Neuropsychological performance, impulsivity, and comorbid psychiatric illness in patients with pathological gambling undergoing treatment at the CORE inpatient treatment center.

Introduction: Pathologic gambling is a disorder with features that implicate abnormal functioning in brain regions involved in addiction, mood, anxiety, and impulse control disorders. Our goal was to examine brain function with neurocognitive tasks that target these brain regions in patients with pathologic gambling.

Methods: Patients were evaluated for comorbid psychiatric disorders, impulsivity, and performance on reversal-learning and reward-based decision-making cognitive tasks.

Results: Patients had higher impulsivity scores and significant deficits on both cognitive tasks compared with controls. All subjects also had comorbid psychiatric disorders, including mood, anxiety, psychotic, and substance abuse.

Conclusion: The cognitive deficits and impulsivity are consistent with abnormal activity in orbitofrontal-limbic networks. The high level of comorbidity is consistent with the overall severity in these inpatients undergoing treatment, and adds weight to the concept of a fundamental abnormality in this network.

Key Words: pathologic gambling, addiction, impulse control, orbitofrontal cortex, comorbidity

**********

Pathologic gambling (PG) is an impulse control disorder with recurrent maladaptive gambling behavior as its main feature. (1) It is estimated to affect up to 15 million Americans, (2) cost 5 billion dollars annually in direct and indirect costs to society, (3) and has become a significant problem as legalized gambling proliferates. (4) While the aforementioned epidemiologic data indicate that proliferation of legalized gambling is impacting the prevalence of this disorder, we know very little about the underlying neurobiology of PG. The essential components of addiction and dependence to a substance are present, if one considers the act of gambling as the exogenous substance. Previous work shows that gamblers have characteristics that are very similar to those addicted to other substances. (5) Previous studies of addictive behavior implicate abnormal functioning in a network of brain regions involved in reward-based decision-making (RBDM) that include the orbital and medial prefrontal cortex. (6-9) The Iowa Gambling Task has been shown to discriminate patients with lesions of the ventromedial prefrontal cortex (VMPFC) from normal subjects, (10) as well as activate this same region of the prefrontal cortex (PFC) during functional neuroimaging. (11) A previous study used this task to examine RBDM in PG patients, and found that patients had deficits in performance compared with controls. (12) This study highlighted the involvement of ventromedial PFC in RBDM, and abnormalities of this process in PG patients. This report also points out that the VMPFC is involved in other disorders, such as obsessive-compulsive disorder and addiction. There have been many reports which find that the VMPFC, orbitofrontal cortex, and ventral anterior cingulate are commonly affected in other psychiatric disorders, many of which are often comorbid in persons with pathologic gambling (see 13, 14 for review), complicating both evaluation and treatment. These include mood disorders, such as major depression (15-16) and bipolar disorder, (15, 17, 18) as well as anxiety disorders like obsessive-compulsive disorder, (19) post-traumatic stress disorder, (20) and panic disorder. (21) Furthermore, these same brain regions have been implicated in attention-deficit hyperactivity disorder, which has significant components of impulse control deficits. (22) Finally, two studies done with PG patients find abnormalities in the ventromedial PFC with an attention task (23) and also in the orbitofrontal cortex and ventral anterior cingulate cortex when viewing gambling scenes. (24)

It is therefore likely that patients with PG have deficits in the function of these brain regions, as there are components of impulsivity, addiction, and reward-based decision making in this disorder. There are at least two tasks that target the reward network that can be used to examine this network's function. The reversal learning (RL) task is a neuropsychological task that targets the reward system of the brain, as shown in a functional neuroimaging study of normal subjects. (25) Another task known to target this network is the Iowa Gambling Task mentioned above. (10) Our hypothesis was that patients diagnosed with PG will: 1.) have high levels of comorbid psychiatric diagnoses, 2.) have deficient performance on the two reward-based tasks compared with controls, and 3.) report more impulsivity than control subjects. The rationale for this study was to better understand the underlying neurobiological abnormalities in the reward-based network associated with the cognitive and psychiatric pathology found in patients with pathologic gambling. We predicted that our population would have significant comorbid mental illness, consistent with previous studies. (26-28) Furthermore, we expected that gamblers would have more impulsivity than controls, and have greater difficulty with the two neuropsychological tasks. The preliminary results presented here represent an initial endeavor with a small number of patients, and further studies with more refined cognitive tasks and functional imaging are required to build upon these data to provide supporting evidence. However, these initial findings support the hypotheses of high comorbidity, deficient performance on reward-based tasks, and high impulsivity in this population of patients receiving inpatient treatment for PG.

