t step, we administered the International Intrusive Thoughts Interview Schedule (IITIS)  and determined each participant’s the most distressing intrusive thought. Next, participants were assigned to the induction of an intrusive thought or the induction of a neutral thought prior to the executive function tasks. Three executive function tasks were administered: the number-letter task   and the local-global task   probed mental flexibility and the go/no-go task probed response inhibition. The objective was to compare the effect of the most distressing intrusive thought versus a neutral thought on executive function. We hypothesized that (i) OCD patients would present an executive deficit relatively to control participants, and (ii) the most distressing intrusive thought would have a greater impact on executive tasks in patients with OCD than in non-clinical control participants.
2. Materials and Methods
We assessed 22 patients with OCD (mean ± standard deviation age: 34.68 ± 9.50; range: 17 - 47) attending the Savoie Mont-Blanc University Psychological Consultation Center (Chambéry, France) and the Grenoble University Medical Center (Grenoble, France) during one year. The diagnosis of OCD was based on the DSM IV criteria, following a semi-structured Mini International Neuropsychiatric Interview (MINI) with trained psychologists  . The Vancouver Obsessional Compulsive Inventory (VOCI)  was used to check for the presence of OCD. Individuals with past or current neurological disease (e.g. epilepsy, brain tumor, traumatic head injury or stroke) or past or current schizophrenia were excluded from the study. The demographic and clinical characteristics of the participants in the OCD and control groups are summarized in Table 1.
Furthermore, 22 control participants (mean ± standard deviation age: 37.86 ± 12.95; range 21 - 61) were recruited by advertisements in Savoie Mont-Blanc University and elsewhere, and were selected to match the patient groups for age, gender ratio and educational level (expressed as years of formal education, from primary school onwards) (Table 1). Again, we screened for exclusion criteria (notably the presence of any neurological disease, psychiatric disorder or medication use (e.g. antipsychotics and sedatives) that could affect cognition during the MINI.
We also administered the Beck Depression Inventory (BDI)  to control for depressive symptoms that might have interfered with cognitive performance. The IITIS (IITIS)  was used to identify each participant’s most distressing intrusive thought (for more details, see  ) from among seven categories: contamination/dirt/disease intrusions, harm/injury/aggression intrusions, doubting intrusions, unwanted religious or immoral intrusions, unwanted sexual intrusions, intrusions of being a victim of violence, and “other intrusions” not falling
Table 1. Demographic and clinical characteristics of the patients with OCD and the healthy control participants.
into the other categories (Table 2). Written informed consent was obtained from each participant. The study was approved by the Savoie-Mont-Blanc University’s investigational review board (reference: 20145).
Materials and Procedures
The Behavior Rating Inventory of Executive Function Adult version (BRIEF-A)  self-questionnaire was used to check for executive dysfunction in everyday life.
The executive function tasks were administered after intrusive thought induction or neutral thought induction. For the induction of intrusive thoughts, the participant was told to write down the most distressing intrusive thought determined in the IITIS. He/she was then told to close his/her eyes and concentrate on this phrase for 30 seconds. For the induction of neutral (control) thoughts, the participant was told to write down the name of a noise that he/she hears frequently in his/her favorite or local park. He/she was then told to close his/her eyes and concentrate on this phrase for 30 seconds. Control and OCD participants were assigned to both intrusive thought/neutral conditions.
Before the three executive function tasks, each participant had to score his/her levels of anxiety, worry and guilt on 10-point Likert scales ranging from 1 (“not at all”) to 10 (“extremely”). All participants then performed three computer-based cognitive tasks assessing the executive functions of inhibition (the go/no go task) and mental flexibility (the number-letter task  and a local-global task   . Superlab 2.0 software was used to present stimuli and record responses for all three tasks.
The order of the tasks was counterbalanced for each participant and each induction condition.
The Behavior Rating Inventory of Executive Function Adult version.
