a0017387.
Posner, M. (1980). Orienting of attention. The Quarterly Journal of Experimental Psychology: QJEP, 32, 3–25. http://doi.org/10.1080/00335558008248231.
Potter van Loon, R. J. D., van den Assem, M. J., & van Dolder, D. (2015). Beyond chance? The persistence of performance in online poker. Plos One, 10(3), e0115479. http://doi.org/10.1371/journal.pone.0115479.
Rinehart, N. J., Bradshaw, J. L., Moss, S. A., Brereton, A. V., & Tonge, B. J. (2008). Brief report: Inhibition of return in young people with autism and asperger’s disorder. Autism: The International Journal of Research and Practice, 12(3), 249–260. http://doi.org/10.1177/1362361307088754.
Samuel, A. G., & Kat, D. (2003). Inhibition of return: A graphical meta-analysis of its time course and an empirical test of its temporal and spatial properties. Psychonomic Bulletin & Review, 10(4), 897–906. http://doi.org/10.3758/BF03196550.
Satel, J., Wilson, N. R., & Klein, R. M. (2019). What neuroscientific studies tell us about inhibition of return. Vision, 3(4), 58. http://doi.org/10.3390/vision3040058.
Schiavella, M., Pelagatti, M., Westin, J., Lepore, G., & Cherubini, P. (2018). Profiling online poker players: Are executive functions correlated with poker ability and problem gambling? Journal of Gambling Studies, 34(3), 823–851. http://doi.org/10.1007/s10899-017-9741-z.
Schlicht, E. J., Shimojo, S., Camerer, C. F., Battaglia, P., & Nakayama, K. (2010). Human wagering behavior depends on opponents’ faces. Plos One, 5(7), e11663. http://doi.org/10.1371/journal.pone.0011663.
Snijders, T. A. B., & Bosker, R. J. (1994). Modeled variance in two-level models. Sociological Methods & Research, 22(3), 342–363. http://doi.org/10.1177/0049124194022003004.
Spielberger, C. D. (1999). Manual for the state-trait anger expression inventory-2. Odessa, FL: Psychological Assessment Ressources.
Talberg, O. N. (2019). Learning poker in different communities of practice: A qualitative analysis of poker players? Learning processes and the norms in different learning communities. Journal of Gambling Issues, 42(42), 8–41. http://doi.org/10.4309/jgi.2019.42.2.
Tang, Y., Li, Y., Zhuo, K., Wang, Y., Liao, L., Song, Z., … Liu, D. (2015). Neural correlates of the preserved inhibition of return in schizophrenia. Plos One, e0119521. http://doi.org/10.1371/journal.pone.0119521.
Vivas, A. B., Estevez, A. F., Moreno, M., Panagis, G., & Flores, P. (2012). Use of cannabis enhances attentional inhibition. Human Psychopharmacology: Clinical and Experimental, 27(5), 464–469. http://doi.org/10.1002/hup.2248.
