Study Rationale: Freezing of gait (FOG) is a symptom that causes many people with Parkinson’s disease (PD) to suddenly lose their ability to take a step while walking, creating a high risk for falling. Despite worldwide efforts, the severity of FOG is still very difficult to measure, especially during walking at home, creating an enormous barrier for developing and testing new treatments. To remedy this issue, we are developing an artificial intelligence-based detection system, AID-FOG, which can automatically measure FOG severity from small wearable sensors worn while walking in daily life. The system could open up new avenues for assessing PD treatments.
Hypothesis: We hypothesize that our AID-FOG algorithm will be able to detect FOG in the home with high precision, as verified against human expert ratings of FOG scored from video recordings, and we predict that AID-FOG will form a reliable basis for providing online therapeutic interventions.
Study Design: For this study, we have recruited 66 people with PD and FOG from Belgium, Germany and Israel. We will assess FOG in participants’ homes using instructed walking tasks, as well as free-living gait without instructions. Using participants’ input about the wearable sensors, we will adjust the system design to be user-friendly and fit the needs of the people with PD. We will use video recordings of the participants’ walking to verify that the AID-FOG algorithm is able to accurately detect and measure FOG severity during everyday life mobility.
Impact on Diagnosis/Treatment of Parkinson’s disease: AID-FOG will automatically capture FOG severity as experienced at home, providing doctors with the best information for optimizing clinical management. AID-FOG also enables further development of treatments that can tackle each episode immediately when it arises. Hence, this project can enhance existing and novel treatments for this devastating symptom.
Next Steps for Development: We will develop a web-based platform that runs our algorithm, allowing clinicians around the world to obtain accurate FOG measures to guide clinical practice. Having a wearable system that can detect FOG in real time can also facilitate development of new, episode-specific interventions in collaboration with neurosurgical and rehabilitation teams.