Obsessive-Compulsive Disorder and Ventromedial Frontal Lesions: Clinical and Neuropsychological Findings

In a minority of cases, obsessive-compulsive disorder (OCD) is refractory to all pharmacological and psychotherapeutic treatments. Extensive, but mostly uncontrolled, evidence suggests that such cases may respond to various types of psychosurgical operations. Developed in the late 1930s and extensively used in the 1940s, early radical forms of frontal leukotomy offered subjects therapeutic gains but at the price of frequent side effects, especially those related to the frontal lobe syndrome (1). More-refined methods, especially the use of stereotaxy, reduced but did not eliminate the risks of personality change and cognitive malfunction after surgery. Unfortunately, evaluation at this time was mainly restricted to assessment of the psychiatric condition treated by surgery, and little attempt was made to study alterations in other aspects of behavior. Only a minority of outcome studies used formal testing (2).

Today a limited number of such operations are still performed; they involve capsulotomy (3), cingulotomy (4, 5), subcaudate tractotomy (6), or combined orbitomedial/cingulate lesions (7). Until the 1980s modified bimedial frontal leukotomy was performed as well (8). Although efficacy and risks vary slightly over the different procedures (2), reports have confirmed their antiobsessional effect in at least 30% of subjects (4–8). Adverse side effects were rare. However, a few subjects were reported to show adverse personality changes of the frontal lobe type (9). Overall cognitive functioning remained surprisingly unaffected (7, 10, 11). Deficits frequently occurred on the Wisconsin Card Sorting Test, a commonly used measure of the ability to form and shift concepts (10, 11).

However, even with the currently used operations, little effort is made to determine the specific lesion sites mediating antiobsessional effects or possible adverse side effects or neuropsychological deficits. During subcaudate tractotomy, radioactive rods are placed beneath and in front of the caudate nucleus (6). Capsulotomy is performed by producing stereotactic lesions within the anterior limb of the internal capsules (3). Cingulotomy involves stereotactic lesions of the anterior cingulate cortex (4). In none of the published studies of these operations were lesion analyses of the respective subjects performed, although the surgical techniques most probably resulted in individually differing lesion sites and lesion sizes.

The aim of the present study was to elucidate the effects of specific lesion sites in OCD subjects with ventromedial frontal leukotomy in order to anatomically specify the antiobsessional effect and to better define the structure-function relationships of cognitive and behavioral frontal lobe functions. A long-term follow-up of subjects who had undergone ventromedial frontal leukotomy for refractory OCD during the 1970s was undertaken. This approach further offered the advantage of determining the long-term effects of the psychosurgical interventions. Usually, follow-ups of such subjects cover at most a few years, thus posing serious difficulties for the evaluation of the stability of therapeutic effects (1).

METHOD

Subjects

Sixteen subjects who had undergone bilateral ventromedial frontal leukotomy as a last-resort treatment for refractory OCD were evaluated in 1992. The 16 subjects were among 62 whose records had been reviewed. Of this total, current addresses of only 37 could be traced. The remainder had died (N=9), had moved to unknown new locations (N=7), or were excluded because of old age (>85 years, N=3) or presence of an additional psychiatric diagnosis (e.g., schizophrenia, N=6). Of the 37 leukotomized subjects available, 16 agreed to take part in the investigation, which involved 2 days of hospitalization in order to complete tests, interviews, and a magnetic resonance imaging (MRI) scan. Another 13 subjects accepted interviews by telephone and completed questionnaires. Those who objected to these kinds of evaluation (N=8) did so mainly because they did not want to be reminded of the worst time of their lives and feared new exacerbation of symptoms. However, these subjects provided self-ratings of improvement. The leukotomized patients were compared with seven nonleukotomized well comparison subjects (table 1) recruited from hospital inpatients and outpatients who had severe OCD of at least 10 years' duration. After complete description of the study to the subjects, written informed consent was obtained.

