Widespread electrical cortical dysfunction in schizophrenia

Abstract

The purpose of this study was to compare slow cortical electrical activity between healthy and schizophrenic individuals using 123-channel EEG and current density reconstruction (CDR). Twenty-nine healthy subjects and 14 drug-free patients performed three visual paired-associate tasks (verbal, pictorial and spatial). We modeled the generators of the slow potentials (SPs) at their peak amplitude by Lp-norm minimization using individual MRIs to model the volume conductor and source. Activity in each architectonic area of Brodmann was scored with respect to individual maximum current by a percentile method. Resulting scores by cortical area were analyzed by multivariate analysis of variance (MANOVA) with planned comparisons, to search for differences among levels. Results showed a multifocal pattern of current density foci comprising the SP generators, including frontal and posterior cortices in all subjects. A few cortical areas, not exclusively frontal, were observed to significantly differ between groups. Moreover, changes in patients' frontal activity were not exclusively to lower scores or ‘hipofrontality’: overall effects (all tasks collapsed) included increased electrical activity in right area 10, left 38 and 47 bilaterally, and decreased activity in right area 6 and left areas 39, 21 and 19. A few additional areas showed significantly altered activity only in particular tasks. We conclude that the present method, by preserving individual anatomical and functional information, indicates bidirectional patterns of altered electrical activity in specific cortical association areas in schizophrenia, which are not compatible with the exclusive ‘hipofrontality’ hypothesis. Our results agree with the hypothesis of schizophrenia as a syndrome resulting from abnormalities in multiple encephalic foci.

Introduction

The view that schizophrenia consists of a syndrome associated with possibly unspecific pathology of multiple interconnected regions of the encephalon is becoming prevalent in the literature. It appears that multiple, including genetic, developmental and acquired factors, contribute to the pathological alterations (for reviews, see Arnold, 1999, Arnold and Trojanowski, 1996, Powers, 1999), which on their turn appear sufficient to generate the core symptoms of schizophrenia. Thus, neuropathological alterations have been found in various cortical association areas pertaining to the prefrontal, temporal and parieto-temporo-occipital junction or the corresponding territories in the basal ganglia and thalamus, in addition to the dopaminergic projection cells of the ventral tegmental area. The syndrome would be a ‘final common pathway’ resulting from of such alterations, whose exact distribution would account for the particular subsyndromes of schizophrenia, being also reflected in brain metabolism (Schröder et al., 1996). Although originally claimed to be a general feature of schizophrenia, functional alterations of the prefrontal cortex in the form of ‘hypofrontality’ are now challenged and more likely to be conditions that reflect only the negative symptoms syndromes Schröder et al., 1996, Manoach, 2003. In some recent functional studies where active engagement in experimental tasks are required by patients, even opposite effects, hyperactive frontal lobes, may be observed (reviewed by Manoach, 2003).

Slow potentials (SPs) are event-related electromagnetic variations whose sufficient condition for occurrence is the mere anticipation of stimuli and can thus be recorded during time intervals (up to many seconds) that precede expected stimuli. Such stimuli may not even occur, the tasks that provoke SPs do not require motor responses Ruchkin et al., 1986, Bocker et al., 1994, and more important for clinical applications, some forms of SPs are not related to task performance itself but to preparatory processes such as selective attention or the mere attempt to perform the task Basile et al., 2002, Basile et al., 2003, Basile et al., 2004. SPs represent direct correlates of anticipatory attention, as opposed to the endogenous evoked potentials, which have anticipatory attention as an antecedent condition and are only modulated by attention. Since the discovery of the prototypical slow potential, the contingent negative variation (CNV, Walter et al., 1964; for a review, see McCallum, 1988), SPs were suspected to be originated in the prefrontal cortex Walter and Crow, 1964, Fuster, 1989. However, we have recently provided evidence, in healthy individuals, that slow potentials are generated not only by prefrontal cortical areas, but by an individually variable set of cortical association areas including posterior cortices as well Basile et al., 2002, Basile et al., 2003, Basile et al., 2004. A vast literature covers EEG and ERP alterations in schizophrenia (spectral, and in amplitudes, latencies and topography of ERPs), but most studies do not allow inferences about (cortical) topography, mainly due to the small number of electrodes and the use of an active reference electrode (the mastoids or ear lobes, which are sensitive to temporal lobes). Among SPs, classical CNVs and the post-imperative negative variation (PINV) are well established as altered in amplitude and topography, and the first may even be treated as a stable marker of the condition (Verleger et al., 1999). A PINV-related potential (SPs recorded after a button press, in anticipation of performance feedback stimuli), corresponding to the test of the memorization of the paired-associates here described, is profoundly altered in topography in patients and is described elsewhere (Basile et al., 2004).

Recently, we used three paired-associates memorization tasks (verbal, pictorial and spatial) and preserved individual results, in the form of electrical activity in each cytoarchitectonic area, prior to group analysis for verification of task-specific activity Basile et al., 2003, Basile et al., 2004. Briefly, the choice of those three classes of stimuli was based on traditional or well-founded anatomical distinctions: given that we had used 29 right-handed subjects, we expected the engagement in the verbal task to correspond to verbal-related hemispheric asymmetries (as compared to the nonverbal tasks), and the use of pictorial versus spatial material among the nonverbal domain was based on the anatomical distinction between the dorsal and ventral visual systems, known to be extended to the prefrontal cortices, with predominant connectivity between particular prefrontal areas and either visual stream (Pandya and Yeterian, 1990). SP generator analysis could thus aid in the functional mapping of association cortex, in an individual-based and independent fashion (from metabolic tracing, e.g., fMRI and PET), and could serve as an additional tool in the pathophysiology of psychiatric conditions. In the present work, we used the same three tasks and modern method of current density reconstruction (CDR) calculated on individual cortical anatomy, for the comparison of the generators of SPs obtained in a group of unmedicated schizophrenic patients with those from a group of healthy control subjects. In particular, we intended to verify whether SP generators would indicate hypo- or hyperactivity in prefrontal cortical areas. Given that our tasks, as all tasks involving categorization or memorization, require selective attention as a minimum correlate of engagement (regardless of correct performance), and that cortical (particularly prefrontal) dysfunction impair all those aspects of performance, we expected them to be unspecifically sensitive to any abnormality in patients. That is, as a first approach, the tasks used would maximize the chances of finding some electrophysiological abnormality in patients, even if pinpointing to a restricted psychological correlate were not possible.