Spatial working memory in obsessive–compulsive disorder improves with clinical response: A functional MRI study

Abstract

To date, only a few studies have examined whether executive dysfunctions in obsessive–compulsive disorder (OCD) are state or trait dependent and almost none of these studies have used functional neuroimaging techniques. We conducted a functional MRI study before and after 12 weeks of pharmacological treatment in 14 psychotropic-free patients with OCD without comorbidity. Subjects performed a spatial variant of a working memory task with four increasing levels of difficulty (n-back task). Responders and non-responders did not differ in clinical and demographical characteristics or brain activation patterns before treatment. Performance improved only in responders and was associated with a change in the overall pattern of brain activity during the task. We found no correlations between (changes in) scores on symptom scales, brain activity and performance. Our preliminary findings suggests that spatial working memory deficits in OCD and their functional anatomical correlates, as assessed with a spatial n-back task, are, at least to some extent, state dependent.

Introduction

Obsessive–compulsive disorder (OCD) is characterized by recurrent, persistent and intrusive thoughts or images that cause anxiety or distress (obsessions), and repetitive behaviors or mental acts aimed at reducing this distress or anxiety (compulsions). Patients recognize that the obsessions and compulsions are unreasonable and are products of their own mind. Recent epidemiological data suggest prevalence rates of OCD from 1.5% to 3% (Rasmussen and Tsuang, 1984, Regier et al., 1990, Stein et al., 1997).

Convergent evidence suggests a neurobiological basis for OCD, resulting in the current fronto-thalamo-striatal model of OCD. According to this model, obsessive–compulsive symptoms result from dysfunction of complex frontal, thalamic and striatal circuitry (Saxena et al., 1998, Saxena et al., 2001). This circuitry is also known to be involved in executive functioning, i.e. higher-order processes necessary for effective and appropriate goal directed behavior, based on which disturbances of executive functioning in OCD may be expected.

Indeed, studies examining cognitive and executive functioning in OCD have demonstrated deficits in several domains of executive functioning such as cognitive set-shifting, trial-and-error learning and inhibition capacities (Jones et al., 1998, Gross-Isseroff et al., 1996, Head et al., 1989, Rosenberg et al., 1997). Subtle cognitive deficits in these domains have also been reported in sub-clinical populations (Mataix-Cols et al., 1999, Roth and Baribeau, 1996). More specifically, some recent studies, including the few functional imaging studies with a cognitive paradigm, have shown that OCD may be associated with specific cognitive deficits, taken to reflect fronto-striatal dysfunction (Purcell et al., 1998a, Purcell et al., 1998b, Rauch et al., 1997, Ursu et al., 2003). One of the cognitive domains addressed in these studies is working memory, i.e. a system that regulates manipulation and short-term storage of information (Baddeley, 1992). OCD patients perform poorly on tasks that engage spatial working memory (i.e. the working memory system that processes information on the location or position of items in space), especially when the task becomes more difficult (Purcell et al., 1998a, Purcell et al., 1998b).

In a previous functional MRI study we assessed performance on a working memory task and the associated brain functioning in a carefully selected sample of patients and closely matched controls (van der Wee et al., 2003). The working memory task employed in that study involves multiple levels of difficulty (load) and therefore enables visualization of the way in which the working memory system responds to increasing load. The advantage of this design is that it allows for determination of the relationship between level of difficulty, regional brain activity and performance (Jansma et al., 2000, Callicott et al., 1998, Callicott et al., 2000). We used a spatial variant of the n-back working memory task that has been used in healthy volunteers and patients with schizophrenia (Callicott et al., 1998, Jansma et al., 2000). In this task subjects see four large dots in a diamond shaped diagram with one dot changing color at a time (the stimulus). After each stimulus they have to indicate the location of a stimulus that has appeared a certain number of stimuli earlier (i.e. n-back). This variant of the n-back task consistently elicits activity bilaterally in dorsolateral prefrontal and parietal cortices, and in the anterior cingulate in patients with schizophrenia and in healthy controls. The involvement of these brain areas in (spatial) working memory has also been demonstrated in other functional imaging studies of working memory using other paradigms (Manoach et al., 1999, Manoach et al., 2000, Manoach et al., 2001, D'Esposito et al., 1999). We found that patients with OCD performed poorly at the highest level of task difficulty but engaged the same brain regions as the matched healthy controls. Brain activity in a region covering the anterior cingulate cortex was significantly elevated in patients with OCD at all difficulty levels of this task (van der Wee et al., 2003).

Although several studies have found executive dysfunctions and functional brain abnormalities in patients with OCD, to date only a few studies have examined whether these dysfunctions are state or trait dependent and almost none of these studies have employed functional neuroimaging techniques. We therefore decided to investigate the state or trait dependent nature of the working memory deficit as found with the n-back task in OCD and the associated abnormal activity in the anterior cingulate cortex (van der Wee et al., 2003). We conducted a functional MRI study before and after pharmacological treatment. To exclude the confound of prior treatment effects, we included only patients who were psychotropic-naïve or psychotropic-free for at least 2 weeks and who had not received cognitive behavioral therapy for at least 3 months. We hypothesized that the non-verbal spatial working memory deficit in OCD would improve (i.e. normalize) with successful treatment (and thus be a state dependent phenomenon) and that an improvement would be associated with a decrease of the activity in the anterior cingulate cortex.