Background: Parkinson's disease (PD) is a neurodegenerative disorder causing postural control impairments. Postural control involves multiple domains, such as control of postural sway in stance, automatic postural responses (APRs) and anticipatory postural adjustments (APAs). We hypothesize that impairments in each postural domain is associated with resting-state functional connectivity (rsFC), accounted by predictive modeling and that cortical and cerebellar networks would predict postural control in people with PD (PwPD). Objective: To determine whether rsFC can predict three domains of postural control independently in PwPD and older adults (OA) based on predictive accuracy of models. Methods: The cohort consisted of 65 PwPD (67.7 +8.1 age) tested in their OFF-state and 42 OA (69.7 +8.2 age). Six body-worn, inertial sensors measured postural sway area while standing on foam, step length of APRs to a backward push-and-release perturbation, and magnitude of lateral APAs prior to voluntary gait initiation. Resting state-fMRI data was reported on 384 regions of interest that were grouped into 13 functional brain networks. Associations between rsFC and postural metrics were characterized using predictive modeling, with an independent training (n = 67) and validation (n = 40) dataset. Models were trained in the training sample and performance of the best model was validated in the independent test dataset. Results: rsFC of different brain networks predicted each domain of postural control in PD: Frontoparietal and Ventral Attention rsFC for APAs; Cerebellar-Subcortical and Visual rsFC and Auditory and Cerebellar-Subcortical rsFC for APRs; Ventral Attention and Ventral Multimodal rsFC for postural sway. In OA, CinguloOpercular and Somatomotor rsFC predicted APAs. Conclusions: Our findings suggest that cortical networks predict postural control in PD and there is little overlap in brain network connectivities that predict different domains of postural control, given the rsFC methodology used. PwPD use different cortical networks for APAs compared to OA.
Bibliographical noteFunding Information:
This work was supported by the National Institute of Aging [ AG006457 ] and VA Merit Award [ RX001075 ] the Tartar Trust Fellowship (Miranda-Dominguez), OHSU Parkinson Center Pilot Grant Program (Miranda-Dominguez). The authors thank all participants for generously donating their time to participate, Peter Fino, Carolin Curtze, Mike Fleming, Heather Schlueter, Peter Martin, Graham Harker and Natassja Pal for helping with data collection, and Daniel Peterson and Katrijn Smulders for data collection and help with study procedures.
- Brain networks
- Functional connectivity
- Parkinson's disease
- Postural control
- Prediction accuracy
PubMed: MeSH publication types
- Journal Article
- Research Support, N.I.H., Extramural
- Research Support, Non-U.S. Gov't
- Research Support, U.S. Gov't, Non-P.H.S.