The BRIEF-A is designed to assess the executive behavior of adults in home and work environments   for the French adaptation. This 75-item questionnaire yields a Behavioral Regulation Index (BRI), a Metacognitive Index (MI) and also an overall Global Executive Composite score (GEC). The BRI comprises four non-overlapping clinical scales (Inhibit, Shift, Emotional Control, and Self-Monitor), and the MI comprises five scales (Initiate, Working Memory, Plan/Organize, Task Monitor, and Organization of Materials). Self-reported and informant-reported versions of the BRIEF-A are available; only the self-reported version was administered here. For each participant, T scores
Table 2. The most distressing types of intrusive thought in the two groups
(based on comparisons with a normative sample) were calculated for each of the three main indices. Higher T scores reflect greater impairments in executive functions.
The go/no-go task.
This task examined the participant’s ability to attend to relevant targets and to inhibit the motor response to distractors. Participants were instructed to (i) respond (by pressing on the mouse key) as quickly and as accurately as possible when a red rectangle appeared at the center of a computer screen (the “go” condition) and (ii) not respond (and wait for the next trial) when a green rectangle appeared at the center of the screen (the “no-go” condition). The two conditions were randomly distributed during the task. Omission errors (when the participant failed to respond in the “go” condition) and commission errors (when the participant responded in the “no-go” condition) were recorded. It is thought that commission errors reflect difficulties in motor inhibition.
The number-letter task (adapted from Rogers & Monsell   ).
A number-letter pair (e.g. “7G”) was presented in one of four quadrants of a computer screen. Participants were instructed to indicate (by pressing a button) whether the number was odd or even when the pair was presented in either of the two top quadrants, and whether the letter was a consonant or a vowel when the pair was presented in either of the two bottom quadrants. The task was divided into three blocks, reflecting two main conditions. In the “no-switching” condition, a first block of 32 target trials appeared solely in the two top quadrants and a second block of 32 target trials appeared solely in the two bottom quadrants. The “switching” condition contained a third block of 64 target trials in which the number-letter pair was presented successively in all four quadrants in clockwise order; this required participants to shift between the two types of categorization every two trials.
A “reaction time (RT) shift cost” was calculated from the difference between the average RT in the third block of trials (in which mental shifting was required) and the average RT in the first two blocks of trials (in which mental shifting was not required). Likewise, an “error shift cost” was also calculated from the number of errors in the third block vs. the first two blocks.
The local-global task (adapted from Miyake and colleagues,  ).
A geometric, local-global “Navon figure” was presented on a computer screen. The ‘‘global’’ element (such as a triangle) was composed of other much smaller ‘‘local’’ elements (such as squares)  . Four different geometric figures (a square, a triangle, a circle and a cross) were used as the global or local elements. The figures were displayed in black or in blue. Participants were instructed to state the name of the global figure if the figure was displayed in black and to state the name of the local figure if the figure was displayed in blue. As the color of the stimuli changed from black to blue or vice versa during the series of trials, the participants had to shift from examining the local features to the global features or vice versa. Ninety-six pre-randomized target trials and 36 practice trials were presented. Half the trials required a switch from local to global features or vice versa, and the RTs were measured. A “shift cost” was calculated from the difference between the mean RT in the trials requiring a shift in mental set (i.e. when the color of the stimulus changed) and the mean RT in the trials in which no shift was required (i.e. when the color of the stimulus did not change). A “shift cost” was also calculated for the number of errors in the local-global task.
We used Statistica software (version 10) for all analyses. Intergroup differences in demographic and clinical characteristics were analyzed in independent t tests. The effect of gender as a categorical variable was examined in a chi-squared test. All tests were two-tailed, and the threshold for significance was set to p < 0.05. Univariate analyses of variance were used to assess intergroup differences in the questionnaire scores (Table 1). Effect sizes were calculated by using the partial η2 method.
Demographic and clinical characteristics
In the OCD group, 18 of the 22 participants (82%) presented with comorbid disorders (according to the MINI). Seven of these 18 participants had only one disorder associated with the primary diagnosis of OCD, 8 presented with two related disorders, 2 participants had three related disorders, and 1 had five comorbid disorders. The most common disorders were generalized anxiety disorder (n = 13), panic disorder (n = 6), agoraphobia (n = 5) and social phobia (n = 5). Fourteen of the 22 OCD participants were taking antidepressants or other medications. None of the non-clinical participants presented with disorders or were taking medications.
