Impaired Trail-Making Test-B performance in patients with acute schizophrenia is related to inefficient sequencing of planning and acting

Abstract

In order to investigate the functional basis of the frequently described deficits of schizophrenia patients in Trail Making Test B (TMT-B) performance, the performance process was analyzed based on the interaction of eye and hand movements. In a longitudinal design, 23 acute schizophrenia patients, 17 acute depressive inpatients, and 21 healthy controls were assessed twice within 4 weeks. Computer versions of both TMT-A and TMT-B were used, which require to connect different targets with a cursor, with concurrent infrared oculography. In both TMT-versions, schizophrenia patients demonstrated poorer performance and longer “planning periods”—as defined by fixations outside the cursor area—containing more fixations, which were stable over time. However, these “planning fixations” were only also longer in duration during TMT-B and differed in time and context of their occurrence compared with healthy controls. Schizophrenia patients demonstrated more fixations while the cursor rested between targets. Significant correlations with performance time gave evidence that it is important for short planning periods to be performed parallel to ongoing connection of targets, to obtain a satisfactory result. No relationship between “planning variables” with psychopathology and medication could be found. Accordingly, poorer TMT-B performance in schizophrenia patients was found to be related to insufficient sequencing of planning and acting, which appears to be a trait-like characteristic. Though depressive patients also performed poorer in TMT-B, they did not differ from either of the other groups in the main process variables, which may hint to different underlying causes of the performance deficits in both groups of patients.

Introduction

Neuropsychological research on schizophrenia patients has revealed a wide range of cognitive impairments (e.g. Heinrichs and Zakzanis, 1998), sometimes considered to be the core of schizophrenia (Elvevag and Goldberg, 2000). Beyond generalized deficits, considerable evidence points to an additional differential impairment of executive functions at least in more severely impaired patients (Robbins, 1990, Liddle & Morris, 1991, Morrison-Stewart et al., 1992, Hutton, 1998, Velligan & Bow-Thomas, 1999, Riley et al., 2000, but also see Saykin et al., 1991, Mohamed et al., 1999). Executive functions are controlled by brain systems involving the prefrontal cortex and comprise in particular the planning, sequencing or temporal structuring of behavior (e.g. Perecman, 1987, Daigneault et al., 1992, Fuster, 1999). However, executive functions draw from and rely upon many other cognitive components such as attention, perception and memory (Tranel et al., 1994) partly related to other brain areas. Accordingly, neuropsychological tests of so-called frontal lobe functions, such as the frequently used Wisconsin Card Sorting Test (WCST; Heaton, 1981), the Tower of London (Shallice, 1982), and the Trail-Making-Test (TMT; Reitan, 1958), involve several different basic functions (David, 1992, O’Donnell, 1994). Therefore, global performance measures such as performance time or error scores do not directly reveal the kind of cognitive deficit causing the observed impairment and are insufficient starting points for the investigation of brain–behavior relationships in the growing field of cognitive neuroscience.

It is therefore necessary, in order to pinpoint the underlying dysfunction and to delineate consistent functional syndromes, to develop neuropsychological paradigms more specific to psychological (dys-)functions and to perform a detailed analysis of the process of task performance, i.e. the contribution of and the synergy between functions (Levin et al., 1989),

The recording and analysis of eye (and hand) movement behavior may provide a clue to a better understanding of the performance process and its deviations in so-called frontal lobe tests: frontal structures (dorsolateral prefrontal cortex, frontal eye fields) within the “oculomotor loop” (Alexander et al., 1986) play a crucial role in the control of voluntary eye movements, being modulated via dopaminergic fibres (Gaebel and Wölwer, 1996). There is a great body of evidence pointing to dysfunctional eye movements in schizophrenia patients. Besides disturbances in basic eye movements such as an instability of fixation, dysmetric saccades (Gaebel, 1989, Holzman, 1991) and disturbances of smooth pursuit eye movements (SPEM; Clementz & Sweeney, 1990, Levy et al., 1994) there are also numerous reports regarding dysfunctional scanning behavior (e.g. Gaebel & Ulrich, 1987, Kojima et al., 1981, Kurachi et al., 1994, Phillips & David, 1997).