Experimental Design and Methods

This research study was approved by the LSUHSC Institutional Review Board before initiation. All subjects were voluntary and signed informed consent forms before starting the study. A total of 41 subjects participated, including 23 controls (of whom three were excluded due to mental illness) and 18 patients. Controls were recruited from friends and unrelated family members (in-laws) of the patients to control for differences in socioeconomic status (SES). The majority of the research was conducted at CORE (Center of Recovery), a unique inpatient pathologic gambling rehabilitation center located in Shreveport, LA. All subjects were screened and tested at the CORE facility, and patients in the study were residents of CORE. Patients with pathologic gambling were chosen from CORE facility clients who had been in the inpatient treatment facility for at least 10 days. The aim was to minimize the effect of confounds from acutely admitted subjects on the project, as well as to minimize any detrimental effect of the psychological testing on the newly arrived clients. This is because one of the neuropsychological tasks can actually be considered a "gambling game" (see below). Urine drug screens were performed on all subjects to evaluate for the common drugs of abuse, including marijuana, cocaine, methamphetamine, benzodiazepines, and barbiturates. All subjects included in the study had negative urine drug screens at the time of examination. Subjects with major medical or neurologic illnesses that precluded their involvement (eg, blindness) were also excluded. Controls with major DSM-IV psychiatric diagnoses (including substance abuse disorders) as determined with the MINI (see below) were also excluded. Nicotine and tobacco use was present in both groups, but was not exclusory. All participants were reimbursed with a small gift certificate to a local retail store to decrease the incidence of monetary incentives to problem gamblers.

Rating Scales and Instruments

All screening instruments were administered by a single coinvestigator trained in their delivery, and all screening data was collected in a private office with minimal distractions. The Mini International Neuropsychiatric Interview (MINI) (29) was used to screen for the presence of comorbid DSM-IV Axis 1 psychiatric disorders. The South Oaks Gambling Screen (SOGS) (30) was used to establish the diagnoses of pathologic gambling. A score of 5 or greater defines the threshold for pathologic gambling. Subjects also filled out an employment and education questionnaire to help determine their socioeconomic status (SES). The SES was determined by assigning a numeric value ranging from 1 to 8 to the level of education and employment (higher numbers indicating higher level of education or employment), taking the mean of these two values for each subject, and then determining the grand mean for both groups. Impulsivity was examined with the Barratt Impulsivity Scale (BIS). (31) The BIS is a self-report instrument, and is composed of three subscales: Inp, Im, and Ic. The Inp subscale relates to nonplanning, exemplified by an inability to plan ahead, not thinking things through carefully, and caring only for the present. The Im subscale is a measure of motor impulsiveness, represented by acting on impulse or without thinking. The Ic subscale relates to making quick cognitive decisions, and subjects with deficits on this subscale often demonstrate poor attention and make hasty choices. Significant differences between patients and controls (and males versus females for both groups) were assessed with the Student's t test, as previous studies have shown that results with the BIS are normally distributed. (31)

Cognitive Tasks

The cognitive tasks were written in E-Prime[c] (Pittsburgh, PA). Tasks were delivered to individual subjects in a quiet room with minimal distractions. The workstation was situated on a desk a standard distance away from the subject. The subject was seated in a comfortable chair, adjusted to the appropriate height to view the monitor and perform the task in comfort. Keyboard keys were marked to indicate the required task response. Task order was randomized.

