Brain-gut microbiota multimodal predictive model in patients with bipolar depression

Area of Science:

• Neuroscience
• Microbiology
• Psychiatry

Background:

• The microbiota-gut-brain axis plays a role in bipolar disorder (BD) pathophysiology.
• Specific gut microbiota associations and their predictive potential in BD remain unclear.

Purpose of the Study:

• To investigate the relationship between gut microbiota and neuroimaging in bipolar disorder.
• To develop a multimodal approach for BD diagnosis and treatment response prediction.

Main Methods:

• Collected fecal and MRI data from 103 BD patients and 39 healthy controls (HCs).
• Performed metagenomic sequencing and functional connectivity (FC)/grey matter volume (GMV) analyses.
• Constructed support vector machine (SVM) classifiers using multimodal data.

Main Results:

• Quetiapine responders showed pre-treatment characteristics closer to HCs than non-responders.
• Disrupted functional connectomes in BD correlated with specific gut bacteria.
• Multimodal classifier achieved high accuracy (AUC-ROC 0.9517 for BD vs HC, 0.8292 for responders vs non-responders).

Conclusions:

• Combining gut microbiota and neuroimaging offers a superior approach for BD diagnosis.
• This multimodal strategy can predict quetiapine treatment efficacy in BD patients.