Early warning for human mental sub-health based on fMRI data analysis: an example from a seafarers' resting-data...

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Psychiatry

Background:

• Early detection of mental sub-health is crucial for effective intervention.
• Traditional mental health assessments via questionnaires have limitations.
• The default mode network (DMN) is increasingly recognized for its role in mental health status.

Purpose of the Study:

• To develop and validate an early warning mechanism for mental sub-health.
• To investigate the utility of fMRI data in identifying mental sub-health.
• To propose a novel computational model for mental health assessment.

Main Methods:

• Constructed a structural-functional DMN template using anatomical and independent component analysis (ICA).
• Employed a two-fold support vector machine (SVM) classifier (one-class SVM followed by two-class SVM).
• Utilized correlation coefficients (CCs) of DMN regions as features for classification.

Main Results:

• The proposed model effectively discriminated seafarers with DMN alterations from healthy controls (HC).
• Mental sub-healthy seafarers exhibited significant alterations in DMN functional connectivity.
• Observed decreased CCs within DMN regions and altered regional homogeneity and fractional amplitude of low-frequency fluctuation.

Conclusions:

• fMRI-based analysis of brain functional activity can effectively distinguish between mental health and sub-health states.
• The developed two-fold SVM model shows promise for early warning of mental sub-health.
• Findings highlight specific DMN alterations associated with mental sub-health.