Are relational style and neuropsychological performance predictors of social attributions in chronic schizophrenia?

Abstract

Attributional style is defined as the pervasive tendency to explain the cause of social actions in terms of oneself, or others, or the context of the event. While the clinical correlates of this aspect of social cognition have been widely researched, its links with relationship style and neuropsychological performance, although hypothesised, have received less attention. This study investigated whether attributional style is predicted by variance in either relationship style or neuropsychological performance in schizophrenia. We assessed attributional style (using the Internal, Personal and Situational Attributions Questionnaire [IPSAQ]), relationship style (using Bartholomew and Horowitz's Relationship Questionnaire), and neuropsychological function (using the Wechsler Abbreviated Scale of Intelligence, athe Wechsler Memory Test, and the Cambridge Automated Test Battery) in 73 stabilised outpatients with chronic schizophrenia and 78 controls matched for age and gender. ‘Externalising bias’ (attributing positive rather than negative events to oneself) was predicted by verbal ability in both patients and controls. ‘Personalising bias’ (attributing negative events to others rather than to situational factors) was predicted by higher secure relationship style ratings, but only in the patient group. This study highlights the importance of relationship style and neuropsychological performance for different aspects of attributional style in schizophrenia.

Introduction

Social cognition – the ability to construct representations of the relations between ourselves and others (Adolphs, 2001) – is a key requirement for interacting flexibly in interpersonal situations. In schizophrenia, social cognition has been parsed into the separate (but overlapping) abilities of affect recognition (Penn et al., 1996), theory of mind (Frith, 1992), social knowledge (Corrigan et al., 1992), and attributional style (AS) (Kinderman et al., 1998). Among these, AS has been defined as the pervasive tendency to explain the cause of social actions either in terms of internal or external factors (i.e. attributing the cause of events to oneself, others, or situational factors; Kaney and Bentall, 1989, Kinderman and Bentall, 1996a, Martin and Penn, 2002).

Much of the research into AS to date has focused on its relationship to clinical presentation. An influential model of persecutory delusions by Bentall et al. (1994) hypothesised that patients avoid activation of negative self-beliefs by making externalising and/or personalising attributions for negative events. In particular, the tendency of paranoid patients to blame others, rather than the situation, for negative events has been described as a “Personalising Bias” (PB). Both this bias and an “Externalising Bias” (EB), the tendency to over-attribute positive rather than negative events to oneself, have been hypothesised as protective of self-esteem, particularly during difficult psychological experiences (Kinderman and Bentall, 1996b, Bentall et al., 2001, Whaldheter et al., 2005).

In addition to symptom severity, a recent review by Berry et al. (2007) suggests that AS may also be influenced by attachment or relational style in schizophrenia. Relationship style is defined in terms of an enduring cognitive affective model for interpreting and responding to others developed through early relational experience (Bowlby, 1982). Studies in the general population suggest that individuals with “insecure” style (i.e. avoidant, pre-occupied, or dismissing; Bartholomew, 1990) are more likely to attribute negative events to personal factors (either in oneself or in others) than to situational factors (Collins, 1996). Since insecure relationship styles are frequently reported in psychosis (Dozier, 1990, Dozier et al., 1991), Berry et al. (2007) argue that these are likely to be reflected in patients' attributional biases. This is important because it would highlight an aspect of AS associated with relational experience rather than illness subtype or severity. We are not aware of any studies to date which have empirically investigated this relationship between AS and relational style in schizophrenia.

Adequate investigation of the neuropsychological correlates of AS is also lacking in schizophrenia. This is surprising given the number of neuropsychological studies of other aspects of social cognition (e.g. ‘theory of mind’; Corcoran and Frith, 2003, Bora et al., 2006, Fiszdon et al., 2007, Shamay-Tsoory et al., 2007). In the non-schizophrenia literature, attributional style is typically described in terms of both automatic (perceptual) and controlled (cognitive) processes (Gilbert and Malone, 1995, Lieberman et al., 2002). Studies in non-clinical samples suggest that adequate consideration of situational factors requires greater cognitive control, such that situational attributions are decreased under conditions of increased ‘cognitive load’ (Gilbert et al., 1988). Based on this evidence, Bentall et al. 's (2001) revised model of attributional biases in schizophrenia suggested that patients with more pronounced attentional deficits would demonstrate greater difficulty in adequately attending to situational factors, consideration demonstrated by higher PB scores.

To date, we are aware of only two patient studies which have investigated the relationship between AS and neuropsychological performance. A study by Randall et al. (2003) failed to detect an association with attentional control, although the sample size was small. A first episode psychosis study by Krstev et al. 's (1999) reported a relationship between higher verbal IQ and lower EB scores, but not PB scores (attentional control was not considered separately). Instead of reflecting the specific effects of ‘attentional’ control as suggested by Bentall's theory, a correlation with verbal ability may suggest an alternative hypothesis: that some attributional biases (e.g. EB) represent a more cognitively ‘simplistic’ AS, reflected by an association with lower general cognitive ability. Further investigation of this relationship between AS and both general and domain specific aspects of neuropsychological performance is required to evaluate these alternative hypotheses. Clarifying this relationship is further warranted on the grounds that AS has been suggested as an aspect of social cognition that may be targetable for drug treatment, independent of more basic aspects of cognition (Green et al., 2005), despite the overlapping brain regions likely to be involved in both processes (Lieberman et al., 2002).

This study investigated the relationships between AS, relationship style and neuropsychological performance in both patients with schizophrenia and normal controls, based on four main hypotheses. First, following the review by Berry et al. (2007) we tested the hypothesis that a bias towards making personal rather than situational attributions for negative events (PB) would be predicted by higher insecure ratings. As the corollary of this hypothesis, our second hypothesis was that more secure relationship ratings would predict lower PB scores and higher situationally based attributions for negative events. For neuropsychological performance, we also tested two hypotheses. To test whether a bias towards attributing more positive than negative events to oneself (EB) represented a cognitively simplistic AS, our third study hypothesis was that EB scores would be predicted by poorer general neuropsychological performance. Finally, following Bentall et al.'s (2001) model, our fourth hypothesis was that poorer attentional control would predict lower situationally based attributions for negative events (SN).