Abstract
Obesity presents a critical public health challenge and is commonly associated with a marked decline in quality of life. Existing research exploring the neurological correlates of obesity through electroencephalography (EEG) has predominantly employed traditional statistical approaches, which rely on prior assumptions about brain networks and are limited in their ability to capture complex interactions between neural features. In this study, we conducted a machine learning (ML) analysis using a novel incremental wrapper-based feature selection method (DLI-WFS) to identify neural signatures of obesity in female individuals through functional connectivity-based resting-state EEG classification. Our proposed model demonstrated its efficiency by outperforming other benchmark models in the classification task with only a minimal number of features selected. Neurologically, our results indicate that obesity is associated with a disrupted network, where regions involved in processing self-referential and contextual environmental information exhibit functional impairments. By exercising targeted intervention in the relevant brain connections, it is possible to enhance neurological behaviours associated with obesity.