Impulsivity in depressed children and adolescents: A comparison between behavioral and neuropsychological data

Abstract

Impulsivity at the neuropsychological and behavioral levels was investigated in a sample of drug-naive depressed children and adolescents. The performance of 21 patients with a current diagnosis of mood disorder was compared with that of 21 normal controls on tests of executive functions related to impulsivity (Matching Familiar Figures Test, Continuous Performance Test, Verbal Fluency, Stroop Test, and Walk–Don't Walk) and on impulsive/restless behavior on the Conners' Parent Rating Scale. Depressed children and adolescents showed a pattern of conservative response style, with slow reaction times and attentional problems, similar to that observed in adults, and a general delay/difficulty in response initiation on the Fluency Test. Depressed participants were rated by their parents as being significantly more impulsive/restless than controls. However, there was no evidence of an impulsive cognitive response style in more impulsive/restless patients. Symptom severity (Hamilton Rating Scale for Depression) and subjective mood state (Children's Depression Inventory) were also taken into account.

Introduction

Neuropsychological functioning in depressed patients is generally described as being less than optimal, suggesting a relationship between depressive symptoms and slow processing, impaired executive functions (e.g., set shifting, attention) in effortful tasks and relatively intact automatic processing (Hasher and Zacks, 1979, Purcell et al., 1997, Degl'Innocenti et al., 1998, Fossati et al., 2002). A neuropsychological impairment of this type has been described mainly in adults, who often show a pattern of symptoms that is slightly different from that in children. Although there are more similarities than differences between younger and adult depressed patients, very young patients are more likely to express distress and negative affect by externalizing symptomatology, probably because of their still immature language capabilities (Kolvin and Sadowski, 2001), whereas vegetative symptoms (e.g., weight loss or gain, insomnia or hypersomnia) and endogenous-type symptoms (e.g., melancholia and suicidal attempts) seem to increase with age (Angold and Costello, 1993, Birmaher et al., 1996). Furthermore, a strong association has been found in children and adolescents between mood disorders and disruptive behaviors as measured by the Aggressive Behavior and Delinquent Behavior subscales of the Child Behavior Checklist (Kasius et al., 1997), which identify proneness to anger, impulsivity and low self-regulation in children (Eisenberg et al., 2001). Furthermore, in the general population, children are usually more impulsive than adults, both at the behavioral (e.g., tendency to blurt out answers, inability to wait for their turn) and cognitive levels (Hasher and Zacks, 1979). According to some studies, depressed/dysphoric children have a more impulsive cognitive style than controls (Palladino et al., 1997, Staton et al., 1981). Therefore, our interest is in the relationship between impulsivity/restlessness in depressed children and adolescents and the pattern of slow processing and impaired executive functions described mainly in adult depressed patients. Moreover, it is still not clear if in adult depressed patients, the neuropsychological impairment is affected by severity of symptoms (Austin et al., 1992, Boone et al., 1995, Elliott et al., 1996, Merriam et al., 1999, Koetsier et al., 2002). Depression in children and adolescents is an ideal field in which to explore these questions because it is, by definition, an early-onset depression, and therefore there are no effects of aging. In the present article, we focus on performance on some measures of executive function related to impulse control, inhibition and response initiation, namely the Matching Familiar Figures Test (MFFT), the Verbal Fluency Test, the Walk–Don't Walk Test, the Continuous Performance Test (CPT), and the Stroop Test. The aims of the present study are (a) to provide a neuropsychological profile of drug-naïve depressed children and adolescents on measures related to neuropsychological and behavioral impulsiveness compared with healthy controls (we focus on tests that either have given inconsistent results or have not been previously studied in pediatric depression); (b) to investigate the relationship between executive functions and severity of depression, subjective mood state and behavioral impulsivity/restlessness, and (c) to test the hypothesis that depressed children and adolescents with higher rates of externalized and impulsive behaviors are also more impulsive in their cognitive style.