Leukotomy

The surgery had been performed by F. Roeder and H. Orthner (departments of neurosurgery and neuropathology at the University of Göttingen). Thermocoagulation lesions were made by stereotaxy with the aid of high-pressure pneumoventriculography. The lesions were always made bilaterally, with 13 to 25 coagulations (mean=18) per hemisphere. The intention was to destroy with each coagulation about 230 mm³ of tissue. Basically, the procedure involved coagulations restricted to the mediobasal angle of the frontal white matter, in order to isolate the ventromedial cortex from its subcortical connections. In some subjects, the leukotomy was extended posteriorly into the white matter of the orbital cortex underlying the caudate nucleus and to the fundus of the striatum. In other subjects, the leukotomy was extended dorsally to the ventromedial lesion in order to isolate the precallosal part of the cingular gyrus. All subjects except two were operated on between 1970 and 1976. Subjects operated on in 1973 and later received lesions extended to the fundus of the striatum, and subjects operated on earlier received ventromedial or dorsally extended lesions. Details of the surgical technique are described in reference 12.

Lesion Groups

In order to investigate the influence of different lesion sites, the 16 leukotomized subjects were divided into three groups according to the main sites of the lesions as determined by current MRI scans. All subjects had ventromedial frontal lesions (Brodmann areas 11/12 and 24/32). Subjects with lesions mainly restricted to these areas constitute group 1 (N=6). Subjects with a posterior extension of the leukotomy also had lesions of the most posterior parts of areas 11/12 and also lesions of the ventral striatum (ventral parts of the putamen, caudate nucleus, and globus pallidus, and nucleus accumbens) and, to a minor degree, of the substantia innominata. This group is referred to as group 2 (N=5). Subjects with dorsally extended lesions, referred to as group 3 (N=5), showed in addition to the ventromedial frontal lesions marked atrophy and necrosis of the dorsolateral convexity (Brodmann areas 6 and 8–10). The lesion characteristics of all subjects are described in table 2. Figure 1 illustrates the lesions of subject 80. The three lesion groups did not differ significantly in the number of coagulations performed.

Neuropsychological Testing

The following tests were administered to each subject according to standard procedures. By using the German version of the Wechsler Adult Intelligence Scale—Revised (WAIS-R), measures of full-scale IQ, verbal IQ, and performance IQ were derived with a short form that comprised the subtests comprehension, similarities, arithmetic, block design, picture completion, picture arrangement, and digit symbol. Mnemonic functions were assessed with a German translation of the Wechsler Memory Scale—Revised. This battery includes subtests of visual and verbal memory. The memory quotient was obtained on the basis of the standardized scoring systems. Further, a paired-associate learning test (13) was applied. This measure requires subjects to subsequently learn two lists of semantically associated word pairs. Intrusions from the first list elicit proactive interference tendencies. The Wisconsin Card Sorting Test (14) was administered to study impairment of abstract conceptualization and shifting ability. Furthermore, a sorting test for emotional stimuli (pictures from the Ekman series) was applied (paper in preparation by Weniger et al.). For testing attentional performance and psychomotor speed, the Trail Making Test (parts A and B) and a computer-driven test battery (15) comprising various subtests (divided attention, go/no-go, flexibility, working memory) were used.

Clinical Assessment

A German version (16) of the Structured Clinical Interview for DSM-III-R—Patient Version (SCID-P) (17) was administered to assess current and previous psychiatric diagnoses. Obsessive-compulsive symptoms were rated for the present and were retrospectively assessed for the preoperative state by using the German authorized translation (18) of the Yale-Brown Obsessive Compulsive Scale (19). Subjects in groups 1, 2, and 3 completed a German version (short form) of the MMPI (20).

Each subject had a comprehensive psychiatric and psychosocial assessment by means of two independent interviews (performed by E.I. and E.R.). A detailed history of the pre- and postoperative course of illness and the current level of psychosocial functioning were recorded. The subjects rated severity and improvement of obsessive-compulsive symptoms on visual analogue scales. Further, obsessive-compulsive symptoms and other complaints were self-rated by the subjects for the entire lifespan, and the subjects were asked to comment on possible causes for major changes. The information obtained was validated by the detailed case notes and by information obtained from current psychotherapists. For each psychiatric diagnosis there was complete correspondence between the two interviews, the SCID-P, and the case notes and information obtained from current psychotherapists.

Statistical Analyses

Statistical computations were based on scaled scores (WAIS-R, revised Wechsler Memory Scale) or raw scores. Because of the small number of subjects in each group, only nonparametric statistical methods were used (Kruskal-Wallis one-way analysis of variance, Wilcoxon test, Mann-Whitney U test, Spearman rank correlation). Frequencies were compared by using the binomial test because of the small number of subjects. All analyses were two-tailed, and the alpha was defined as 0.05. Bonferroni corrections were applied for neuropsychological measurements (e.g., measures of the Wisconsin Card Sorting Test). All statistical comparisons were performed by using SPSS for Windows, version 6.0.1.

