Objective: This study aims to mitigate the adverse effects of FOG through a novel, wearable device designed for real-time detection and intervention.

Background: One of the most debilitating motor symptoms of Parkinson’s Disease is Freezing of Gait (FOG), which manifests as sudden and temporary episodes of an inability to initiate or continue walking. FOG significantly increases the risk of falls and subsequent injuries, thereby reducing the quality of life for individuals with PD.

Method: As publicly available datasets for FOG detection are scarce, we first developed a device capable of collecting accelerometer and gyroscope data from PD patients to establish a dedicated dataset. After obtaining informed consent, we securely attached the device below the knee of each participant. Among the 15 patients included in our study, 7 experienced FOG episodes, during which our device successfully recorded acceleration and gyroscopic data, thereby generating a dataset suitable for training a deep learning model.

Results: Using the collected dataset, we trained a neural network-based detection algorithm, which achieved a 93% accuracy in real-time FOG recognition. The trained model was subsequently integrated into a mobile application that communicates with the wearable device via Wi-Fi. Upon detecting a FOG episode, the device delivers vibratory stimuli, serving as an external cue to facilitate gait continuation and mitigate freezing episodes.

Conclusion: Our findings indicate that this system holds significant potential as an assistive technology for PD patients, offering a non-invasive and responsive approach to addressing FOG. Further clinical validation with a larger cohort is planned to refine the detection algorithm and optimize the intervention strategy.

« Back to 2025 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/freezing-of-gait-detection-and-warning-system-for-parkinsons-patients/