The two groups did not differ in terms of sociodemographic characteristics, gender distribution (Χ2 = 0.83, p = 0.36), age (t(42) = 0.92, p = 0.36) or educational level (t(42) = 1.52, p = 0.14) (Table 1).
The VOCI total score differed significantly when comparing patients with OCD and control participants (F(1, 42) = 54.70, p < 0.001, η2p = 0.57). Furthermore, univariate tests on each of the 6 VOCI factors revealed significant intergroup differences for checking (F(1, 42) = 36.78, p < 0.001, η2p = 0.47), contamination (F(1, 42) = 32.97, p < 0.0001, η2p = 0.44), hoarding (F(1, 42) = 20.30, p < 0.001, η2p = 0.33), “just right” (F(1, 42) = 26.36, p < 0.001, η2p = 0.39), and indecisiveness (F(1, 42) = 25.45, p < 0.001, η2p = 0.38). No intergroup difference was found for obsessions (F(1, 42) = 0.42, p = 0.52, η2p < 0.01). With the exception of the latter factor, patients with OCD had significantly higher VOCI factor scores than control participants. The main effect of group was also significant for the BDI (F(1, 42) = 25.86, p < 0.001, η2p = 0.38), with higher total scores for patients with OCD than for control participants (Table 1). The most distressing type of intrusive thought for patients with OCD was contamination (n = 8, 36%), followed by doubt (n = 7; 32%). The most distressing type of intrusive thought for non-clinical participants was doubt (n = 10; 45%) (see Table 2).
As can be seen in Table 3, overall T scores in the BRIEF-A were higher for patients with OCD than for control participants. The main effect of group was significant for all three indexes, i.e. the BRI: F(1, 42) = 21.76, p < 0.001, η2p = 0.34; the MI: F(1, 42) = 4.74, p = 0.04, η2p = 0.10; and the GEC: F(1, 42) = 12.57, p < 0.001, η2p = 0.23).
We measured the effect of thought induction on the mean anxiety, worry and guilt scores in the two groups. Significant effects of group and thought induction were observed for anxiety (F(1,40) = 9.51, p < 0.01, η2p = 0.19 for group and F(1,40) = 6.53, p < 0.05, η2p = 0.14 for thought induction), worry (F(1,40) = 5.69, p = 0.02, η2p = 0.12 for group and F(1,40) = 16.13, p < 0.001, η2p = 0.29 for thought induction), and guilt (F(1,40) = 4.89, p = 0.03, η2p = 0.11 for group and F(1,40) = 4.89, p = 0.03, η2p = 0.11 for thought induction). None of the interactions was significant. In both groups, intrusive thought induction yielded higher anxiety, worry and guilt scores than neutral thought induction did. However, the overall effect size was greater in patients with OCD (Table 4).
According to the BDI, patients with OCD were more depressed than healthy controls. Given that depression might have influenced the dependent variables (i.e. the executive function scores), the BDI total score was used as a covariate in the analyses. For each executive task, a two-way analysis of covariance was performed with group (OCD, healthy) and thought induction (intrusive, neutral) as between-participants factors and the BDI total score as a covariate. Effect sizes were also calculated, using the partial η2 method. Table 5 summarizes the executive task scores for patients with OCD and control participants as a function of
Table 3. Mean ± standard deviation T-scores for control participants and patients with OCD in the self-reported BRIEF-A
Table 4. Means ± standard deviation anxiety, worry and guilt scores (range: 1 to 10) as a function of the group and the type of thought induction
Table 5. Mean ± standard deviation executive task scores for control participants and patients with OCD, as a function of the type of thought induction
the type of thought induction.
The go/no-go task.
There was a significant intergroup difference in the commission error rate (errors in the no-go condition; F(1, 39) = 5.15, p = 0.03, η2p = 0.12); patients with OCD had a higher commission error rate than healthy participants. The size effect was moderate. No difference in the commission error rate was observed when comparing intrusive thought and neutral thought inductions (F(1, 39) = 0.01, p = 0.91, η2p < 0.001), and there was no group x thought induction interaction (F(1, 39) = 0.71, p = 0.40, η2p < 0.001). Hence, neither group shown an effect of thought induction (see Table 5).
