Movement Recovery Lab Publishes New Bayesian Modeling Study in Brain Stimulation

A new paper published in Brain Stimulation reports a hierarchical Bayesian modeling approach that provides faster and more reliable estimates of how muscles respond to increasing levels of stimulation delivered to the brain or spinal cord. By gradually increasing stimulation intensity and recording muscle activity, researchers generate “recruitment curves” that show how the nervous system facilitates movement.

These curves are critical in neurostimulation studies of spinal cord injury and recovery, but collecting enough data can be time-consuming and uncomfortable for study participants. The new method aims to make this process more efficient while better accounting for uncertainty in the measurements.

The method was developed by the Movement Recovery Lab’s data scientist Vishweshwar Tyagi, MS, together with James McIntosh, PhD and colleagues. Their model is designed to perform well with limited data and helps clinicians reduce experiment duration and participant discomfort while maintaining accuracy. An accompanying open-source Python software library (hbMEP) allows other researchers to use the method in their studies.

This work builds on the lab’s ongoing efforts to improve the analysis of stimulation-evoked responses and to understand how the brain and spinal cord interact during recovery. The team, led by Jason B. Carmel, MD, PhD, continues to investigate how precise measurements and modeling can help evaluate therapies aimed at restoring movement after spinal cord injury.