Machine learning-based identification of abnormal functional connectivity in obesity across different metabolic...

Area of Science:

• Neuroscience
• Metabolic Health
• Brain Imaging

Background:

• Obesity is a significant health issue linked to chronic diseases.
• Neurological studies often overlook metabolic state variations, limiting understanding of brain function.
• Data-driven approaches can reveal complex neural interactions in obesity.

Purpose of the Study:

• To identify brain connectivity patterns associated with obesity.
• To analyze these patterns across different metabolic states (fasting to satiety).
• To utilize a data-driven methodology, avoiding hypothesis bias.

Main Methods:

• Electroencephalography (EEG) data collected from women with and without obesity.
• Functional connectivity analysis performed on source-localized EEG signals.
• Machine learning framework with feature selection to identify discriminative connectivity patterns.

Main Results:

• Six connectivity features accurately classified obesity (95% accuracy) across metabolic states.
• Obese individuals showed reduced connectivity in food-reward processing regions.
• The dorsal anterior cingulate cortex (dACC) was identified as a central hub in altered connectivity.

Conclusions:

• Disrupted brain connectivity is a fundamental aspect of obesity.
• The dACC plays a crucial role in maladaptive reward processing in obesity.
• Targeting the dACC via neuromodulation may be a viable obesity treatment strategy.