NCT03655171

A Pilot Study to Evaluate the Severity of Motor Dysfunction in Parkinson's Disease Based on AI Video Analysis

Huashan Hospital Affiliated to Fudan University

• Status: Completed
• Phase: —
• Study type: Observational
• Enrollment: 30 (actual)
• Sponsor: Huashan Hospital · Academic / Other
• Sex: All
• Age: 18 Years – 80 Years
• Healthy volunteers: Not accepted

Summary

This study is a pilot study, mainly to explore the potential application value of specific pattern movement after video-based quantitative methods in the early recognition and assessment of Parkinson's disease. According to UPDRS III, a series of motion indicators related to the characteristics of the disease were determined and quantitatively analysed. Motor function scores were given by the senior physicians and the AI video analysis team separately to evaluate the accuracy of the scores of AI video analysis compared with that of senior specialists' team of movement disorders.

Detailed description

In this study, participants are asked to perform video-recorded UPDRS-III test, and the unmarked motion feature identification and quantitative analysis based on the video were conducted. The investigators have collected standardized motor function videos of patients with Parkinson's disease from outpatient clinics and follow-up since August 2017. Based on the UPDRS III motor function test, the investigators screened five pattern actions that are both disease-characteristic and easy to visualize, including finger tapping, hand movements, pronation-supination movements of hands and gait test. With the assistance of the artificial intelligence team, the characteristic value quantification (frequency, distance, angle, etc.) of the above pattern actions has been initially implemented. Considering that the characteristic values of the pattern action under the identification of video are all continuous variables and cannot be directly compared with the discrete UPDRS rating scale, the investigators initially explored the construction of deep learning algorithm based on UPDRS rating in the previous work to verify the effectiveness of video analysis in motor function evaluation. In this study, the investigators plan to include patients with Parkinson's disease with different disease severity and analyse the consistency between UPDRS scores evaluated by specialists of movement disorders and video quantification score. The results of this study will hopefully lay a good foundation for launching a large-scale, multi-centre clinical trial in the future.

Conditions

• Parkinson Disease

Interventions

Type	Name	Description
DIAGNOSTIC_TEST	Video-based quantitative methods	The AI video analysis is used to automatically identify motion details and quantify motion features such as amplitude and angular velocity, then an algorithm will be applied to generate the degree of motor dysfunction.

Timeline

• Start date: 2018-08-20
• Primary completion: 2018-10-20
• Completion: 2018-11-20
• First posted: 2018-08-31
• Last updated: 2019-01-22

Locations

1 site across 1 country: China