Objective: This study explores AI-based video analysis to characterize the Timed Up and Go (TUG) test in Parkinson’s disease. Using MediaPipe Pose, skeletal models are extracted from RGB videos to derive phase-specific functional metrics and investigate their association with clinical scales of axial impairment.

Background: Axial symptoms, including postural abnormalities and instability, are major contributors to disability and risk of falls in PD. The TUG test is widely used to assess functional mobility measuring the total completion time, but potentially overlooking relevant kinematic features. Recent advances in markerless tracking allow objective movement analysis from RGB-only videos and the extraction of quantitative metrics to support clinical assessment.

Method: Forty-four PD patients (H&Y 2-4) were recruited from two Italian Movement Disorders Centers and evaluated using standard clinical scales to assess motor and non-motor impairment. TUG was recorded using a synchronized dual-camera setup to ensure frontal and lateral views. Videos were then processed using GMP to extract 3D skeletal models from the RGB streams. TUG sub-phases were identified from the frontal view, while functional parameters were extracted from the lateral view.

Results: Preliminary analysis of 26 patients was conducted. TUG duration and sit-to-stand, forward gait, and turning sub-phases correlated with H&Y (ρ=0.60, 0.56, 0.67 and 0.43, respectively). TUG sub-phases also correlated with BERG (sit-to-stand, ρ=-0.49), MDS-UPDRS IV (forward gait, ρ=0.56) and MDS-UPDRS item 3.10 (turning, ρ=0.47). Among the forward gait parameters, the mean stride length correlated with LEDD (ρ=-0.36), variability with H&Y (ρ=0.51), and symmetry with BERG (ρ=-0.37) and MDS-UPDRS IV (ρ=0.46). Stance and swing symmetries correlated with H&Y (ρ≥0.40), while trunk lateral and forward flexion with MDS-UPDRS item 3.13 (ρ=0.36 and ρ=0.60).

Conclusion: AI-based RGB-only video analysis using GMP provides objective characterization of TUG. Meaningful correlations were observed between phase-specific metrics and clinical scales, suggesting that video-derived parameters could complement conventional time-based TUG measures and provide valuable insights into PD axial impairment.

References: This study was supported by “Objective monitoring of axial symptoms in Parkinson’s disease: quantitative assessment in daily life based on the use of wearables, video sensing and artificial intelligence (OMNIA-PARK)” project, funded by European Union – Next Generation EU within the PRIN 2022 PNRR program (D.D.1409 del 14/09/2022 Ministero dell’Università e della Ricerca). This abstract reflects only the authors’ views and opinions, and the Ministry cannot be considered responsible for them.

« Back to 2026 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/objective-assessment-of-the-timed-up-and-go-test-in-parkinsons-disease-using-ai-based-markerless-video-analysis/