RESULTS

Neuropsychological Testing

The neuropsychological test results of all groups are summarized in table 3. No significant group differences were found for verbal or full-scale IQ (WAIS-R) among any of the groups, including the comparison group. However, group 3 (with dorsally extended lesions plus ventromedial frontal lesions) had a significantly lower performance IQ than the comparison group (U=5.0, p<0.05). Three of the five subjects in group 3 performed poorly on performance-related tasks, achieving performance IQ scores of more than two standard deviations below average.

There was no significant group difference for the overall memory quotient yielded by the revised Wechsler Memory Scale, although group 3 scored significantly below the comparison subjects on one subtest of nonverbal memory—visual paired associates, immediate recall (U=5.5, p<0.05). Significant group differences were seen for the number of intrusion errors in the paired-associate learning task (χ²=7.8, df=2, p<0.05). Post hoc analyses revealed that group 3 made significantly more intrusion errors than group 1 (U=1.5, p<0.04) and group 2 (U=0.0, p<0.03). All of the subjects in group 3, one subject in group 2, and none of the subjects in group 1 had scores on this test below the 16th percentile of a normal comparison population.

All three lesion groups showed marked difficulties with the Wisconsin Card Sorting Test. Significant group differences were seen for the percentage of conceptual level responses (χ²=8.2, df=3, p<0.05). The subjects in group 2 (lesions of the ventral striatum plus ventromedial frontal lesions) were the most impaired, producing significantly fewer categories (U=3.0, p<0.02) and more perseverative and total errors than the comparison subjects. However, on a sorting test for emotional stimuli (pictures from the Ekman series), the performance of the group 2 subjects was similar to that of the comparison subjects (group 0), but group 1 (with ventromedial frontal lesions only) performed significantly worse than group 0 (U=3.0, p<0.02) and group 2 (U=2.0, p<0.02).

All groups, including the comparison group, showed reaction times below the 16th percentile of a normal control population on the computerized tests of attentional performance, except for the go/no-go subtest. The subtest for divided attention yielded the worst results; i.e., 16 of the 22 study subjects (including six out of the seven comparison subjects) showed reaction times below the 16th percentile. The performance on the Trail Making Test was for most subjects within the normal range for a comparison population. However, the group 3 subjects were significantly slower than the comparison subjects (part A: U=5.5, p<0.05; part B: U=5.0, p<0.05). Group 3 also had longer reaction times than group 2 on two of the computerized tests of attentional performance, divided attention (U=2.0, p<0.04) and go/no-go (U=1.5, p<0.02). Reaction time on the go/no-go task was also significantly longer for group 1 than for group 2 (U=1.5, p<0.02).

Clinical Assessment

On the basis of detailed reviews of case notes, the SCID-P, and two independent psychiatric interviews, it could be established that at the time of their operations all of the leukotomized subjects had met the DSM-III-R criteria for severe and chronic OCD and that the OCD had been refractory to extensive trials with nonsurgical psychiatric treatments (e.g., antidepressant medication, various kinds of psychotherapy, electroconvulsive shock, sleep therapy). The clinical records of eight subjects contained reports of at least one suicide attempt. Three subjects (numbers 47 and 81 in group 1 and number 80 in group 3) additionally met the DSM-III-R criteria for obsessive personality disorder. All of the subjects in the comparison group met the criteria for OCD, and none had an additional diagnosis of obsessive personality disorder.

According to Wilcoxon tests, the three lesion groups had significant improvements in their scores on the Yale-Brown Obsessive Compulsive Scale between the (retrospectively rated) presurgical state and the current postsurgical state (group 1: z=–2.2, p<0.03; group 2: z=–2.0, p<0.05; group 3: z=–2.0, p<0.05) (see table 1). The subjects in the comparison group significantly improved as well (z=–2.02, p<0.05). Groups 1 and 2 had strong, similar median levels of improvement on the Yale-Brown Obsessive Compulsive Scale (73% and 75%, respectively), but group 3 improved less (56%). However, these differences did not reach conventional levels of significance. The scores of the subjects with both OCD and obsessive personality disorder (N=3) differed from those of the subjects with OCD only (N=13), and the subjects with both OCD and obsessive personality disorder showed significantly less improvement (U=1.0, p<0.01) (table 4). Two of the OCD subjects with obsessive personality disorder (subjects 47 and 80) and two subjects with OCD (subjects 66 and 77) were receiving clomipramine at the time of the study. The present scores on the Yale-Brown Obsessive Compulsive Scale of the two OCD subjects (16 and 10, respectively) were well within the range of the scores for the other OCD subjects, lending further support to the idea that the clinical outcomes of the leukotomized subjects were mainly related to the surgery.

