AI Science Tracker
EN

Yale School of Medicine

1 discovery from 2024

Discoveries

Post hoc neural gait analysis detects a tandem-walk metric difference in an ataxia trial

Source check: Source checked. Not independent scientific confirmation.

Finding: A deep pose-estimation pipeline extracted gait measurements from a small convenience subset of a randomized ataxia trial. Post hoc analysis found a tandem-walk metric difference between troriluzole and placebo, while the natural-walk comparison was nonsignificant and treatment did not reduce falls. The result supports a bounded retrospective measurement finding, without confirmatory evidence of clinical benefit.

Technology: MPII-trained multi-stage deep neural network for human pose estimation; derived Pose Dispersion Index

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
The learned pose model extracted body coordinates from trial videos, enabling a source-specific gait metric and treatment-group comparison that standard clinician ratings did not directly provide. Trial randomization does not make this later AI-derived endpoint a locked prospective evaluation.
Test and result
Among 218 randomized participants, interpretable paired videos were available for 67 tandem-walk and 56 natural-walk participants. The pretrained deep network tracked body keypoints used to calculate pose dispersion. An adjusted treatment-by-visit mixed model reported a week 48 tandem-walk coefficient 0.584, 95% CI 0.137–1.031, p=0.010. The natural-walk difference was nonsignificant, and the source explicitly reports no treatment effect on falls.
Limitation / still unknown
The video endpoint was post hoc and was not the original recording purpose. Convenience and quality selection retained only 30.7% and 25.7% of participants, respectively, introducing selection and measurement limitations. Only baseline/final recordings were analyzed, supplementary metric implementation was not independently reconstructed, and the finding does not establish prespecified efficacy, reduced falls or clinical utility.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://link.springer.com/article/10.1007/s40120-024-00625-6

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Read actual primary trial/video eligibility, pretrained pose-model and derived-index Methods, post hoc mixed/Poisson evaluation, tandem/natural-walk Results, no-falls-benefit statement and convenience-video limitations. Supplementary metric implementation not independently reconstructed. No prospective locking or clinical-benefit claim adopted. Exact original article front DOI/title and fresh PubMed primary relationship metadata checked..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open the interactive profile for the timeline, or browse all organizations and model families.