Reversal Learning (RL) Task

See Figure 1 for a schematic representation of the RL task. In this task, four geometric shapes were shown (circle, square, diamond, triangle), each a different color (red, green, blue, yellow), and in four locations (left, left-middle, right-middle, right). Thus there were three "rules": shape, color, and location. The subjects were instructed to press one of four buttons on the keyboard (D, G, J, L), marked for visibility. They either heard a buzz indicating a wrong answer, or a ding indicating a correct answer. The symbols varied such that the color and location changed in a random fashion after every button press by the subject. The goal was to determine which rule from a set of rules was currently correct. For example, if the rule was "square", then the correct choice was to press the button representing the square, irrespective of where the square was, or what color. After ten correct choices, the rule changed, and the subject had to then determine the new rule. This switch from one rule to another could be either an intradimensional shift (IS) such as from green to blue, or an extradimensional shift (ES) such as green to triangle. This continued for a total of 150 choices. Behavioral data were collected for later analyses. Initially during the task the choices are made randomly, but subjects typically learn how to choose and reduce the number of choices required to successfully complete a trial of 10 sequential correct choices. Subjects were assessed on the number of completed trials (eg, how often they identified a rule), and the amount of time and number of choices that it took to complete a trial. These were assessed for both IS and ES trials. Significant differences between patients and controls were assessed using analysis of variance.

[FIGURE 1 OMITTED]

Iowa Gambling Task

Our version of this task is based upon the original version used to discriminate patients with lesions of the ventromedial prefrontal cortex from normal subjects. (10) On each trial, four symbols are displayed on a screen for one second and then flicker, indicating that a choice can be made (Fig. 2). The subject selects a symbol by pressing a key on the keyboard. After the subject chooses a symbol, feedback is given in the form of two numbers displayed on the computer monitor one above the other. The top number represents the amount of points won or lost on the last trial (reward), and the bottom number the points accumulated over time (sum). The goal is to learn which two symbols earn the most points in the long run to win the game, defined by keeping at least the 500 points that the subject begins with. Unknown to the subjects, two symbols have large rewards but larger penalties, and will lose if chosen consistently (long-term loss). The other two game symbols have small rewards, but even smaller penalties such that in the long run the subject will win if these symbols are chosen consistently (long-term gain). The frequency of wins and losses for the four symbols vary, and symbol valence changes from run-to-run in a pseudorandom fashion. There were 150 trials per experiment. Subjects were familiarized with the layout before the experiment began, but as the task involved learning, no extensive training was done. Behavioral data were collected for later analyses, and assessed for the number of choices made from the correct and incorrectly identified symbols. Significant differences between patients and controls were assessed using analysis of variance.

[FIGURE 2 OMITTED]

To minimize the risk of activating symptoms of pathologic gambling in the patient group, the features of the gambling task (which is essentially a gambling game in its original form, see 10) were kept as bland as possible. The task was composed of white symbols on a black background, instead of using images of cards or colored symbols. Only numerical feedback (no verbal responses or other sound) was given on the selected choice and the reward and sum were presented as "points" and not "dollars."

Results

Education Level and Demographics

There was no significant difference on the SES between groups using ANOVA (patients: 3.9 [+ or -] 1.5, controls: 3.1 [+ or -] 1.4, P = 0.12). There was also no significant difference in age using ANOVA (patients: 45 [+ or -] 13, controls: 41 [+ or -] 16, P = 0.39), or gender (Chi-squared distribution, P = 0.16).

Rating Scales

The Table shows the results of the MINI. All patients had a comorbid psychiatric disorder. Of these, 15 (83%) had a mood disorder, which included major depressive episode (n = 3), and bipolar disorder, either type I or type II (n = 12). One of the latter patients also met criteria for dysthymia. Most of the patients (n = 17, 94%) had an anxiety disorder, and many of these met criteria for more than one type of anxiety disorder. This included panic disorder with or without agoraphobia (n = 7), social phobia (n = 9), obsessive-compulsive disorder (n = 9), post-traumatic stress disorder (n = 3), and generalized anxiety disorder (n = 9). These numbers add up to greater than 17 due to patients having more than one anxiety disorder. Also, six patients (33%) had a history of having psychotic symptoms, but five of these were in the context of previous bipolar manic episodes, and one was reported to be substance induced. Only five patients (22%) were diagnosed with alcohol abuse (3) or dependence (2) in the past year, and the latter two subjects also had a history of substance abuse in the past year. Nicotine and tobacco use information was collected with the demographic data. In the patient group, 9 subjects (50%) smoked, and 4 (25%) of the controls smoked. In the control group, 7 of 23 were diagnosed with psychiatric disorders or symptoms. Of these, three were severe enough to warrant exclusion: two had bipolar disorder (one with psychotic features), and one had major depressive disorder and panic disorder with agoraphobia. These subjects were excluded from the study. Four controls mentioned mood or anxiety symptoms that were recorded on the MINI, but did not warrant diagnosis based on DSM-IV criteria, and thus were not excluded from the study.

