Structural-functional network decoupling in early stage amyotrophic lateral sclerosis reveals cell-type specific...

Amyotrophic lateral sclerosis (ALS) involves widespread brain network dysfunction, yet the molecular mechanisms linked to these alterations remain poorly understood. We investigated macroscopic structural-functional coupling abnormalities in early-stage ALS (ALS-ES) and their underlying transcriptomic signatures.

We analyzed multimodal MRI data from 73 patients with sporadic ALS-ES and 74 age- and sex-matched healthy controls. Structural-functional (SC-FC) coupling was quantified using diffusion tensor imaging and resting-state functional MRI. Machine learning models were constructed to distinguish patients from controls based on network features. Coupling alterations were spatially correlated with neurotransmitter receptor maps and gene expression profiles from the Allen Human Brain Atlas. Key transcriptomic findings were validated using independent single-cell RNA sequencing datasets.

While structural connectivity remained largely preserved, functional connectivity was significantly reduced in the somatomotor network (SMN). This mismatch manifested as significant SC-FC network decoupling, particularly within the SMN (p_FDR = 0.001). A gradient boosting machine model accurately classified patients, identifying SC-FC coupling in the left precentral gyrus as a primary statistical contributor to the classification model. Decoupling spatially correlated with 5-HT2A and mGluR5 receptor distributions. Imaging-transcriptomics linked network failure to a gene signature enriched for synaptic pathways and microglial markers. Single-cell analysis identified FMN1 as a candidate gene whose glial expression spatially associates with network decoupling.

Early-stage ALS is characterized by significant structural-functional network decoupling, primarily in motor systems. This macroscopic failure is linked to specific microglial dysregulation, particularly FMN1 downregulation, providing a multiscale framework bridges statistical neuroimaging signatures with potential cellular pathology.