Geo-Mamba: Geometry-informed state-space learning of functional brain organization.
Geo-Mamba is a geometry-informed dual-path state-space model that operates on Riemannian manifolds of fMRI/EEG functional connectivity to produce compact SPD representations and improves classification robustness across multiple neuroimaging datasets, including three Parkinson's cohorts.
What the AI sees
Geo-Mamba is a geometry-informed dual-path state-space model that operates on Riemannian manifolds of fMRI/EEG functional connectivity to produce compact SPD representations and improves classification robustness across multiple neuroimaging datasets, including three Parkinson's cohorts.
Research significance
Although it does not address molecular mechanisms or interventions, the method has moderate translational value by potentially improving PD biomarker detection, patient stratification, and cross-modal neuroimaging analyses for clinical studies and trial endpoints.
Source abstract
Functional magnetic resonance imaging (fMRI) derived functional connectivity (FC) is represented as graphs and as correlation or covariance matrices that live on non-Euclidean spaces, cortical graphs and the Riemannian manifold of symmetric positive-definite (SPD) matrices, thus conventional Euclidean sequence models are misspecified. To this end, we introduce Geo-Mamba, a geometric variant of Mamba formulated on Riemannian manifolds. Geo-Mamba employs a dual-path selective state-space design, (1) a stacked path performs hierarchical modeling by aggregating pyramid multi-granular features to capture short- and long-range dependencies; and (2) a distillation path combats redundancy in high-dimensional SPD inputs via progressive, geometry-aware dimensionality reduction (operating in the manifold spaces) to produce compact states without violating Riemannian constraints. Their complementary outputs are fused through the tailored GeoMix operator to yield a compact, discriminative SPD representation. Geo-Mamba is evaluated on seven public fMRI datasets, including two Alzheimer's disease cohorts, three Parkinson's disease cohorts, one Autism dataset, as well as a longitudinal single-site, single-scanner study designed for detecting subtle changes in the brain due to a season of playing contact sports. To further evaluate the cross-modal applicability and scalability of the model, we apply Geo-Mamba to three electroencephalography (EEG) datasets. Across these benchmarks, it delivers consistently competitive accuracy and robustness, supporting the value of dual-path manifold modeling for neuroimaging and its potential for clinical translation. The code is released at https://github.com/acmlab/Geo-Mamba.