By definition, patients entering the CORE treatment program were severe pathologic gamblers who needed an inpatient treatment regimen. Thus patients had significantly higher scores (14.3 [+ or -] 2.91) on the SOGS as compared with the controls 90.22 [+ or -] 0.073).

The BIS results are shown in Figure 3. This graph shows the result for each of the subscales, and the total of all three subscales, for both patients and controls. Higher scores indicate more reported impulsivity. Patients reported significantly more impulsivity than controls in all three domains, including motor, cognitive, and nonplanning, as well as the overall score (Inp: P = 1.85 X [10.sup.-10], t = 8.85; Im: P = 5.12 X [10.sup.-7], t = 6.13; Ic: P = 3.69 X [10.sup.-09], t = 7.80; Total: P = 8.47 X [10.sup.-12], t = 10.00). Furthermore, the BIS subscales Inp and Ic were positively correlated with the severity of the SOGS rating scale data from patients. As there have been reports of gender differences in disorders such as ADHD, especially involving impulsivity measures, (32,33) the BIS data was examined based on gender. As both these studies find more impulsivity in males, a one-tailed t test was thought to be indicated, however there were findings of changes in the opposite direction on one scale (Ic--controls) and so a two-tailed t test was used. While there were no significant differences found, there were trends toward higher impulsivity reports in the males in the patient group on all scales, as well as the Im scale of the control group in females. The distribution of the SOGS and the BIS subscales were sufficiently normal to warrant linear regression. Figure 4 shows the relationship of Inp (nonplanning impulsiveness) to the SOGS. Both the Inp (r = 0.453, P = 0.030) and Ic subscales (r = 0.431, P = 0.037) were positively correlated with the SOGS, as was the total BIS score (r = 0.434, P = 0.036); however, only the Inp data are shown graphically. The Im subscale was not significantly correlated with SOGS (r = 0.248, P = 0.161).

[FIGURE 3 OMITTED]

Cognitive Tasks

Results from the RL task are shown in Figure 5. The RL task performance measure is separated into the ES and IS components for patients and controls. Panel A (left) shows the average number of choices needed to complete a successful trial, panel B (middle) shows the average time required to complete a successful trial, and panel C (right) shows the average number of successful trials completed. Because there are three sets of rules (color, shape, location) and four rules in each set, and because the switch from one rule to another was random, the frequency of ES switches was higher than IS switches. Patients required significantly more choices (P = 0.0036, F = 3.91) as well as more time (P = 0.024, F = 3.91) on ES trials compared with controls, and this resulted in a significant decrease in the number of successful trials completed in patients (P = 0.008, F = 4.10). There were no significant statistical differences with IS trials in any of the three measures.

[FIGURE 4 OMITTED]

Results from our version of the Iowa Gambling Task are shown in Figure 6. The time course of the task was divided up into 5 segments, and results averaged over each portion for both patients and controls. These results show that patients did significantly worse than controls toward the end of the task in segment D (P = 0.015, F = 3.967), but there were no significant differences in other parts of the task.

