Alterations in whole-brain white matter structural network among females with abdominal obesity by appetite subtypes

Abdominal obesity (AO) in women exhibits clinical heterogeneity, yet its neural mechanisms remain unclear. Using diffusion tensor imaging (DTI), this study conducts the first systematic investigation of whole-brain white matter (WM) structural network topological properties and female-specific neural mechanisms in AO patients stratified by appetite subtypes (strong appetite, SA; moderate appetite, MA). This study enrolled 60 female AO patients (stratified into 30 SA and 30 MA) and 30 age-matched healthy controls (HCs). Whole-brain WM structural networks were constructed, with global and local topological properties analyzed through graph theoretical approaches. Statistical correlations between network metrics and clinical measures were assessed. Global network analysis demonstrated intact small-world properties ( γ > 1, σ > 1) across all groups. At the nodal level, compared with HCs, SA patients showed reduced degree centrality (DC) in the right temporal pole: superior temporal gyrus (TPOsup. R) ( p = 0.017) and lower nodal local efficiency (NLE) in the right amygdala (AMYG. R) ( p = 0.007). Notably, TPOsup. R DC positively correlated with appetite scores ( r = 0.408, p = 0.025). Additionally, the SA group exhibited reduced betweenness centrality (BC) in the right rectus gyrus (REC. R) compared to the MA group ( p = 0.041), which negatively correlated with appetite scores ( r = −0.439, p = 0.015). Conversely, MA patients showed increased NLE in the right medial orbital superior frontal gyrus (ORBsupmed. R) ( p = 0.005). Whole-brain connectome analysis identified a hypo-connected subnetwork specific to MA involving 23 nodes and 24 edges, highlighting weakened right amygdala (AMYG. R)–right superior occipital gyrus (SOG. R) ( r = 0.464, p = 0.01) and right lingual gyrus (LING. R)–right fusiform gyrus (FFG. R) ( r = 0.365, p = 0.047) connections that showed positive correlation with waist circumference. In summary, this study provides the first evidence that female patients with abdominal obesity exhibiting different appetite phenotypes possess specific white matter structural network patterns. The SA subtype is characterized by aberrant nodal topological properties within prefrontal-limbic circuits, reflecting a localized disruption associated with heightened appetite. In contrast, the MA subtype is defined by widespread structural disconnectivity between occipital sensory areas and limbic regions, appearing as a neuroanatomical consequence linked to abdominal adiposity severity.