Walking towards the future: Exploring OpenTUG's validity in automatic walking activity analyses and the relationship to cognition in vestibular patients.

Vestibular patients required significantly more time and steps to complete the TUG than age-matched controls at normal and slow walking speeds.

• At normal speed, patients took 7319 ms versus 5976 ms for controls (p = .002).
• Differences were significant at normal and slow speeds but not reported as significant at fast speed.
• Sample consisted of 48 vestibular patients and 26 mean age-matched controls.
• The Timed Up and Go (TUG) task was performed at three speeds: normal, slow, and fast.

Automated OpenTUG analyses were highly correlated with manual analyses for both time and step count measures.

• Time correlations: all r > .94, p < .001 across all walking speed conditions.
• Step count correlations: all r > .74, p < .001 across all walking speed conditions.
• OpenTUG uses 2D human pose estimation to automatically analyze walking activity.
• These results support the feasibility and validity of the automated approach for extracting gait parameters.

Vestibular patients reported significantly higher dizziness-related handicap than controls.

• Median DHI (Dizziness Handicap Inventory) total score was 30 for patients versus 0 for controls (p < .001).
• The difference was statistically significant, indicating substantially greater self-reported dizziness-related handicap in patients.

Vestibular patients reported significantly higher psychological symptoms than controls.

• Median HADS (Hospital Anxiety and Depression Scale) total score was 12 for patients versus 8 for controls (p = .005).
• The HADS assesses anxiety and depression symptoms.
• Psychological symptoms were assessed as part of quality-of-life questionnaires.

TUG performance measures showed moderate correlations with attentional performance in exploratory analyses.

• Spearman's r ranged from .30 to .54 for correlations between TUG measures and attentional performance.
• Statistical significance ranged from p < .05 to p < .01.
• Neuropsychological evaluations assessed attention as part of the study protocol.
• These were described as exploratory analyses.

TUG performance measures showed moderate correlations with postural stability indices.

• Correlations with postural stability indices ranged from r = .38 to .57 (p < .05 to p < .01).
• A specific example cited was the unstable platform with eyes open condition (r = .38-.57, p < .05-.01).
• Postural stability was assessed as part of the vestibular evaluation protocol.
• These correlations were identified through exploratory analyses.