INTRODUCTION

A number of studies have demonstrated the existence of information-processing abnormalities in obsessive-compulsive disorder (OCD), especially cognitive inhibition deficits (see Tallis, 1995). The term “cognitive inhibition” is used to describe a variety of potentially different processes (Hasher et al., 1999), especially those restricting access to strong but situationally inappropriate responses and suppressing the activation of no longer relevant information. A related concept, interference, refers to the likelihood of decreased performance in the presence of distracting stimuli. Inhibition deficits refer to difficulties in inhibiting irrelevant information in general; they must be distinguished from inhibition bias, which refers to difficulties inhibiting information with certain specific content (e.g., related to the patient's primary concerns). The existence of such inhibition deficits in OCD has been identified in several studies, mainly using two different procedures: the Stroop task (Stroop, 1935) and the negative priming procedure (Tipper, 1985).

In a series of studies, Enright and Beech (1990, 1993a, 1993b) showed that OCD patients, unlike patients with other anxiety disorders, exhibited reduced negative priming effects for identity. The negative priming paradigm was originally developed to assess and quantify the inhibitory component of selective attention. It refers to the disruption (usually slowing) of the response to an item if it has previously been ignored. The inhibition explanation of the negative priming effect is that an inhibitory process blocks the representation of a distractor from access to the response systems. If the distractor subsequently appears as a target, the inhibition will take time to dissipate, evidenced by a delay in responding. In a typical negative priming paradigm, participants have to select a target from among one or more distractors (the prime trial). In a subsequent trial (the probe trial), an item that was a distractor in the previous trial becomes the target. The change from distractor to target can be based on the identity of the stimulus (e.g., participants are shown a prime trial of a target picture printed in red, with a distracting picture printed in green, followed by a probe trial in which the red target picture to be named is the same as the distracting picture from the prime trial) or on the spatial location of the stimulus (e.g., the location of the target in the probe trial is identical to the location of the distractor in the prime trial). Enright et al. (1995) also found that the decreased negative priming effect observed in OCD patients was greater in checkers than in noncheckers, especially when the stimulus presentation time was short (Enright et al., 1995). However, these findings were challenged by MacDonald et al. (1999), who observed a similar amount of negative priming in nonclinical controls and in two OCD groups (checkers and noncheckers) by using two negative priming procedures for identity, one based on a perceptual feature (i.e., color) and one based on a semantic feature (i.e., referent size). Similarly, Hartston and Swerdlow (1999) used a spatial negative priming task and also reported a normal negative priming effect in OCD patients. More recently, McNally et al. (2001) found only a marginally significant negative priming deficit in OCD patients, relative to controls. In addition, checkers were not especially impaired. Finally, Hoenig et al. (2002) demonstrated that the negative priming effect was differentially impaired in OCD subgroups (checkers and noncheckers) depending on the response-stimulus interval (RSI).

A few studies using the Stroop task observed higher interference costs in OCD patients (Hartston & Swerdlow, 1999; Moritz et al., 1999). In Hartston and Swerdlow's (1999) study, greater Stroop interference was found despite normal negative priming. However, none of these studies included an anxious control group; consequently, it cannot be determined whether the results obtained were specific to OCD. In addition, two studies (Moritz et al., 2002, 2004) did not replicate the earlier finding of enhanced Stroop interference in OCD participants (Moritz et al., 1999). Similarly, Nakao et al. (2005) showed that OCD patients and normal controls did not differ on a Chinese character Stroop task. However, a study conducted by Bannon et al. (2002) revealed that OCD patients exhibit inhibition deficits, as assessed by a Stroop test (as well as a Go/No Go task), when compared with patients presenting panic disorders.