The MFFT (Kagan et al., 1964, Cairns and Cammock, 1978) is the most widely used test to evaluate cognitive style, and it requires the subject to decide on the similarity between a standard figure and six variants, five of which are similar and one identical to the standard one. Cognitive style is defined according to two parameters: the total number of errors made before the correct choice (Accuracy), and the mean time taken by the subjects to decide on the first response (Latency). Cognitive style has been described as the preferential mode of processing information a subject manifests in a large variety of tasks, and it is defined by conceptual tempo and number of errors in performing cognitive skills. Based on performance on the MFFT, subjects can be classified as “reflective”, with longer response latency and fewer errors; “impulsive”, with fast response times and higher error rate; “fast and accurate”; and “slow–inaccurate”. Furthermore, as noted by Keller and Ripoll (2001), cognitive style is best defined as the ability to adapt response latency to the task request. Under time constraints, reflective children are faster than impulsive children because they can adapt the response time to the context and thus be more efficient at problem solving. Therefore, these authors suggest that the MFFT Reflective–Impulsive taxonomy does not predict a generalized impulsive style in all types of tasks, especially when time constraints are introduced. In fact, the slow–inaccurate classification on the MFFT seems to better predict impulsivity and inefficient performance on other cognitive measures than the impulsive classification (Victor et al., 1985).

When cognitive style is explored in relation to depressed mood, results are very inconsistent: some studies found short response time latencies in depressed/dysphoric children (Staton et al., 1981, Palladino et al., 1997), while other studies found longer latencies (Schwartz et al., 1982). Last, no differences between depressed and healthy controls were found in Fuhrman and Kendall (1986) and Kendall et al. (1990). Differences were probably due to selection criteria and group size: in some studies dysphoric children were selected through rating scales (Palladino et al., 1997, Schwartz et al., 1982, Fuhrman and Kendall, 1986) and no formal diagnosis was made through clinical assessment. In the study of Staton et al. (1981), the sample of depressed children was small (9 depressed children), there was no control group, and a comparison was only made between performance before and after pharmacological treatment. Finally, in the study of Kendall et al. (1990), 17 sixth graders who were diagnosed as depressed based on a diagnostic semi-structured interview (K-SADS) were compared with healthy controls. To disentangle this inconsistency and examine the cognitive style of depressed children and adolescents, we administered the MFFT (the MFFT-20 version) to a larger sample of drug-naïve patients with a DSM-IV diagnosis of depression.

The Verbal Fluency Test has been related to response initiation, access to semantics, cognitive strategy and attention, and it is particularly sensitive to depressive mood (Ravnkilde et al., 2002, Fossati et al., 2003). Deficits in semantic fluency in depression have been reported in several studies (Tarbuck and Paykel, 1995, Fossati et al., 2003), while conflicting results have been found on phonemic fluency. In the present study, we focused on response initiation to explore the relationship between behavioral impulsiveness and response initiation, but we also looked for differences between semantic and phonemic fluency in children and adolescents with depression.

The Walk–Don't Walk Test is taken from the Test of Everyday Attention for Children (TEA-Ch; Manly et al., 2001). It measures auditory attention and impulsive motor response to an auditory stimulus. To our knowledge, it has not previously been used in studies on depressed patients.

The Stroop Test has been related to attention and effortful processing (Cohen et al., 1990) because of its requirement to counteract an automatic, interfering response. The Stroop Test was chosen to test the hypothesis that more impulsive/hyperactive children and adolescents might also be less able to inhibit interference.

Finally, the Continuous Performance Test (CPT) is an attention task that has been widely used in psychiatric research. It provides a measure of vigilance (sustained attention over a long period to incoming stimulation). In the CPT, the subject monitors a continuous presentation of stimuli and reacts as fast as possible to the occurrence of a critical target (Conners, 1995). The computerized system allows for recording of omissions (number of non-responses to targets), commissions (number of responses to non-targets), and mean reaction times. Some studies have shown that depressed patients generally have a conservative response style characterized by increased reaction times and more omissions than healthy controls. Koetsier et al. (2002) found significantly impaired performance on the CPT in depressed, drug-free adult patients; they also found a significant relationship between performance and subjective mood state and a non-significant correlation to severity of depression, as measured by the Hamilton Rating Scale for Depression (Ham-D; Hamilton, 1967). Findings in adults show significant attentional impairment, but not impulsivity, as indicated by short reaction times and a high number of commission errors.