The two groups did not differ significantly in terms of the omission error rate (i.e. errors in the “go” condition): F(1, 39) = 1.45, p = 0.24, η2p = 0.03. The main effect of thought induction was not significant: F(1, 39) = 3.01, p = 0.09, η2p = 0.07. There was no group x thought induction interaction: F(1, 39) = 0.71, p = 0.40, η2p = 0.02.
The number-letter task.
We observed an intergroup difference in the RT shift cost (F(1, 39) = 5.75, p = 0.02, η2p = 0.13), with a greater RT shift cost for patients with OCD than for healthy controls. A trend towards an effect of thought induction (F(1, 39) = 4.03, p = 0.052, η2p = 0.09) was observed, although there was no group x thought induction interaction (F(1, 39) = 0.74, p = 0.40, η2p = 0.02).
The main effect of group was not significant for the error shift costs (F(1, 39) = 0.04, p = 0.84, η2p = 0.001) There was neither an effect of thought induction (F(1, 39) = 0.12, p = 0.74, η2p = 0.003) nor a group x thought induction interaction (F(1, 39) = 0.05, p = 0.83, η2p = 0.001).
The local-global task.
In the local-global task, the main effect of group failed to achieved significance: (F(1, 39) = 0.006, p = 0.94, η2p < 0.001), and there was no main effect of thought induction on the RT shift cost (F(1, 39) = 0.76, p = 0.39, η2p = 0.02). The patients with OCD were not slower than the healthy controls (see Table 5). Importantly, the two-way group x thought induction interaction was not significant (F(1, 39) = 0.09, p = 0.77, η2p = 0.002).
When considering the error shift cost, there was no intergroup difference (F(1, 39) = 0.05, p = 0.82, η2p = 0.001) and no effect of thought induction (F(1, 39) = 0.03, p = 0.87, η2p < 0.001). The group x thought induction interaction was not significant (F(1, 39) = 2.87, p = 0.10, η2p = 0.07).
The objectives of the present study were (i) to evidence an executive impairment in OCD patients and (ii) to compare the effect of the most distressing intrusive thought on executive function (response inhibition and mental flexibility) in patients with OCD and in healthy control participants. As expected, the patients with OCD had significantly higher scores than the control group participants for all questionnaires (the VOCI and the BDI), with the exception of the VOCI obsession subscale. This finding might be due to the fact that the VOCI was completed after the IITIS (with the latter highlighting the intrusive thought). Two types of intrusive thought (contamination and doubt) were prevalent in OCD participants, whereas doubt was prevalent in control participants. This is in agreement with the literature data   . We found an effect of group on the behavioral aspects of executive function in the self-reported BRIEF-A. Patients with OCD generally considered that they had poor executive function (for both behavioral and cognitive dimensions) in everyday life. The effect size was moderate for the BRI and the GEC, and modest for the MI.
We measured the effect of inducing the intrusive or neutral thought on anxiety, worry and guilt. As expected, we observed a greater effect of intrusive thoughts on anxiety, worry and guilt in the OCD group than in the control group. However, the difference between neutral thought induction and intrusive thought induction was the same in the two groups for all three variables (i.e. there was no group x intrusive thought interaction). Generally, obsessions (intrusive thoughts) are more distressing and interfere more with everyday life for patients with OCD than for non-clinical participants   . Inducing the participant’s most distressing intrusive thought does not have the same effect as obsessions or spontaneous intrusive thoughts do. OCD participants had higher anxiety, worry and guilt scores than control participants.
