Cluster analysis reveals abnormal hippocampal neurometabolic profiles in young people with mood disorders

Abstract

While numerous studies have employed magnetic resonance spectroscopy (MRS) to determine in vivo neurometabolite levels associated with mood disorders the findings in both unipolar depression and bipolar disorder have been mixed. Data-driven studies may shed new light on this literature by identifying distinct subgroups of patients who may benefit from different treatment strategies. The objective of the present study was to utilize hierarchical cluster analysis in order to generate new hypotheses with respect to neurometabolic profiling of mood disorder. Participants were 165 young persons (18–30 yrs) with a mood disorder and 40 healthy controls. Neurometabolite levels were recorded via proton-MRS (1H MRS). The ratios (relative to creatine) of glutamate (GLU), N-acetyl aspartate (NAA) and myo-inositol (MI) measured within the hippocampus. Self-reported and clinician rated symptoms as well as cognition were also measured. The unipolar depression (N=90) and bipolar disorder (N=75) groups did not significantly differ (from each other or controls) in their levels of GLU, NAA or MI. Cluster analyses derived four subgroups of patients who were distinguished by all three metabolites. There was a pattern of positive association between NAA and GLU, whereby clusters were abnormally increased (clusters 1, 2) or normal (cluster 4) or abnormally decreased (cluster 3) in these neurometabolites. These findings suggest that there are neurometabolic abnormalities in subgroups of young people with mood disorder, which may occur despite diagnostic similarities. Such evidence highlights that the underlying neurobiology of mood disorder is complex and MRS may have unique utility in delineating underlying neurobiology and targeting treatment strategies.