Resting-state EEG source localization and functional connectivity in schizophrenia-like psychosis of epilepsy

Abstract

Background: It is unclear whether, like in schizophrenia, psychosis-related disruption in connectivity between certain regions, as an index of intrinsic functional disintegration, occurs in schizophrenia-like psychosis of epilepsy (SLPE). In this study, we sought to determine abnormal patterns of resting-state EEG oscillations and functional connectivity in patients with SLPE, compared with nonpsychotic epilepsy patients, and to assess correlations with psychopathological deficits.

Methodology/principal findings: Resting EEG was recorded in 21 patients with focal epilepsy and SLPE and in 21 clinically-matched non-psychotic epilepsy controls. Source current density and functional connectivity were determined using eLORETA software. For connectivity analysis, a novel nonlinear connectivity measure called "lagged phase synchronization" was used. We found increased theta oscillations in regions involved in the default mode network (DMN), namely the medial and lateral parietal cortex bilaterally in the psychotic patients relative to their nonpsychotic counterparts. In addition, patients with psychosis had increased beta temporo-prefrontal connectivity in the hemisphere with predominant seizure focus. This functional connectivity in temporo-prefrontal circuits correlated with positive symptoms. Additionally, there was increased interhemispheric phase synchronization between the auditory cortex of the affected temporal lobe and the Broca's area correlating with auditory hallucination scores.

Conclusions/significance: In addition to dysfunction of parietal regions that are part of the DMN, resting-state disrupted connectivity of the medial temporal cortex with prefrontal areas that are either involved in the DMN or implicated in psychopathological dysfunction may be critical to schizophrenia-like psychosis, especially in individuals with temporal lobe epilepsy. This suggests that DMN deficits might be a core neurobiological feature of the disorder, and that abnormalities in theta oscillations and beta phase synchronization represent the underlying neural activity.

Figure 1. Averaged eLORETA solutions (current density at cortical voxels) of EEG sources for each frequency band.

The maximum current density values for each frequency are given below the corresponding column. SLPE, schizophrenia-like psychosis of epilepsy; nPE, nonpsychotic epilepsy; L, left; R, right.

Figure 2. eLORETA statistical maps of theta oscillations.

Colored areas represent the spatial extent of voxels with significant difference (red-coded for p<0.05; yellow-coded for p<0.01, corrected for multiple testing) in source current density in psychotic vs. nonpsychotic epilepsy patients. Significant results are projected onto a fiducial cortical surface (top panel) and a brain MRI template (bottom panel). The MRI slices are located at the MNI-space coordinates indicated in the figure that correspond to the voxel of highest significance (i.e., left posterior cingulate/precuneus cortex). The color scale represents log F-ratio values (threshold: log-F = 1.03, p<0.05). L, left; R, right; A, anterior; P, posterior.

Figure 3. eLORETA wire diagram illustrating significantly increased functional connectivity (lagged phase synchronization) in the beta2 frequency band in right frontotemporal circuits in patients with schizophrenia-like psychosis of epilepsy vs. those with nonpsychotic epilepsy (t _max = 5.6; p<0.01, corrected).

The significant connectivity wires are shown inside a transparent cortical surface with axial views from the top (left) and bottom (right), as well as coronal (top middle) and right sagittal view (bottom middle). The red color of the wire indicates relative increase in phase synchronization (as opposed to a blue wire that would indicate connectivity decrease). Increased beta2 (21–30 Hz) phase synchronization was observed between signals of the Brodmann area (BA) 32 (anterior cingulate/medial prefrontal cortex) and those of BAs 32/28 (medial temporal cortex) and BA21 (middle temporal gyrus in the lateral convexity), and between BA 32 and BA 9 (dorsolateral prefrontal cortex). The points to which the lines are connected represent the center of mass of the BAs. The table at the bottom shows the phase synchronization values of the significantly different connections between groups. ^a Regions in the right hemisphere; SLPE, schizophrenia-like psychosis of epilepsy; nPE, nonpychotic epilepsy; BA, Brodmann area; ACC, anterior cingulate cortex; mPreF, medial prefrontal; DLPFC, dorsolateral prefrontal cortex.

Figure 4. eLORETA wire diagram and scatterplots of significant correlations between functional connectivity (lagged phase synchronization values) in beta2 (21–30 Hz) frequency band and psychopathology scores (i.e., positive symptoms) in patients with schizophrenia-like psychosis of epilepsy at specific regions (r = 0.79; p<0.05, corrected).

The significant connectivity wire is projected onto a transparent fiducial cortical surface. a) axial view from the top; b) sagittal view from the right; c) coronal view from the front d) axial view from the base. The points to which the lines are connected represent the center of mass of the BAs. BA32_34, connectivity between Brodmann areas 32 and 34.

Figure 5. eLORETA wire diagram and scatterplots showing a significant correlation between interhemispheric beta2 (21-30 Hz) phase synchronization of right auditory cortex and the Broca's area with auditory hallucination scores among patients with schizophrenia-like psychosis of epilepsy (r = 0.82; p<0.05, corrected).

The significant connectivity wire is projected onto a transparent fiducial cortical surface. a) axial view from the top; b) sagittal view from the left; c) coronal view from the front d) sagittal view from the right. The points to which the lines are connected represent the center of mass of the BAs. BA 42_45, connectivity between Brodmann areas 42 and 45.