Effects of virtual reality on spatiotemporal gait parameters and freezing of gait in Parkinson’s disease

OARC’s Statistical Methods and Data Analytics team collaborated with Dr. Katy Cross of UCLA Health on a project that explored whether virtual reality (VR) could be used to study or treat a phenomenon in Parkinson’s disease called “freezing of gait,” where patients suddenly cannot walk. Dr. Cross’s research investigated whether VR technology could allow for the assessment and rehabilitation of patients’ motor skills in a safer and less stressful environment than real-life scenarios, for example by simulating crossing a road with moving cars.

Dr. Cross’s research team tracked freezing episodes in both VR and real-life environments with varying triggers, like turning or facing obstacles. In total, there were 271 freezing episodes during the experiment, affecting 12 out of 16 participants. Most people had several freezing moments, with a typical person having 17, but some had as few as 1 or as many as 78; four people didn’t freeze at all.

Experts from OARC’s Statistical Methods and Data Analytics team modeled the probability and length of freezing episodes as a function of environment (VR vs real-life) simultaneously in a multilevel, zero-inflated lognormal linear model. This model addresses the challenges of analyzing these data, repeated measurements of individuals, where the measurements are mixtures of zeroes (no freezing) and positive values (freezing), but is still relatively straightforward to interpret.

This study showed that individuals with Parkinson’s disease are just as likely to freeze in VR environments as they are in real-life for any given trigger, though the freezing episodes tended to last longer in VR. These findings could help improve how VR could be a useful alternative to real-world scenarios for understanding and treating freezing of gait in Parkinson’s patients.