Brain structural signatures of negative symptoms in depression and schizophrenia

Abstract

Negative symptoms occur in several major mental health disorders with undetermined mechanisms and unsatisfactory treatments; identification of their neural correlates might unveil the underlying pathophysiological basis and pinpoint the therapeutic targets. In this study, participants with major depressive disorder (n = 24), schizophrenia (n = 22), and healthy controls (n = 20) were assessed with 10 frequently used negative symptom scales followed by principal component analysis (PCA) of the scores. A linear model with the prominent components identified by PCA was then regressed on gray and white-matter volumes estimated from T1-weighted magnetic resonance imaging. In depressed patients, negative symptoms such as blunted affect, alogia, withdrawal, and cognitive impairment, assessed mostly via clinician-rated scales were inversely associated with gray matter volume in the bilateral cerebellum. In patients with schizophrenia, anhedonia, and avolition evaluated via self-rated scales inversely related to white-matter volume in the left anterior limb of internal capsule/anterior thalamic radiation and positively in the left superior longitudinal fasiculus. The pathophysiological mechanisms underlying negative symptoms might differ between depression and schizophrenia. These results also point to future negative symptom scale development primarily focused on detecting and monitoring the corresponding changes to brain structure or function.

Keywords: cerebellum; depression; negative symptoms; schizophrenia; white matter.

Figure 1

Boxplots of scores in cognition/expression component and pleasure/motivation component are shown. Both patient groups are significantly different from the control group in these two components (Table 1).

Figure 2

Neural signatures of negative symptoms in schizophrenia and depression groups are shown. Significant association of white-matter volumes with the principal component coding for “pleasure/motivation” was found in the schizophrenia group: (A) left superior longitudinal fasiculus (SLF) with cluster size = 105 voxels and peak p  = 0.017. (B) Left anterior limb of internal capsule/anterior thalamic radiation with cluster size = 114 voxels and peak p  = 0.014. (C) Significant association of gray matter volumes with the principal component coding for “cognition/expression” was found in the depressed group: cerebellum with a cluster of size = 2212 voxels and peak p = 0.002. Peak voxels shown with corresponding Montreal Neurological Institute (MNI) coordinates. Scatter plots: controlling for age, gender, total gray or white-matter volume, partial correlation coefficients (r) between the extracted volumes and principal component scores are calculated within group and are shown here [(A): r = 0.85, (B) r = −0.84, (C) r = −0.79]. Please note that the partial correlation coefficients are expected to be high as the regions in which we quantify that these correlations were defined as being regions with significant associations between volume and principal component expression.

Figure 3

Inter-regional correlation in schizophrenia group. Linear regression was performed with a dependent variable of extracted volumes in the left superior longitudinal fasiculus or left anterior limb of internal capsule/anterior thalamic radiation and covariates of age, gender, and total white-matter volumes. Pearson correlation coefficient between the residuals of extracted volumes from these two regions is r = −0.66 with p < 0.001 in the schizophrenia group.