Children’s “catastrophic Responses” to Negative Feedback on Cantab’s Id/ed Set-shifting Task: Relation to Indices of a Depressive Temperamental Style

This study assessed whether a specific form of motivational impairment, over-sensitivity to perceived failure, described previously in depressed adults, would impact healthy children’s performance on neuropsychological tasks as a function of individual differences in negative affectivity. Healthy children completed the Cambridge Neuropsychological Test and Automated Battery (CANTAB) as part of a large-scale study of cognitive development. The tendency to respond to perceived failure with subsequent item failure was calculated for each child on the basis of his/her performance on the CANTAB’s ID/ED set-shifting task, during which trial-by-trial feedback is provided. Children were divided into those with low versus high tendencies to react to failure with subsequent item failure. One year later, their parents completed child temperament ratings, using Rothbart’s Child Behavior Questionnaire (CBQ) as part of another study where cortisol was also measured. Children with increased tendencies to exhibit heightened responses to failure on the ID/ED set-shifting task were rated by their parents as higher in sadness and slightly higher in overall Negative Affectivity. These children performed worse on several CANTAB subtests, including memory span, pattern recognition memory, and set shifting. Relations to daily cortisol rhythms were examined but did not yield strong effects. Abnormally sensitive responses to negative feedback have been discussed as a trait marker of affective disorder. These findings suggest that this motivational style might impact children as a function of temperamental characteristics that might lead to vulnerability to later internalizing psychopathology. Findings are discussed in relation to neurobehavioral * Corresponding address: Monica Luciana, Department of Psychology, 75 East River Road, University of Minnesota, Minneapolis, MN 55455, USA Email: [email protected] Iunie 2005 • Cogniţie, Creier, Comportament 343 models of feedback processing in affective disorder and developmental psychopathology. KEY-WORDS: temperament, depression, cognition, negative affect, children Cognitive theories of depression and anxiety are among the most influential psychological models advocated by clinicians to explain internalizing disorders (Beck, 1972; 1979; Crews & Harrison, 1995). These models posit that depression and anxiety are exacerbated by negative self-directed attributions that bias the manner in which vulnerable individuals perceive themselves, their interactions with others, and outcomes of their behaviors across situations. They also imply that individuals who have been previously depressed, but are currently remitted, remain vulnerable to relapse due to the trait-like stability of negative biases, which are automatic in nature, involuntary, and perseverative across situations (Beck, 1972). Cognitive-behavioral interventions based on these theories are perhaps the most successful non-pharmacological treatments for both depression and anxiety, particularly in children and adolescents (Compton et al., 2004). Cognitive functions have been examined from various perspectives in internalizing psychopathology through the use of psychological inventories (e.g., the Beck Depression Inventory: Beck, Ward, Mendelsohn, Mock, & Erbaugh, 1961) and through direct measures of discrete abilities as measured by neuropsychological tests (reviewed by Elliott, Sahakian, Herrod, Robbins & Paykel 1997). These investigations have revealed, on the one hand, that depressed individuals tend to feel pessimistic or hopeless about the likelihood of positive events occurring in the future (Abramson, Seligman & Teasdale, 1978; Crews & Harrison, 1995). On the other hand, brain-based evaluations of cognition have revealed the presence of neuropsychological deficits in individuals with depression and/or anxiety (George et al., 1997; Trichard et al., 1995). These deficits include impairments in attention, memory, and executive functions under conditions where high effort is demanded. Despite some degree of consistency in the types of impairments reported, no single neurocognitive dysfunction has been described to characterize all individuals with internalizing psychopathology, and links between these deficits and negative self-attributional biases that have traditionally been conceptualized as more “psychological” in nature have not been proposed. Recently, it has been suggested that the general affective demeanor of anxious and/or depressed individuals might undercut their performance across structured cognitive tasks regardless of task content, perhaps contributing to failures observed during neuropsychological testing, especially when direct feedback is provided to participants in a trial-by-trial manner. For example, Beats, Sahakian