Evaluating Voxel-Level and Network-Level Approaches for Functional Connectivity MRI in Gait Disorders

Functional movement disorder (FMD) is a common neuropsychiatric disorder in which patients present with involuntary motor symptoms and altered brain function in the absence of structural lesions. Among FMD subtypes, Functional Gait Disorder (FGD) is one of the most disabling. However, despite its clinical relevance, the underlying neurobiological mechanisms remain poorly understood. Although previous resting-state fMRI studies have identified altered functional connectivity in FMD, their findings are limited both methodologically and by clinical heterogeneity. Most of these studies have used seed-based methods, which are not sensitive to large-scale network connectivity or hierarchical brain organization. In addition, some have included heterogeneous clinical samples that pool different motor phenotypes, thereby reducing the specificity of the results and limiting the ability to identify unique neurobiological features, while leaving subgroups such as FGDs largely unexplored. This study addresses these limitations by applying a multi-scale approach in a clinically homogeneous population of patients with FGD. By moving beyond localized regional analysis, this work provides a whole-brain characterization of altered network integration and its relationship with clinical symptomatology, with a focus on the cortical hierarchy revealed by functional gradient analysis. Resting-state Functional Magnetic Resonance Imaging (fMRI) data were obtained for 15 patients with FGD and 15 healthy controls matched for age and sex (3T MRI, TR = 2s, scan duration ≈ 6 min). The analysis included seed-based functional connectivity (FC) using precuneus, pre- and post-central gyrus, supplementar motor area (SMA), the right temporo-parietal junction (TPJ) and amygdala, as well as network-level connectivity based on Schaefer's parcellation (200 regions and 17 networks). In addition, functional gradient mapping using Laplacian embedding, was used to investigate the compression or dispersion of the unimodal-transmodal cortical hierarchy. Clinical severity was measured with S-FMDRS and questionnaires for fatigue, anxiety, depression and alexithymia, and these clinical measures were subsequently correlated with gradient-derived distance metrics. The analysis showed hyper-connectivity between amygdala and several motor brain regions in patients, increased between-network connectivity of the Dorsal Attention Network together with reduced integration of the Default Mode Network with visual, somatomotor and limbic systems, and a pronounced reorganization of cortical gradients, characterized by increased gradient distance and a shift of left-SomMotB, right-SalVentAttnA and right-ContC networks toward transmodal regions. Gradient distance in the right-SalVentAttnA and left-SomMotB correlated positively with anxiety scores, whereas distance in the right-ContC and left-SomMotB network was negatively correlated with reduced motivation scores and duration of symptoms. Our findings indicate that FGD results from a loss of functional segregation and compression of the cortical hierarchy, rather than isolated regional deficits. Therefore, multi-scale connectivity and gradient-based metrics appear to be promising tools for objective characterization, longitudinal monitoring, and development of tailored therapeutic strategies for functional gait disorders.