Outcome categories were formed on the basis of pre- and postoperative scores on the Yale-Brown Obsessive Compulsive Scale, as described by Jenike et al. (5). Marked improvement (>75%) was shown by 44% of the subjects (N=7), and another 25% (N=4) showed moderate improvement (>50%). Only one subject (6%) experienced less than 25% improvement in obsessive-compulsive symptoms. Eleven (85%) of the 13 subjects with OCD but not obsessive personality disorder were convinced that the surgery was the cause for all or most of their improvement, which appeared in eight subjects 3–12 months after surgery and in three subjects 1–2 years postoperatively. Two of the subjects with both OCD and obsessive personality disorder experienced less intense improvement 3–8 months postoperatively, and one subject with OCD and obsessive personality disorder experienced no operation effect.

Table 4 contains the rates of improvement for the 16 leukotomized study subjects and for the contacted subjects who did not receive neuropsychological examinations. The OCD subjects completing only questionnaires showed strong improvements (86%) (Wilcoxon test: z=–2.4, p<0.02) on the Yale-Brown Obsessive Compulsive Scale similar to those of the study subjects (82%) (z=–3.2, p<0.01). The self-ratings of outcome on the visual analogue scale were similar for the study subjects, the subjects completing only questionnaires, and the subjects not available for these kinds of evaluation (table 4). The improvement reported by self-rating was significantly related to the improvement in Yale-Brown Obsessive Compulsive Scale scores for the study subjects (r_s=0.68, N=16, p<0.005) and for the subjects completing only questionnaires (r_s=0.68, N=13, p<0.01).

Changes on measures of psychosocial functioning can be seen from table 5. The postoperative scores of the OCD subjects on the Global Assessment of Functioning Scale (axis V of DSM-III-R) were significantly better than the preoperative scores (determined retrospectively). Further, there was a significant increase after surgery in the number of OCD subjects who were able to work. Also, significantly fewer OCD subjects suffered from dysthymic disorder after leukotomy. Sexual problems were reduced postoperatively in five of the seven subjects complaining of them preoperatively. However, for three subjects (numbers 32, 56, and 59) a loss of drive and initiative clearly occurred after surgery. These subjects, as well as those with combined OCD and obsessive personality disorder, also had higher depression scores (t value >70) on the MMPI after surgery.

There was a significant increase in the number of subjects who became addicted to alcohol (N=6) or to alcohol and benzodiazepines (N=2) after psychosurgery, and there was no case of substance dependence before surgery (table 5). The addicted subjects started to drink between 3 and 9 years postoperatively and in most cases during severe psychosocial stress (e.g., divorce, accident). Five of these subjects experienced remissions of their substance dependence during long-term inpatient psychotherapy. The remaining three subjects (numbers 10, 56, and 59) had extraordinarily severe courses of chronic alcohol disorder, being refractory to numerous and protracted inpatient psychotherapies. Subjects 56 and 77 had family histories of alcoholism and had occasionally showed signs of substance abuse before psychosurgery.

Lesion analysis revealed that all of the addicted subjects had lesions in the ventral striatum. Of the 11 subjects with lesions of the ventral striatum (table 2), eight developed substance dependence postoperatively (p<0.01, binomial test). There was no case of postoperative addiction in the subjects without striatal lesions. Accordingly, the extent of ventral striatal lesions (as indicated in table 2) was significantly different for the addicted and nonaddicted subjects (U=3.0, p<0.002).

DISCUSSION

Before the results are discussed it seems appropriate to mention the limitations of this study. First, the number of subjects was relatively small, especially as three different types of lesions were studied. Second, the preoperative data were limited to case records; no formal prospective assessment was made. Thus, all measures of improvement were based on the retrospective judgment of the patients and may therefore be confounded. Third, as psychosurgery was a treatment of last resort, obtaining a comparison group retrospectively that matched the leukotomized subjects in terms of severity and chronicity of illness was extremely difficult. As a consequence, our comparison group mostly contained subjects who were not recommended for leukotomy because of less severe OCD or better responsiveness to other kinds of therapy.

