Machine learning identified common brain network signature between depression diagnosis and symptoms of depression

Summary

  Ayumu Yamashita of Advanced Telecommunications Research Institutes International in Kyoto, Japan, and his colleagues developed a brain network marker for symptoms of depression, which is generalizable across imaging sites, based on resting-state fMRI (rs-fMRI) data. 75% of the previous translational neuroimaging studies that used machine learning techniques aimed to predict “diagnoses” by psychiatrists classifying patients from healthy controls. However, an increasing number of studies have highlighted the difficulty of finding a clear association between existing clinical diagnostic categories and neurobiological abnormalities.

  In the present study, researchers determined resting state functional connections (FCs), which is coordinated activity between different parts of the brain, related to depression symptoms. The researchers identified the FCs in a data-driven, unbiased manner, and validated them using data from several independent imaging sites. They then compared the important FCs with FCs related to major depressive disorder (MDD) diagnosis to strengthens neuroscientific understanding, and to aid future diagnosis and treatment of MDD. As a result, the FCs related to the insula were common between MDD diagnosis and symptoms of depression. Identification of biomarkers that determine therapeutic targets (theranostic biomarkers) could allow more personalized treatments of MDD than the current situation.

Article