Is aberrant functional connectivity a psychosis endophenotype? A resting state functional magnetic resonance imaging study

Abstract

Background: Schizophrenia and bipolar disorder share overlapping symptoms and risk genes. Shared aberrant functional connectivity is hypothesized in both disorders and in relatives.

Methods: We investigated resting state functional magnetic resonance imaging in 70 schizophrenia and 64 psychotic bipolar probands, their respective first-degree relatives (n = 70 and 52), and 118 healthy subjects. We used independent component analysis to identify components representing various resting state networks and assessed spatial aspects of functional connectivity within all networks. We first investigated group differences using five-level, one-way analysis of covariance (ANCOVA), followed by post hoc t tests within regions displaying ANCOVA group differences and correlation of such functional connectivity measures with symptom ratings to examine clinical relationships.

Results: Seven networks revealed abnormalities (five-level one-way ANCOVA, family-wise error correction p < .05): A) fronto-occipital, B) midbrain/cerebellum, C) frontal/thalamic/basal ganglia, D) meso/paralimbic, E) posterior default mode network, F) fronto-temporal/paralimbic and G) sensorimotor networks. Abnormalities in networks B and F were unique to schizophrenia probands. Furthermore, abnormalities in networks D and E were common to both patient groups. Finally, networks A, C, and G showed abnormalities shared by probands and their relative groups. Negative correlation with Positive and Negative Syndrome Scale negative and positive scores were found in regions within network C and F respectively, and positive correlation with Positive and Negative Syndrome Scale negative scores was found in regions in network D among schizophrenia probands only.

Conclusions: Schizophrenia, psychotic bipolar probands, and their relatives share both unique and overlapping within-network brain connectivity abnormalities, revealing potential psychosis endophenotypes.

Keywords: Bipolar; endophenotype; relatives; resting state; schizophrenia; within-network connectivity.

Figure 1

One-way ANCOVA main effect of group difference (p < 0.05, Family-wise error corrected, cluster threshold k = 10 voxels): (A) fronto-occipital, (B) cerebellum/midbrain, (C) fronto-limbic, (D) meso/paralimbic, (E) posterior default mode network, (F) fronto-temporal/paralimbic and (G) sensorimotor network. All the brain slices are in transverse view with their corresponding MNI slice in mm.

Figure 2

Bar plot of mean loading coefficients in HC, BP, SZ, BP-Relative (BP-REL) and SZ-Relative (SZ-REL) in fronto-occipital network: A(i) lingual gyrus with peak at (26, -46, -4), A(ii) cuneus with peak at (-8, -78, 4), cerebellum/midbrain network: B(i) left pons with peak at (-10, -38, -40), B(ii) right pons with peak at (12, -38, -38), fronto-limbic network: C(i) right thalamus with peak at (12, -14, 10), C(ii) right putamen with peak at (22, 12, -2), meso/paralimbic network: (D) uncus/superior temporal gyrus with peak at (-30, 6, -30), posterior default mode network: (E) posterior cingulate gyrus/precuneus with peak at (2,-54, 30), fronto-temporal/paralimbic network: (F) medial temporal gyrus/inferior temporal gyrus with peak at (-56, -10, -24), sensorimotor network: G(i) superior frontal gyrus/medial frontal gyrus with peak at (14, 22, 50), G(ii) medial frontal gyrus/precentral gyrus with peak at (22, -18, 64) and G(iii) supplementary motor area/superior frontal gyrus with peak at (12, 2, 70). Coordinates reported here are in MNI. Error bar represents standard error of mean. *Groups with significant aberrant functional connectivity when compared to HC (p < 0.0125). BP: Bipolar probands, SZ: Schizophrenia probands, BP-REL: Relatives of bipolar probands, SZ-REL: Relatives of schizophrenia probands, HC: Healthy controls.