In view of these limitations, our conclusions are tentative. However, the subjects were thoroughly assessed by means of an extensive battery of neuropsychological tests. All subjects were personally interviewed, and the interviewers had not been involved in the treatment of the subjects. The comparison subjects did match the leukotomized subjects well in terms of age, sex, education, duration of illness, and current (although not preoperative) severity of symptoms.

Neuropsychological Findings

A comprehensive neuropsychological test battery was administered to the leukotomized subjects and comparison subjects. We found largely preserved intellectual functions (see table 3), a finding that is consistent with previous results for subjects who had been treated with subcaudate tractotomy (10) or cingulotomy (7). However, group 3 (with dorsolateral frontal lesions) had a significantly lower performance IQ (median=67) than the OCD comparison subjects (median=110) (see table 3). This finding is consistent with the literature, which indicates that loss on WAIS-R scales is more often associated with lesions of the dorsolateral convexity (21). It must, however, be noted that poor performance on visual-spatial tasks, which mainly contribute to the Wechsler performance scale, has been found relatively consistently in OCD subjects who have not received psychosurgery (see, for instance, references 22 and 23), a finding that we could not confirm in our study.

With regard to the data for memory tasks, we could not find a general memory impairment in the leukotomized subjects. However, the dorsolateral group (group 3) performed significantly poorer than the nonleukotomized comparison subjects on a visual and a verbal paired-associate learning task. Our finding of significantly more intrusion errors, thus greater susceptibility to proactive interference, in group 3 is in line with previous reports on effects of frontal lesions on memory and learning. Resistance to interference is often reported to be the aspect of mnestic activity that is most likely to be disturbed after dorsolateral prefrontal pathology (21, 24).

Poor performance on the Wisconsin Card Sorting Test is consistently reported in OCD subjects not receiving surgery (22), as well as in those who do receive surgery (7, 10, 11). We could not confirm the deficit in our nonleukotomized comparison subjects but did find such a deficit in varying degrees for the three lesion groups, with group 2 (lesions of the ventral striatum plus ventromedial frontal lesions) showing the poorest performance. This finding was somewhat unexpected since Wisconsin Card Sorting Test performance has been associated with dorsolateral prefrontal activation of regional cerebral blood flood (25, 26). Additionally, the group 2 subjects showed the best results in a task requiring the sorting and categorization of emotional stimuli (see table 3). Probably the planning and sequencing behavior required for the Wisconsin Card Sorting Test was especially impaired in our group 2 subjects because of their striatal lesions. Lesions of the caudate nucleus have been reported to impair these kinds of behavior, including performance of the Wisconsin Card Sorting Test (27).

Clinical Outcome

In the majority of our subjects obsessive-compulsive symptoms were greatly improved after surgery and so was social adjustment (see tables 4 and 5). Improvement of more than 50% on the Yale-Brown Obsessive Compulsive Scale was found for 69% of the subjects. This improvement rate roughly matches rates found in other studies. In a prospective study of 26 OCD subjects who underwent psychosurgery involving orbitomedial and/or cingulate lesions, Hay et al. (7) demonstrated obvious improvement in 38% of the subjects and mild improvement in an additional 23% according to a 6-point global rating scale based on the Goktepe scale (28).

Jenike et al. (5) retrospectively studied 35 OCD subjects two decades after cingulotomy and established that at least 30% benefited substantially from the procedure, i.e., showed more than 50% improvement on the Yale-Brown Obsessive Compulsive Scale. In a review of all patients who had received stereotactic subcaudate tractotomy in Great Britain, Bridges et al. (6) found 33%–45% of OCD subjects to be in outcome category I or II of Goktepe's 5-point rating scale.

Our study has demonstrated the long-term antiobsessional effect of bifrontal leukotomy in severe OCD. Most subjects show clearly reduced obsessive-compulsive symptoms and improved psychosocial function. Both were sustained for more than two decades. Restricted ventromedial frontal leukotomy should therefore be discussed as a last-resort treatment for severe and refractory OCD but not obsessive personality disorder. However, it seems difficult to judge whether the long-term outcome of leukotomized OCD subjects is significantly better than the natural course of the illness (1). Since psychosurgery was always a treatment of last resort, obtaining a comparison group without psychosurgery that matched our leukotomized subjects in terms of severity and chronicity of illness was extremely difficult. However, most of our leukotomized subjects were convinced that they would have tried to commit suicide if their severe illness had persisted (eight subjects had attempted suicide before their operations).

