Attentional network task performance in patients with schizophrenia–spectrum disorders: Evidence of a specific deficit

Abstract

Although attentional deficits are frequently displayed by schizophrenia and spectrum patients, the precise nature of the impairment is unclear. The present study investigates attentional performance in 26 schizophrenia–spectrum outpatients and 24 healthy controls using the Attentional Network Task (ANT). We assessed the efficiency of the segregated alerting, orienting, and executive control networks by measuring how response latencies were influenced by alerting cues, spatial cues, and flanking stimuli. In overall ANOVAs we found main effects of cue condition and flanker congruency. The data also revealed a significant interaction of group and flanker type, with schizophrenia–spectrum patients taking longer on average to resolve conflict. These results suggest that compared to the healthy nonpatients, schizophrenia–spectrum patients show a deficit in their executive control network. Considered together, these findings support the notion that schizophrenia–spectrum patients have a specific attentional deficit, rather than a global one. The significance of these findings is considered from both experimental and clinical perspectives.

Introduction

Attention is comprised of several different functions, including orienting, detecting signals, preparing and maintaining alertness, inhibiting competing signals, and monitoring and adapting to conflict. Rather than involving a single unified system, Posner and colleagues (Posner and Boies, 1971, Posner and Petersen, 1990) assert that attentional functioning is executed by separate independent systems. Posner and Petersen (1990) proposed a model in which attention is conceptualized as consisting of three anatomically defined brain networks, namely, the alerting, orienting, and executive control networks.

The alerting network governs the capacity to sustain an alert cognitive state, whereas the orienting network regulates the focused identification and selection of sensory stimuli. According to the Posner and Petersen (1990) model, the alerting function is associated with frontal and parietal regions of the right hemisphere. The orienting function is associated with the parietal and frontal lobes, as well as subcortical regions such as the thalamus. The executive control network orchestrates the capacity to decide among conflicting or competing responses based upon a principle or goal. Executive control has been associated with midline frontal areas (especially the anterior cingulate) and the lateral prefrontal cortex. Functional imaging studies of healthy controls conducted during the performance of the Attention Network Task (ANT) suggests that the alerting, orienting, and conflict resolution (executive) functions of attention differentially activate distinct cortical and subcortical anatomical networks (Fan et al., 2005). Such findings buttress further support for Posner and Petersen's (1990) model.

Not only is attention a network of different operations, but these components can be parsed using various experimental paradigms (Posner and Petersen, 1990, Elvevag and Goldberg, 2000). Alerting can be tested using vigilance tasks such as warning signal tasks and continuous performance tasks. Orienting functioning can be studied using visual search tasks and spatial cueing paradigms. Executive functioning can be assayed using paradigms such as the Stroop task and flanker tasks.

Attentional deficits have been consistently observed in schizophrenia patients (Elvevag and Goldberg, 2000). To date, the precise nature of the attentional impairment displayed by patients with schizophrenia is unclear. However, it is likely that the impairment is a primary deficit, rather than an epiphenomenon of medication, chronicity, or hospitalization (Elvevag and Goldberg, 2000). Parsing the aspects of attention that are impaired in schizophrenia may assist us in gleaning further insights about the underlying pathophysiology of schizophrenia (MacDonald and Carter, 2002).

Several studies have tested the functional integrity of the alerting, orienting, or executive control systems in schizophrenia patients. Early studies (cf. Zahn et al., 1963), based on broader definitions of schizophrenia, indicated that schizophrenia patients exhibit impaired performance on tasks involving responses to warning signals. These findings are consistent with the notion that schizophrenia patients have an alerting deficit. More recent studies (cf. Cornblatt and Malhotra, 2001), using the Identical Pairs version or degraded stimulus version of a Continuous Performance Test (CPT), demonstrate that patients with schizophrenia and/or schizophrenia–spectrum disorders display sustained attention deficits. Although this can be interpreted as indicative of an alerting impairment, successful performance on the more difficult forms of the CPT involves several processes, including working memory, which is an executive control function (Michie et al., 2000).

In cued reaction time tasks, subjects have to respond as quickly as possible to the appearance of a target at a peripheral location in either the left or right visual field. Spatial cues are used to direct orienting to the stimuli, though the target that follows the cue may or may not appear at the cued location. Thus, cued reaction time tasks assay attentional engagement and disengagement. Studies (cf. Moran et al., 1992, Strauss et al., 1992, Fuentes and Santiago, 1999) indicate that schizophrenia patients, like controls, display cue effects, i.e., the RT is faster in validly cued trials and longer on invalidly cued trials. It is noteworthy that some investigators (cf., Sapir et al., 2001) assert that the abnormality in schizophrenia patients' visual attention is due to a deficit in inhibitory processes, rather than a deficit in orienting.