Given that test performance depends on the processing of continuously and externally available information, and brings about overt results, scan path analysis give clues about the mode of cognitive processing. The TMT as a visual search paradigm with a manumotor component, is a prototypical test fulfilling these criteria. However, to our knowledge, scan paths during TMT performance have never been investigated. Form A of the test requires searching and connecting irregularly arranged targets containing digits in ascending order; in form B, digits and letters have to be connected alternately. Efficient visual search, visuospatial sequencing, visuomotor integration and—in form B—executive components such as the ability to alternate between and maintain two sets of stimuli are needed to perform the task successfully (desRosiers & Kavanagh, 1987, Asarnow & MacCrimmon, 1982). More generally, sequencing and temporal structuring of behavior seem to be crucial particularly for TMT-B performance. Despite the variety of psychological functions and brain areas involved (Lorig et al., 1986), TMT-B performance seems to be most sensitive to lesions of the frontal lobe (Reitan & Tarshes, 1959, Stuss & Benson, 1986). Nevertheless, the functional specificity of the test, i.e. the kind of functional impairment underlying the performance deficit, remains unclear.

Impaired TMT-performance in schizophrenia—especially in the B-version of the test—has been reported for different phases of the illness (e.g. Gatynker & Harvey, 1992, Saykin et al., 1994, Hoff et al., 1992, Levander et al., 1985, Grawe & Levander, 1995, Goldberg et al., 1988, Levin et al., 1989, Breier et al., 1991, Liddle & Morris, 1991) as well as for healthy relatives of schizophrenia patients (Franke et al., 1993, Pogue-Gelie, 1991), and schizotypal personality disorders (Thaker et al., 1991, Trestman et al., 1995), suggesting a trait characteristic or even a status as vulnerability marker. Findings of a correlation to potential vulnerability markers such as SPEM (Grawe and Levander, 1995) may support this hypothesis, but results remain contradictory (Bartfai et al., 1985, Litman w et al., 1991, Radant et al., 1997).

Related to the obscure functional specificity, the nosologicat specificity of schizophrenia patients’ TMT-B deficit with regard to other psychiatric patients is not clear either: whereas Franke et al. (1993) found impairments in schizophrenia patients in a variety of “frontal tasks” with significant differences between schizophrenia patients and depressives only in TMT-B, others reported no significant differences between schizophrenia patients and depressives (Goldberg et al., 1993) or schizophrenia patients and alcohol dependent patients (Nixon et al., 1996). However, comparable impairments in performance time may rely on completely different functional disturbances in schizophrenia patients and other psychiatric patients.

Therefore, the aim of the present study was to provide a means for the analysis of the performance process, in order to generate hypotheses on the critical functional components in the performance process underlying schizophrenia patients’ difficulties in the TMT. For this purpose, scan paths during TMT performance were assessed using infrared-oculography. The continuous recording of both the movement of the eyes and the hand allows their interaction to be examined and for indicators critical for task performance to be extracted. Such process variables could be used in future for more differentiated analyses of the TMT-performance by combining them with additional experimental variations of the task conditions. Thus, in this first step it was of special interest, whether the poorer TMT-performance described in schizophrenia patients is due to difficulties in connecting targets themselves (e.g. more resting, accessing wrong targets, slower connection) or in the superordinate scheduling of the performance process (e.g. point of time to search targets during performance, duration to find next target, integration of searching and connecting targets during performance), or both. With respect to the hypothesis of executive dysfunctioning in schizophrenia patients, particularly the scheduling of the performance process—in the sense of planning, sequencing or temporal structuring of behavior—should be affected. To control for nosological and state/trait specificity of the process variables extracted, TMT performance in schizophrenia patients was compared with that in a depressive and a healthy control group with all subjects assessed twice within 4 weeks.