Discussion

We used the BIS (31) to examine impulsivity in this study. The preliminary results presented here indicate that pathologic gamblers may be more impulsive compared with controls. This is consistent with the categorization of PG as an impulse control disorder, as well as previous reports which find higher impulsivity in this population. Impulse control problems are a well known component in attention deficit hyperactivity disorder (ADHD), and significant attention problems consistent with attention deficit disorders have been reported in patients with PG. (34) One prospective longitudinal study that examined impulsivity in adolescents found that problem gambling was predicted by higher impulsiveness. (35) In the present study, patient self-report scores on all three BIS subscales were significantly higher, which may be indicative of deep-seated impulsivity problems, consistent with the degree of pathologic gambling in this population. This is bolstered by a previous report that used principal component analysis of multiple scales to examine problem gamblers, and found that the degree of impulsivity was closely related to the severity of pathologic gambling (36) consistent with the results presented here (see Fig. 4). It is important to note that ADHD was not specifically evaluated in the present study, thus it is not clear whether any patients actually met criteria for comorbid adult ADHD. However, it is worth noting that the BIS does look at many of the same behavioral elements as instruments used in the diagnosis of ADHD.

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

While there was not a significant difference in the frequency of males and females between the patients and controls, 56% of the patients were male compared with 35% of the controls. This was enough of a difference to be of concern in regard to impulsivity, as there have been gender differences reported in ADHD, with males having higher rates of the disorder as well as higher rates of impulsive symptoms. (32,33) However, we found no significant differences between males and females in either the three subscales or the total BIS score for either group (data not shown).

Two cognitive tasks were used in this study (RL and RBDM) as these tasks have been shown to target the ventromedial prefrontal and orbitofrontal cortex, (25) regions of the brain thought to be involved in reward-based decision-making (37) and addiction. (6) Our results show that patients with PG have significant deficits on the RL task, including fewer successful trials and taking more time and choices for successful trials, compared with controls. These results illustrate that pathologic gamblers have trouble switching from task to task, and are consistent with previous research showing that patients with ADHD have abnormal performance measures on a reversal-learning task. (38)

Patients with PG had slight but significant deficits on our version of the Iowa Gambling Task compared with controls. The lack of strong findings with this task in our population may be a result of making the task bland to prevent over stimulation of sensitive inpatients in the early stages of recovery. However, the presence of weak but significant findings with this task in pathologic gamblers are consistent with previous studies that indicate patients with PG have abnormal brain activity in the VMPFC in response to a Stroop task of attention (23) as well as the orbitofrontal cortex and ventral anterior cingulate when viewing gambling scenes. (24) One previous study used the original version of the Iowa Gambling Task in PG patients. (12) The patients examined in that study had no comorbid axis I illnesses, and had SOGS scores of 15.8 [+ or -] 3.6, very similar to our results on the SOGS (14.3 [+ or -] 2.91), and they found robust differences on the Iowa Gambling Task in patients with PG. These more robust results are most likely due to the use of the original version of the Iowa Gambling Task in their study, while the one used in the present study was made intentionally bland. Another major difference between the previous study (12) and the present study was the presence of comorbid psychiatric disorders. However, it is unlikely that this caused the differences between the two studies, as other reports indicate that deficits on the Iowa Gambling Task can be associated with other addictive disorders, such as substance abuse. (6,39)

In this preliminary study, we provide evidence that indicates our inpatient population with pathologic gambling disorder had significant comorbid mental illness. We expected a large degree of comorbidity in this population, as previous studies have reported this finding. One recent study recruited 30 subjects with compulsive gambling from the community and examined comorbidity using the Diagnostic Interview for DSM disorders. (26) They found that 60% had a current comorbid mental illness. This study used a SOGS > 5 as an inclusory criterion, however the mean SOGS score (and thus the level of severity) for the group of patients was not reported. Another study that used the Structured Clinical Interview for DSM disorders (SCID) examined 69 outpatients enrolled in a specialized treatment program. (27) They found 62% comorbidity, but this included personality disorders along with Axis I disorders. The overall SOGS was 13.3 in those with a comorbid psychiatric disorder, and 11.7 in those without any comorbidity. Only a small percentage was actually determined to have mood (8.7%) or anxiety (4.3%) disorders. A third study used the SCID to examine 40 outpatients with pathologic gambling, and found 38% had a current mood disorder, and 23% had a current anxiety disorder. (28) The patients in this study also had a mean SOGS score of 13.7. Our study reported higher comorbid axis I disorders than these previous reports. This may be due to differences in the rating instrument, methodological issues, or the patient population. The pathologic gamblers evaluated in our study had significant gambling problems, as indicated by their requiring inpatient treatment and the mean SOGS score of 14.3 (range 10-18).