The study by Basso and colleagues  showed that executive function impairments were more related to co-morbid depressive symptoms than to OCD per se. Given that our patients with OCD had a higher BDI score and that depression affects executive function, we used the BDI score as a covariate in our analysis. Overall, we expected to observe a group effect in each cognitive task (i.e. impaired executive function in patients with OCD, relative to control participants) and a greater effect of intrusive thought induction in the patients with OCD than in the controls (i.e. with a significant group x intrusive thought interaction). In the literature, there is some evidence of impaired response inhibition in OCD, as assessed in go/no-go tasks   . It is usually assumed that commission errors (i.e. a response when none is required, such as no-go errors and false alarms) reflect impulsivity and impaired inhibitory control. In our experiment, the go/no-go task revealed a significant intergroup difference in the commission error rate; patients with OCD made more commission errors than controls. Furthermore, there was no effect of intrusive thought induction in the OCD group or the control group. Our findings are in line with the results of Abramowitch and colleagues meta-analysis  , which showed a small effect on commission errors in inhibitory control tasks. Commission errors are thought to reflect impaired motor inhibition, whereas omission errors in the go/no-go test may reflect attentional control problems (inattention). We did not observe a difference between the OCD and control groups, regardless of the type of induction. This result is consistent with the literature data  . A shifting impairment in OCD has been reported  . In the present study, we administered two tasks that required switching between tasks or mental sets: the number-letter task  and the local-global task  . In both tasks, shift costs were calculated for RTs and error rates by subtracting the score in the “no switching” condition from the score in the “switching” condition. An intergroup difference in RT shift costs (but not in error shift costs) was observed in the number-letter task (with greater shift cost for patients with OCD and a moderate effect size). There was a small, non-significant difference in RTs between neutral and intrusive thought induction for both groups, and the effect size was moderate. In the literature (as in the present study), intergroup differences in the RT shift cost tend to be greater than intergroup differences in the error shift cost. Likewise, no effect on the RT shift cost or the error shift cost was observed in the local-global task; there were no intergroup differences, no effects of thought induction, and no interactions. In patients with OCD, the local-global task may be less sensitive than the number-letter task for assessing mental set-shifting. The symptoms in OCD may result from over-focused attention at the local level, which affects the stimulus process  . This hypothesis was confirmed in a local-global paradigm: OCD participants were impaired in a global task, relative to control participants  . However, the latter researchers did not control for depressive symptoms. It is also possible that the composition of the OCD group (i.e. checkers vs. washers, etc.) differed from that of the present study. Our present results in the local-global task and for the omission error rate in the go/no-go task did not confirm the previously reported excessive focus on local detail in a “heterogeneous” OCD group. One can hypothesize that over-focused attention is specific to “checkers”. When controlling for depression, we observed an intergroup difference in the commission error rate (in the go/no-go task) and a significant effect of group on the RT shift cost of the number-letter task. OCD participants performed less well than healthy controls in both executive tasks. In both groups, we also observed a trend towards an effect of thought induction in the number-letter task. More importantly, intrusive thought induction did not have a greater effect on the OCD group than on the control group in either task. With regard to the categories of intrusive thoughts, most patients with OCD are washers and then checkers. There are few studies of OCD subcategories and executive function. In the study by Van der Linden, Ceschi, Zermatten, Dunker, and Perroud  , patients with OCD (checkers and washers) had more difficulty in inhibiting a prepotent response (in the Hayling response inhibition task) than control individuals, although there was no difference between checkers and washers. In another study  , checkers had poorer results than washers in the Stroop test, the trail-making test, category fluency, and commission errors in the go/no-go test. There were also significant differences between checkers and washers in the inhibition and the cognitive flexibility scores. The presence and severity of hoarding symptoms in OCD was found to be associated with impaired decision-making  . It is possible that checkers and washers perform differently in the executive function tasks used in our research. Our OCD group consisted of washers and checkers. Ideally, this heterogeneity should have been taken into account in the statistical analysis but sample size in each category was too small. It may be important to consider the OCD symptom dimensions when studying neuropsychological task performances.
This study has several limitations: 1) the sample size could be more important; 2) in order to determine the extent to which executive deficits are specific to OCD or a general impairment relative to a clinical condition, a control group with another disorder could be studied; 3) the effects of medication on executive functioning have not been considered.
Patients with OCD reported that they have poor executive function in everyday life, with impaired mental flexibility (in the number-letter task) and response inhibition (in the go/no-go task) than control participants. The local-global task (a test of mental flexibility) did not reveal a difference between patients with OCD and control participants, and so appears to be less effective for assessing mental set-shifting in the context of OCD. The induction of an intrusive thought had no effect on patients with OCD (relative to controls) when considering anxiety, worry, guilt and executive function scores. As a recommendation, we could say that ecological tests (such as BRIEF-A) seem more useful to differentiate OCD patients from control participants than laboratory tests.