Paretic stance-phase biomechanical changes associated with gait speed improvement 4 weeks after ankle robot-assisted gait training in community-dwelling individuals with post-stroke hemiparesis: an exploratory study
Background Improving walking ability is an important rehabilitation goal for individuals with post-stroke hemiparesis. Robot-assisted gait training (RAGT) with an ankle robot can assist ankle motion during stance; however, the relationship between the resulting biomechanical changes and improvements in gait speed remains unclear, as do the gait pattern changes that should be targeted during ankle RAGT. We therefore conducted an exploratory analysis to examine which stance-phase biomechanical changes after ankle RAGT were associated with changes in gait speed from baseline to the 4-week follow-up. Methods Community-dwelling individuals with post-stroke hemiparesis completed 15 sessions of RAGT using an ankle robot driven by pneumatic artificial muscles. The primary outcome was gait speed, measured at baseline, immediately after the intervention, and at the 4-week follow-up. Biomechanical variables during level walking were analyzed before and immediately after the intervention using three-dimensional motion analysis and force plates and were calculated at initial contact and the ends of the loading response and terminal stance. Spearman’s rank correlation analysis was used to examine associations between changes in these variables and changes in gait speed immediately after the intervention (ΔVpost) and at the 4-week follow-up (ΔVFU); multiple linear regression analysis was then used to identify variables independently associated with ΔVFU. Results Mean gait speed was 63.7 cm/s (SD, 25.2) at baseline and increased by 6.4 cm/s (SD, 10.8) immediately after the intervention and by 7.3 cm/s (SD, 11.4) at the 4-week follow-up; 10 and 12 of the 30 participants, respectively, exceeded the clinically meaningful change threshold of > 10 cm/s. More biomechanical changes at the end of the loading response than at the other two events were significantly associated with ΔVFU. In a regression model of these variables (adjusted R² = 0.521, p < 0.001), increases in the paretic hip adduction angle and braking force were independently associated with ΔVFU. Conclusions These findings suggest that biomechanical changes related to paretic-limb weight acceptance during loading response are associated with improvement in gait speed at the 4-week follow-up after ankle RAGT. These exploratory findings are hypothesis-generating and require prospective confirmation. Trial registration UMIN Clinical Trials Registry, UMIN000031194. Registered 7 February 2018.
Authors
- Tomoyuki Noda (ORCID: https://orcid.org/0000-0002-1055-5080)
- Kimihiko Mori (ORCID: https://orcid.org/0000-0002-3806-4400)
- Masanori Wakida (ORCID: https://orcid.org/0000-0002-6582-2203)
- Yuta Chujo (ORCID: https://orcid.org/0000-0002-9356-1151)
- Meguru Taguchi (ORCID: https://orcid.org/0009-0002-7478-6440)
- Takayuki Kuwabara (ORCID: https://orcid.org/0009-0008-6980-3502)
- Naoto Mano (ORCID: https://orcid.org/0000-0002-1770-4229)
- Hisayuki Ogura (ORCID: https://orcid.org/0000-0001-9004-3837)
- Tatsuya Teramae (ORCID: https://orcid.org/0000-0001-7957-5576)
- Takanari Kubo (ORCID: https://orcid.org/0000-0003-1049-5302)
- Kimitaka Hase (ORCID: https://orcid.org/0000-0002-0379-6513)
- Fumiya Suzuki
Institutions
- Kansai Medical University (JP)
- Advanced Telecommunications Research Institute International (JP)
- Mie University Hospital (JP)
Publication Details
- Journal
- Journal of NeuroEngineering and Rehabilitation
- Published
- 2026-11-21
- DOI
- https://doi.org/10.1186/s12984-026-02233-x
- Primary Topic
- Stroke Rehabilitation and Recovery
- Type
- article
- Field-Weighted Citation Impact
- 0.00