Although we show that patients with PG have abnormal responses on two cognitive tasks, this is confounded by the high degree of comorbid mental illness. Performance on a neuropsychological task is likely to be affected adversely by comorbid psychiatric disorders, especially the mood disorders, (40) and the high degree of anxiety may have affected performance as well. These results require replication in a population of pathologic gamblers who are carefully screened for comorbid mental illness. Then again, this latter type of study would also have confounds, as such a population would not be representative of the typical patient with pathologic gambling, since there normally is a high degree of comorbidity. (26-28) This, in and of itself, is suggestive of the potential for pathologic gambling, impulsivity, and other mood and anxiety disorders mentioned earlier to all occur as components of a single multifaceted underlying neurobiological abnormality. Although such a concept is speculative at best at the present time, the overlap in brain regional abnormalities found in mood, (15-18) anxiety, (19-21) addiction, (6-9) and ADHD (22) provide support for such a conceptualization. If such a model one day proves to be true, it may affect the way we think about disease processes and distinct segregation of disorders. These findings are also important in the matters of diagnosis and treatment for patients with such comorbid disorders. Careful diagnostic methods are indicated as a simple diagnosis of PG will miss major psychiatric disorders in some patients. Treatments for addiction, craving, impulse control, mood and anxiety symptoms must all be considered, and medications or therapies which cross the boundaries between illnesses would likely be more appropriate for such a population of patients.

There are multiple drawbacks to the present study. This work is preliminary in a small number of patients, and requires further work in a larger population to provide support for the data presented here. As mentioned above, PG patients with and without comorbid mental illness need to be examined. Furthermore, controls with comorbid mental illness also need to be evaluated to determine the performance of that population with these measures. The issue of comorbid substance abuse needs to be further investigated, as the rate of alcohol and illicit drug abuse in our patient population was surprisingly low. While patients who come to CORE for help can have comorbid substance abuse problems, it is primarily for PG. There is no medical staff to handle detoxification from alcohol or other substances, thus the environment may be selective for subjects without other substance abuse issues. Cigarette smoking was tracked, but future work will be needed to study the effect on smoking in this population, as the time from last smoking to cognitive testing is an important factor, and was not measured in this study. Another issue is the effect of different psychiatric diagnoses on the various results. The sample size for this study was too small to make relevant comparisons, eg, patients with PG and generalized anxiety versus patients with PG and major depressive disorder. Moreover, measuring psychiatric symptom severity with ratings scales would be important to assess their effect on the results. This preliminary study did not include extensive neuropsychological testing, but future work should examine the effect of IQ on the cognitive tests.

Conclusion

This small study provides further evidence that patients with PG have significant problems with psychiatric comorbidity as well as higher degrees of impulsivity and problems performing cognitive tasks which target a network of brain regions involved in RBDM, further implicating this network in the pathophysiology of PG. Further studies are necessary to delineate whether specific components are involved in isolated PG versus comorbid disorders, and larger studies where different combinations of diagnoses can be evaluated.

Acknowledgments

This work was supported by a grant from the Louisiana Association on Compulsive Gambling. Thanks go to Rachel Alcorn for her excellent technical assistance with this project.

References

1. American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 4th edition, text revision. Washington, DC: American-Psychiatric Press, 2000.

2. Unwin BK, Davis MK, DeLeeuw JB. Pathologic gambling. Am Fam Physician 2000;61:741-749.

3. National Gambling Impact Study. National Gambling Impact Study Commission, Final Report; 1999. Available at:http://govinfo.library.unt.edu/ngisc/index.html.

4. Petry NM, Armentano C. Prevalence, assessment, and treatment of pathological gambling: a review. Psychiatr Serv 1999;50:1021-1027.

