Aberrant brain dynamics of large-scale functional networks across schizophrenia and mood disorder

Abstract

Introduction: The dynamics of large-scale networks, which are known as distributed sets of functionally synchronized brain regions and include the visual network (VIN), somatomotor network (SMN), dorsal attention network (DAN), salience network (SAN), limbic network (LIN), frontoparietal network (FPN), and default mode network (DMN), play important roles in emotional and cognitive processes in humans. Although disruptions in these large-scale networks are considered critical for the pathophysiological mechanisms of psychiatric disorders, their role in psychiatric disorders remains unknown. We aimed to elucidate the aberrant dynamics across large-scale networks in patients with schizophrenia (SZ) and mood disorders.

Methods: We performed energy-landscape analysis to investigate the aberrant brain dynamics of seven large-scale networks across 50 healthy controls (HCs), 36 patients with SZ, and 42 patients with major depressive disorder (MDD) recruited at Wakayama Medical University. We identified major patterns of brain activity using energy-landscape analysis and estimated their duration, occurrence, and ease of transition.

Results: We identified four major brain activity patterns that were characterized by the activation patterns of the DMN and VIN (state 1, DMN (-) VIN (-); state 2, DMN (+) VIN (+); state 3, DMN (-) VIN (+); and state 4, DMN (+) VIN (-)). The duration of state 1 and the occurrence of states 1 and 2 were shorter in the SZ group than in HCs and the MDD group, and the duration of state 3 was longer in the SZ group. The ease of transition between states 3 and 4 was larger in the SZ group than in the HCs and the MDD group. The ease of transition from state 3 to state 4 was negatively associated with verbal fluency in patients with SZ. The current study showed that the brain dynamics was more disrupted in SZ than in MDD.

Conclusions: Energy-landscape analysis revealed aberrant brain dynamics across large-scale networks between SZ and MDD and their associations with cognitive abilities in SZ, which cannot be captured by conventional functional connectivity analyses. These results provide new insights into the pathophysiological mechanisms underlying SZ and mood disorders.

Keywords: Default mode network; Energy-landscape analysis; Psychiatric disorder; Resting-state fMRI.

Fig. 1

Procedures performed in energy-landscape analysis. (A) Regions of interest from the Schafer atlas. (B) Blood oxygen level-dependent signals of seven functionally different brain networks were extracted, and (C) they were binarized by the mean of each time series. (D) The distribution of the frequency of activity patterns was fitted by a pairwise MEM model, and (E) the energy landscape was constructed from the MEM model. Abbreviation; MEM, maximum entropy.

Fig. 2

Identification of the local minimum and comparison of the energy-landscape structures. (Upper) The activity patterns of local minima identified in the energy landscape analysis. (Bottom) Disconnectivity graphs of the energy-landscape structure. Lower energy reflects more stable and more frequent occurrence of a local minimum pattern, while higher energy reflects less stable and less frequent occurrence. HC, healthy controls; SZ, patients with schizophrenia; MDD, patients with major depressive disorder.

Fig. 3

Group comparison of the duration and occurrence across brain states. The durations of states 1, 3, and 4, and the occurrence of states 1 and 2 were significantly different among the HC, SZ, and MDD groups (ANCOVA including age and sex as nuisance covariates: p < 0.001). Tukey’s multiple-comparison test was performed to investigate which pairs of groups showed significant differences: * corrected p < 0.05. HC, healthy controls; SZ, schizophrenia; MDD, major depressive disorder.

Fig. 4

Group comparison of the direct transition probabilities across the brain states. The direct transition probabilities from state 3 to state 4 and state 4 to state 3 were significantly different among the HC, SZ, and MDD groups, including age and sex as nuisance covariates (p < 0.001). Tukey’s multiple-comparison test was performed to investigate which pairs of groups showed significant differences: *corrected p < 0.05. HC, healthy controls; SZ, schizophrenia; MDD, major depressive disorder.

Fig. 5

Association between brain dynamics and the scores of BACS for SZ. The duration of state 4 was positively correlated with the verbal memory scores in the BACS, while the occurrence of state 1 and the direct transition frequency from state 1 to state 4 were negatively correlated with the verbal memory scores in the BACS. BACS, Brief Assessment of Cognition in Schizophrenia; SZ, schizophrenia.