Comparing the efficacy of the different psychosurgical procedures has proved difficult. However, the existing data suggest that the different procedures (i.e., tractotomy, capsulotomy, cingulotomy) have roughly comparable degrees of efficacy. Also, we could not find significantly different effects of the different surgical approaches, despite detailed analyses of the lesions of each subject. Nevertheless, group 2 (lesions of the ventral striatum plus ventromedial frontal lesions) tended to show the most improvement (scores on the Yale-Brown Obsessive Compulsive Scale are given in table 1).

Evidence is converging that the projection of allocortical frontal regions onto the ventral striatum (reference 29 provides an overview) is a critical route for obsessive-compulsive behavior: the serotonergic innervation of the most anterior ventral striatum is profound (30), and serotonin reuptake inhibitors have been shown to be effective for many OCD subjects (31). Successful behavior therapy or pharmacotherapy with fluoxetine was shown to reduce the metabolic activity of the caudate nucleus and orbitofrontal cortex of OCD subjects (32). On the other hand, provocation of obsessive-compulsive symptoms may increase the orbitofrontal cortex activity of OCD subjects (33). Thus, the common feature of our three lesion groups may be that all lesions effectively interrupted the frontal/basal ganglia/thalamic/frontal loop in our subjects at some stage. However, quantitative morphometric studies point to widely distributed structural brain abnormalities in the OCD subjects as well, e.g., low white matter volume, high ventricle-to-brain ratio (34).

Surprisingly, the ventromedial frontal lesions of our subjects were followed by many fewer deficits in the domain of emotion and personality than would be assumed from human neuropsychological lesion studies (e.g., references 21, 35). Our psychiatric assessment indicated that many subjects had difficulties in relating their emotional state to self-cognition, but none of these subjects seemed to be obviously impaired in his or her personal life. An explanation for this could be that the ventromedial frontal cortex of our OCD subjects was hyperfunctional in the preoperative state and that lesioning part of this cortex returned the activity of its functional loops toward normal levels. Similar proposals have been made by the neurosurgeons performing the lesions in our subjects.

Substance Dependence

A serious finding was the occurrence of addiction in eight subjects (50%) after surgery; the addictions of three subjects were untreatable. Although there is a generally higher than normal risk for OCD subjects to become addicted in the long course of a debilitating illness, the reported prevalence rates are not higher than 12%–15% (36, 37). Our figures are clearly higher than this. Other reports do not mention any case of postoperative-onset addiction among OCD subjects (3, 5, 7–9), even after subcaudate tractotomy.

All of the addicted subjects in our study had lesions of the ventral striatum (see table 2 and Results section, “Clinical Assessment”). Research suggests that the ventral striatum (including the nucleus accumbens) is the main link between the limbic system and motor output (29). Information derived from limbic system circuits may reach the caudate nucleus and putamen by several routes, including inputs from the amygdala/hippocampus complex and from the reward-related dopamine-containing neurons in the substantia nigra pars compacta and the ventral tegmental area (29, 38). The ventral striatal lesions in our subjects most probably led to a disturbance or interruption of information flow between the limbic and motor systems.

There is now an extensive literature from animal research that points to the involvement of the ventral striatum in the regulation of motivation and reward and its role in addiction (reference 39 provides an overview). For example, creation of lesions in the region of the nucleus accumbens in rats leads to a decrease in the self-administration of cocaine (40), which possibly reflects an anhedonia due to disturbance of the mesocorticolimbic dopamine system. It might be speculated that our subjects with ventral striatal lesions developed anhedonia, facilitating drug abuse and dependence in the presence of psychosocial stress. However, little is known about the role of the ventral striatum and nucleus accumbens in humans because specific lesions of this area are extremely rare. Our findings could indicate a role of this area in human addictive behavior, although the exact mechanisms remain to be elicited.

TABLE 1

Enlarge table

TABLE 2

Enlarge table

TABLE 3

Enlarge table

TABLE 4

Enlarge table

TABLE 5

Enlarge table

Enlarge table