In the flanker task (Eriksen and Eriksen, 1974), there is a central target which must be identified by a manual response. The target (X) is flanked by distractors that are either congruent (e.g. XXX) or incongruent (e.g. YXY) with the target. The flanker congruency effect is the phenomenon whereby despite attempting to ignore the distractors, the subject's reaction times are slower when the target and flanker are incongruent than when the target and flanker are congruent. There have been relatively few studies that have looked at the flanker effect in schizophrenia patients. Overall, it appears that schizophrenia patients do not differ significantly from controls in terms of displaying the effect of context (Elkins and Cromwell, 1994, Kopp et al., 1994, Kopp and Rist, 1999, Yucel et al., 2002) under conditions in which the flanker task varies on only one dimension. That is, using modified flanker tasks, e.g., using arrowheads or different colors instead of letters as flankers, the schizophrenia patients displayed shorter reaction times for the response-compatible conditions in relation to the response-incompatible conditions.

There are some findings that suggest that schizophrenia patients' performance on flanker tasks is more nuanced than early findings indicated. Kopp and Rist (1999) administered a modified flanker task to a mixed group of schizophrenia inpatients and outpatients while simultaneously obtaining electroencephalographic (EEG) recordings. The investigators were particularly interested in the error-related negativity (N_E) component of the event-related responses, a late negative-going deflection that peaks approximately 400 ms posttarget, because it is considered an index of response monitoring. Although the paranoid patients displayed the expected effect of visual context and normal error correction performance on the flanker task, compared to the controls, they also showed an attenuation of the N_E. It is possible that the behavioral task was not sufficiently sensitive to demonstrate the patients' differential deficit in terms of response monitoring. This interpretation is consistent with the intriguing findings of Yucel et al. (2002). When Yucel et al. (2002) administered a flanker task that varied on only one dimension, they observed the flanker effect in the schizophrenia patients. However, when they used a bidimensional flanker task, the schizophrenia patients displayed a lack of interference from the flanker stimuli. These findings suggest that schizophrenia patients have difficulty in resolving multiple and simultaneous response conflicts.

In summary, previous studies have assessed schizophrenia patients' attentional functioning. These studies varied in terms of the hospitalization status of the patients, the chronicity of the patients, and the experimental manipulations that were employed in their assays.

The Attention Network Test (ANT; Fan et al., 2002) was developed to assay the three attention networks, namely, alerting, orienting, and executive control, within a single experiment. The ANT is a combination of a cued reaction time task (Posner, 1980) and a flanker task (Eriksen and Eriksen, 1974). The ANT is advantageous for the study of attention in schizophrenia for several reasons. First, it is a theoretically derived test. Secondly, because it is thought to involve dissociable cognitive functions, it allows identification of a specific deficit, i.e., a reliable disturbance in a specific cognitive domain (Chapman and Chapman, 1973, Chapman and Chapman, 1978) rather than simply indicating the presence of an overall disturbance in performance across several interrelated tasks. To date, the ANT has been used in several studies of normal controls as well as in studies of individuals with various disorders, including chromosome 22q11.2 deletion syndrome (Sobin et al., 2004, Bish et al., 2005) and borderline personality disorder (Posner et al., 2002, Fertuck et al., 2005).

Recently, Wang et al. (2005) administered the ANT to a large sample of chronic schizophrenia inpatients (N = 77) and compared them with normal controls. Interestingly, the normal controls in their sample did not differ significantly from the patient group in terms of IQ scores, memory performance, or word fluency. However, the ANT data revealed that the schizophrenia patients differed significantly from the controls in terms of their conflict resolution, reflecting a selective executive attention deficit. Additionally, the patients showed a smaller albeit significant difference in terms of their orienting network scores. Altogether, the findings of Wang et al. (2005) suggest that schizophrenia inpatients display impairments of the executive and orienting networks, though the alerting network appears functionally intact.

The purpose of the present study was to assess different aspects of attention functioning in schizophrenia–spectrum patients using the ANT. We wanted to investigate whether the findings of Wang et al. (2005), which are based upon an inpatient sample, would extend to outpatients with schizophrenia. More specifically, we sought to determine whether schizophrenia and/or schizophrenia–spectrum outpatients would display generalized attentional deficits or whether they would show specific deficits limited to the executive and orienting attentional networks. Administering the ANT to a sample of schizophrenia outpatients extends the literature because the inpatients, who are more acutely ill than their outpatient peers, may display performance that is particularly fragile. Thus, it is reasonable to suspect that there might be performance differences between the outpatients and inpatients.

Given prior findings of impaired performance by schizophrenia and schizoaffective patients on measures of executive functioning such as the Wisconsin Card Sorting Test and delayed match-to-sample tasks (Park and Holzman, 1992, Gooding and Tallent, 2002, Gooding and Tallent, 2004) we hypothesized that the schizophrenia–spectrum patients would display deficits in the executive control of attention. Given Fernandez-Duque and Posner's (2001) assertion that, deficits of the executive network may interact with the parietal system to produce abnormalities of orienting, we hypothesized that the patient group might also reveal subtle deficits in the orienting network. Thus, we hypothesized that the schizophrenia–spectrum patients would display specific deficits on the ANT, limited to the executive attention and orienting networks. We did not expect the patients and healthy controls to differ in terms of their efficiency of the alerting network. Finally, we also hypothesized that the patients, like healthy controls, would display an independence of the three attentional networks.