MRI textural plasticity in limbic gray matter associated with clinical response to electroconvulsive therapy for psychosis

Abstract

Electroconvulsive therapy (ECT) is effective against treatment-resistant psychosis, but its mechanisms remain unclear. Conventional volumetry studies have revealed plasticity in limbic structures following ECT but with inconsistent clinical relevance, as they potentially overlook subtle histological alterations. Our study analyzed microstructural changes in limbic structures after ECT using MRI texture analysis and demonstrated a correlation with clinical response. 36 schizophrenia or schizoaffective patients treated with ECT and medication, 27 patients treated with medication only, and 70 healthy controls (HCs) were included in this study. Structural MRI data were acquired before and after ECT for the ECT group and at equivalent intervals for the medication-only group. The gray matter volume and MRI texture, calculated from the gray level size zone matrix (GLSZM), were extracted from limbic structures. After normalizing texture features to HC data, group-time interactions were estimated with repeated-measures mixed models. Repeated-measures correlations between clinical variables and texture were analyzed. Volumetric group-time interactions were observed in seven of fourteen limbic structures. Group-time interactions of the normalized GLSZM large area emphasis of the left hippocampus and the right amygdala reached statistical significance. Changes in these texture features were correlated with changes in psychotic symptoms in the ECT group but not in the medication-only group. These findings provide in vivo evidence that microstructural changes in key limbic structures, hypothetically reflected by MRI texture, are associated with clinical response to ECT for psychosis. These findings support the neuroplasticity hypothesis of ECT and highlight the hippocampus and amygdala as potential targets for neuromodulation in psychosis.

Fig. 1. Significant group-time interactions in the normalized gray-level size zone matrix (GLSZM) large area emphasis of the left hippocampus and the right amygdala.

Group-time interactions in the normalized gray-level size zone matrix (GLSZM) large area emphasis in (A) the left hippocampus and (B) the right amygdala in patients treated with both electroconvulsive therapy (ECT) and medication (blue) and patients treated with medication only (gray) are shown. The rainclouds represent the distribution of the normalized GLSZM large area emphasis in (A) the left hippocampus and (B) the right amygdala at each time point in the ECT group (blue) and the medication-only group (gray), respectively. Lines in lighter shades indicate data from individual participants in the ECT group (blue) and the medication-only group (gray). ECT electroconvulsive therapy, GLSZM gray level size zone matrix.

Fig. 2. Statistically significant repeated-measures correlations between clinical variables and normalized gray-level size zone matrix (GLSZM) large area emphasis of the left hippocampus and the right amygdala in patients treated with both electroconvulsive therapy and medication.

Repeated-measures correlation coefficients and false discovery rate (FDR)-adjusted p values are shown. The blue dots represent data before ECT; the gray dots represent data after ECT. PANSS Positive and Negative Syndrome Scale, CGI-S Clinical Global Impression Severity, FDR False Discovery Rate, GLSZM gray level size zone matrix.