5. Jacobs DF. A general theory of addictions. A new theoretical model. J Gambling Behav 1986;2:15-31.

6. Bechara A, Dolan S, Denburg N, et al. Decision making deficits, linked to a dysfunctional ventromedial prefrontal cortex, revealed in alcohol and stimulant abusers. Neuropsychologica 2001;39:376-389.

7. Breiter HC, Rosen BR. Functional magnetic resonance imaging of brain reward circuitry in the human. Ann NY Acad Sci 1999;877:523-547.

8. Drevets WC, Gautier C, Price JC, et al. Amphetamine-induced dopamine release in human ventral striatum correlates with euphoria. Biol Psychiatry 2001;49:81-96.

9. Wexler BE, Gottschalk CH, Fulbright RR, et al. Functional magnetic resonance imaging of cocaine craving. Am J Psychiatry 2001;158:86-95.

10. Bechara A, Damasio AR, Damasio H, et al. Insensitivity to future consequences following damage to human prefrontal cortex. Cognition 1994;50:7-15.

11. Patterson JC, Ungerleider LG, Bandettini PA. Task-independent functional brain activity correlation with skin conductance changes: an fMRI study. Neuroimage 2002;17:1797-806.

12. Cavedini P, Riboldi G, Keller R, et al. Frontal lobe dysfunction in pathological gambling patients. Biol Psychiatry 2002b;51:334-341.

13. Cunningham-Williams RM, Cottler LB, Compton WM. Taking Chances: Problem Gambling and Mental Health--Results from the St. Louis Epidemiologic Catchment Area (ECA) Study. American Journal of Public Health 1998;88:1093-1096.

14. Crockford DN, el-Guebaly N. Psychiatric Comorbidity in Pathological Gambling: A Critical Review. Can J Psychiatry 1998;43:43-50.

15. Drevets WC, Price JL, Simpson JR Jr., et al. Subgenual prefrontal cortex abnormalities in mood disorders. Nature 1997;386:824-827.

16. Elliott R, Rubinsztein JS, Sahakian BJ, et al. The neural basis of moodcongruent processing biases in depression. Arch Gen Psychiatry 2002;59:597-604.

17. Blumberg HP, Leung HC, Skudlarski P, et al. A functional magnetic resonance imaging study of bipolar disorder: state- and trait-related dysfunction in ventral prefrontal cortices. Arch Gen Psychiatry 2003;60:601-609.

18. Blumberg HP, Stern E, Ricketts S, et al. Rostral and orbital prefrontal cortex dysfunction in the manic state of bipolar disorder. Am J Psychiatry 1999;156:1986-1988.

19. Cavedini P, Riboldi G, D'Annucci A, et al. Decision making heterogeneity in obsessive-compulsive disorder: ventromedial prefrontal cortex function predicts different treatment outcomes. Neuropsychologia 2002a;40:205-211.

20. Zubieta JK, Chinitz JA, Lombardi U, Medial frontal cortex involvement in PTSD symptoms: a SPECT study. J Psychiatr Res 1999;33:259-264.

21. Bystritsky A, Pontillo D, Powers M, et al. Functional MRI changes during panic anticipation and imagery exposure. Neuroreport 2001;12:3953-3957.

22. Hesslinger B, Tebartz van Elst L, Thiel T, et al. Frontoorbital volume reductions in adult patients with attention deficit hyperactivity disorder. Neurosci Letters 2002;328:319-321.

23. Potenza MN, Leung HC, Blumberg HP, et al. An FMRI Stroop task study of ventromedial prefrontal cortical function in pathological gamblers. Am J Psychiatry 2003;160:1990-1994.

24. Potenza MN, Steinberg MA, Skudlarski P, et al. Gambling urges in pathological gambling: a functional magnetic resonance imaging study. Arch Gen Psychiatry 2003b;60:828-836.

25. O'Doherty J, Kringelbach ML, Rolls ET, et al. Abstract reward and punishment representations in the human orbitofrontal cortex. Nature Neurosci 2001;4:95-102.

26. Black DW, Moyer T. Clinical features and psychiatric comorbidity of subjects with pathological gambling behavior. Psychiatr Serv 1998;49:1434-1439.