and Levy (1996) administered a computerized version of the Tower of London planning task to elderly depressed individuals and to healthy controls. It was observed that although the groups were equivalent in their overall levels of Children’s responses to negative feedback 344 perfect problem solutions, individuals in the depressed group tended to deteriorate in their efficiency of performance after making mistakes within a given problem. This basic effect was replicated and extended by Elliott et al. (1996) who administered the Tower of London and a Delayed Match to Sample task to depressed adults versus controls and found that depressed individuals not only performed worse than control subjects overall, but they also exhibited decrements in performance following error trials. Notably, this performance pattern was specific to depressed individuals and was not evident in individuals with other psychiatric disorders (i.e., schizophrenia) or in those with neurological impairments (i.e., individuals with Parkinson’s disease or temporal lobe lesions: Elliott et al., 1997). Moreover, previously but not currently depressed individuals who were tested while they were in states of recovery also showed oversensitivity to perceived failures (Elliott et al., 1997) suggesting that the phenomenon is trait versus state-based. Elliott et al. (1996, 1997) have therefore suggested that oversensitivity to perceived failure is a motivationally-driven trait-based disposition that renders depressed individuals vulnerable to cognitive dysfunction when cognition is measured through the use of structured neuropsychological tasks. These findings suggest novel methodologies for assessing cognitionemotion interactions in the laboratory, because they call for a focus on the motivational processes that individuals use in the course of problem-solving versus strict adherence to categorical evaluations of performance (i.e., correct versus incorrect scores on given problems). It may be that a qualitative analysis of motivational processes is more predictive of which individuals are vulnerable to specific types of pathology. It might also be the case that dysfunctions in common brain regions account for both the distorted responses that depressed and anxious people exhibit following perceived failure as well as their difficulties with attention, memory, and executive function measures. In order to begin to address this issue, it is important to establish an association, on a behavioral level, between “catastrophic” responses and/or distorted cognition in the course of testing, measures of depressogenic traits, and neuropsychological performance. When applied to developmental samples, this type of analysis may contribute in unique ways to our understanding of internalizing disorders and their etiology. Prospective developmental studies would allow us to demonstrate that high-risk individuals who are vulnerable to pathology, but who haven’t yet expressed it, nonetheless demonstrate biased cognition. However, similar but more subtle relations might also characterize healthy (apparently low-risk) individuals who vary in temperamental traits thought to contribute to internalizing tendencies. These traits would include diminished levels of positive affect, which characterizes individuals with depressive tendencies, and high levels of negative affect, which characterizes both individuals with depression as well as those with anxiety (Clark, Watson & Mineka, 1994; Watson & Clark, 1995; Watson & Tellegen, 1985). Accordingly, this project sought to apply the general findings of Elliott and colleagues (1996, 1997) to a developmental sample of children who were free of Iunie 2005 • Cogniţie, Creier, Comportament 345 psychopathology. A low risk sample of children was tested using selected subtasks from the CANTAB neuropsychological testing battery (Sahakian & Owen, 1992), and their responses to measures where performance feedback was provided were tabulated in order to identify individual children who showed a relative tendency to respond to failure with subsequent failure (i.e., exhibiting a so-called “catastrophic response to failure”: Beats et. al., 1996). This tendency was then examined in relation to each child’s temperament as reported by a parent, with a specific focus on aspects of temperament (negative affectivity) that would suggest vulnerability to either anxiety or depression. As an additional dimension of temperament, we also examined whether children high in “catastrophic” responses to failure exhibited high levels of salivary cortisol, a recognized marker of stress-reactivity (Stansbury & Gunnar, 1994). It has been reported recently that generation of a highly variable cortisol rhythm from day to day or within a single day might be a risk factor for internalizing disorders (Halligan, Herbert, Goodyer, & Murray, 2004; Goodyer, Herbert, Tamplin, & Altham, 2000; Peeters, Nicholson, &