Besides the heterogeneity of these results (partly because of procedural differences), another problem with the use of the negative priming procedure and the Stroop task to explore inhibition in OCD is that the interpretation of group differences in terms of inhibition deficits remains controversial. First, there is much debate as to whether the negative priming effect actually reflects inhibition. Indeed, other accounts of the effect, involving episodic retrieval (Neill, 1997) or temporal discrimination (Milliken et al., 1998), have been proposed. In addition, the inhibition account of negative priming does not really clarify the relationships between negative priming and Stroop interference. In light of the dissociations observed in some studies between the two effects, it has been suggested that Stroop interference and negative priming might result from independent inhibitory mechanisms (Stoltzfus et al., 1993). In the interference situation, inhibitory processes should intervene to reduce interference during concurrent response selection; in the negative priming situation, inhibitory processes should act to prevent recently rejected information from influencing the current task. However, evidence supporting this interpretation is controversial (Kieley & Hartley, 1997). Furthermore, although the interference portion of the Stroop task has generally been considered to examine resistance to interference (Nigg, 2000), it might also be viewed as tapping into the deliberate suppression of dominant responses (i.e., reading the words). More importantly, it has also been suggested that the classical interpretations of the Stroop task (i.e., in terms of a problem with either resisting interference or inhibiting a dominant response) are somewhat simplistic and that the mechanisms that underlie the Stroop effect are clearly more complex. In particular, the Stroop effect has been interpreted as reflecting the intervention of a general attentional process that modulates the respective contributions of the word-reading and color-naming processes by increasing the gain for the color pathway and/or decreasing the gain for the word-reading pathway (Balota & Faust, 2001; Barrett et al., 2004). Contradictory findings and interpretation difficulties concerning OCD subjects' performance also exist for other inhibition tasks such as the inhibition of return task (Nelson et al., 1993; Rankins et al., 2004; Moritz & von Mühlenen, 2005). Finally, most studies that identified an inhibition deficit in OCD patients did not examine the correlations between this deficit and obsession/compulsion symptomatology. Furthermore, of the few studies that did calculate such correlations, only one found significant correlations between obsessive and compulsive symptoms and Stroop reaction time (and not with Go/No Go performance), and this was observed exclusively in the less severely affected patients (Bannon et al., 2002).

In view of these heterogeneous results and controversial interpretations, the purpose of our study is to further investigate inhibition processes in OCD by using another task—the Hayling task—developed by Burgess and Shallice (1996) specifically to evaluate the ability to inhibit a prepotent or automatic response. This ability is conceived by Burgess and Shallice to be one of the functions devoted to the Supervisory Attentional System (SAS), an executive system in charge of the control of action and coping with novelty (Norman & Shallice, 1980). The SAS is required in situations where the routine selection of actions is unsatisfactory or unavailable, or where strong habitual responses must be inhibited. In this theoretical framework, an SAS deficit affecting the ability to inhibit a dominant response should prevent OCD patients from inhibiting the automatic cognitive schemas that progressively develop by integrating intrusive thoughts, appraisals, and underlying beliefs, as well as compulsions (Tallis, 1995; Lubman et al., 2004).

In the Hayling task, participants are presented with sentences in which the last word is missing. What this last word should be is strongly cued by the rest of the sentence. Two conditions are administered. In the initiation condition, participants have to complete the sentence by adding the missing word (e.g., “To protect himself against rain, he opened his … umbrella”). This condition would require automatic activation of stereotyped responses. In the inhibition condition, the participant must produce a word that makes no sense in the context of the sentence. Therefore, the participant has to inhibit the automatic response before generating a new one. The Hayling task has frequently been used to explore inhibition abilities in various psychopathological and neuropsychological conditions such as schizophrenia (Marczewski et al., 2001), alcoholism (Noël et al., 2001), Tourette's syndrome (Channon et al., 2004), focal frontal lesions (Burgess & Shallice, 1996; Andrés & Van der Linden, 2001), Alzheimer's disease (Collette et al., 1999), and Parkinson's disease (Bouquet et al., 2003). In addition, it has been shown that inhibition deficits, as assessed by the Hayling task, are related to problems maintaining short-term abstinence from alcohol (Noël et al., 2002) and to auditory hallucinations in schizophrenia (Waters et al., 2003). However, besides response inhibition abilities, it appears the Hayling task also requires the ability to strategically generate nonstereotypical responses; these two capacities stand in a necessary reciprocal causal relationship and are therefore difficult to distinguish (Burgess & Shallice, 1996).

We first postulated that OCD patients would perform poorly in the inhibition condition. Furthermore, to determine whether this possible inhibition deficit is specific to a subgroup of OCD patients, general to OCD, or even general to emotional disorders, we compared two groups of OCD patients (checkers and washers) with healthy controls and patients with social phobia. Finally, we also predicted that the inhibition deficit in OCD patients would be related to the severity of their obsessions and compulsions.