27. Ibanez A, Blanco C, Donahue E, et al. Psychiatric comorbidity in pathological gamblers seeking treatment. Am J Psychiatry 2001;158:1733-1735.

28. Specker S, Carlson G, Edmonson K, et al. Psychopathy in pathological gamblers seeking treatment. J Gambling Studies 1996;12:67-81.

29. Sheehan DV, Lecrubier Y, Harnett-Sheehan K, et al. The Mini International Neuropsychiatric Interview (MINI): the development and validation of a structured psychiatric interview. J Clin Psychiatry 1998;59:22-33.

30. Lesieur HR, Blume SE. The South Oaks Gambling Screen (SOGS): a new instrument for the identification of pathological gamblers. Am J Psychiatry 1987;144:1184-1188.

31. Patton J, Stanford M, Barratt E. Factor Structure of the Barratt Impulsiveness Scale. J Clinical Psychology 1995;51:768-774.

32. Gomez R, Harvey J, Quick C, et al. DSM-IV AD/AD: confirmatory factor models, prevalence, and gender and age differences based on parent and teacher ratings of Australian primary school children. J Child Psychol Psychiatry 1999;40:265-274.

33. Gershon J. A meta-analytic review of gender differences in ADHD. J Atten Disord 2002;5:143-154.

34. Rugle L, Melamed L. Neuropsychological assessment of attention problems in pathological gamblers. J Nerv Ment Dis 1993;181:107-112.

35. Vitaro F, Arseneault L, Tremblay RE. Impulsivity predicts problem gambling in low SES adolescent males. Addiction 1999;94:565-575.

36. Steel Z, Blaszczynski A. Impulsivity, personality disorders and pathological gambling severity. Addiction 1998;3:895-905.

37. Damasio AR. The somatic marker hypothesis and the possible functions of the prefrontal cortex. Phil Trans R Soc Lond B 1996;351:1413-1420.

38. Itami S, Uno H,. Orbitofrontal cortex dysfunction in attention-deficit hyperactivity disorder revealed by reversal and extinction tasks. Neuroreport 2002;13:2453-2457.

39. Grant S, Contoreggi C, London E. Drug abusers show impaired performance in a laboratory test of decision making. Neuropsychologia 2000;38:1180-1187.

40. Elliott R, Sahakian BJ, McKay AP, et al. Neuropsychological impairments in unipolar depression: the influence of perceived failure on subsequent performance. Psychological Med 1996;26:975-89.

James C. Patterson II, MD, PHD, Janet Holland, MS, and Reece Middleton, MS

From the Department of Psychiatry, Louisiana State University Health Sciences Center and the Center of Recovery Treatment, Center for Pathological Gambling, Shreveport, LA.

Reprint requests to James C. Patterson II, LSUHSC--Psychiatry, 1501 Kings Highway, Shreveport, LA, 71103. Email: [e-mail]

This work was supported by a grant from the Louisiana Association on Compulsive Gambling.

Accepted August 26, 2005.

RELATED ARTICLE: Key Points

* Patients with Pathological Gambling Disorder have a high degree of comorbid mental illness in our population.

* Tasks that target the orbitofrontal and ventromedial prefrontal networks were abnormal in this population, compared to controls.

* Because these brain regions are involved in the pathophysiology of mood, anxiety, and addictive disorders, it is impossible to distinguish abnormalities related to pathological gambling versus that related to mood or anxiety disorders.

Table. Results from the Mini International Neuropsychiatric Interview

Subject  Mood  Anxiety  Psychosis  Substance

 1             x
 2       x     x        x (a)      x
 3       x     x
 4       x     x        x (a)
 5       x     x                   x
 6                      x (a)
 7       x     x
 8       x     x
 9       x     x
10       x     x
11       x     x        x (a)      x
12       x     x
13       x     x        x (b)      x
14             x
15       x     x                   x
16       x     x
17       x     x        x (a)
18       x     x
Total    15    17       6          5

(a) History of psychotic symptoms during previous manic episodes.
(b) History of previous substance-induced psychotic symptoms.