After accounting for walking speed, higher ataxia severity (SARA) after subacute cerebellar stroke was associated with smaller whole-limb excursions, and these exploratory findings support interpreting trunk acceleration magnitude and regularity alongside whole-limb movement and clinical severity.
Key Findings
Results
Patients with subacute cerebellar stroke walked more slowly than healthy controls and showed lower anteroposterior and mediolateral trunk acceleration RMS but higher anteroposterior and vertical sample entropy.
Multicenter cross-sectional study of 55 patients with subacute cerebellar stroke and 56 healthy controls.
Inertial measurements during comfortable walking quantified trunk acceleration RMS (without speed normalization) and SampEn.
Higher SampEn indicates less regular trunk acceleration patterns in the anteroposterior and vertical directions.
Results
After speed adjustment, higher SARA scores were associated with smaller peak limb flexion and less-negative peak limb extension angles.
Each five-point higher SARA score was associated with peak limb flexion -1.01° (95% CI -1.99 to -0.02).
Each five-point higher SARA score was associated with extension +1.21° (95% CI 0.08 to 2.34).
Both had Benjamini-Hochberg-adjusted P = 0.045.
Analyses were exploratory, patient-only, speed-adjusted regressions.
Results
None of the six full-cohort group-by-speed interactions met the adjusted significance criterion.
Group-by-speed models were used in the full cohort of patients and controls.
This indicates no clear evidence that the speed–gait variable relationships differed between groups after multiple-comparison adjustment.
Results
In speed-matched comparisons, significant waveform differences between patients and controls were confined to limb angles near the beginning and end of the gait cycle.
One-dimensional statistical parametric mapping was applied with 11 participants per group after speed matching.
Differences were not found across the rest of the gait cycle.
Results
In speed-adjusted rank analyses, vertical trunk RMS correlated with SARA and with limb angles, but the original-scale linear SARA coefficient remained uncertain.
Partial Spearman correlation between vertical RMS and SARA: partial ρ = 0.471.
Vertical RMS was also correlated with limb angles.
The linear regression coefficient for SARA on the original scale remained uncertain.
Results
Vertical sample entropy showed no clear association with SARA or limb kinematics.
Assessed in exploratory patient-only speed-adjusted analyses.
Contrasts with the group difference in vertical SampEn between patients and controls.
Conclusions
The authors conclude that trunk acceleration magnitude and regularity should be interpreted together with whole-limb movement and clinical severity, while noting findings are exploratory.
Higher ataxia severity was associated with smaller whole-limb excursions in speed-adjusted models.
The study is cross-sectional with subacute patients and a small speed-matched subsample (11 per group).
What This Means
This study compared how people walk after a stroke affecting the cerebellum (a brain region important for coordination) with healthy adults. Using wearable motion sensors, researchers measured how much the trunk accelerates, how regular that motion is, and how far the legs swing. 55 patients and 56 healthy people took part. Patients walked more slowly, had smaller trunk accelerations front-to-back and side-to-side, and had less regular trunk motion in the front-to-back and vertical directions.
Because walking speed itself changes gait, the researchers adjusted for it. Even then, people with more severe ataxia (higher SARA scores, a clinical rating of coordination problems) had slightly smaller leg movement ranges, with less hip-to-foot bending and extending during each step. Trunk vertical acceleration also tracked with severity in rank-based analyses, though the exact size of that relationship was uncertain. Differences in gait patterns between patients and controls with matched speeds (only 11 per group) were limited to the start and end of the step cycle.
This research suggests that sensor-based measures of trunk movement and leg motion may add useful information about ataxia severity when interpreted together and with speed considered. The authors describe the findings as exploratory, based on a single time point in a modest sample, so they should be confirmed in larger and longer-term studies before being used in clinical practice.
Yamasaki Y, Mizuta N, Ohnishi S, Kai T, Sato H, Okuma K, et al.. (2026). Trunk Acceleration and Lower-limb Kinematics after Cerebellar Stroke: Associations with Walking Speed and Clinical Ataxia Severity.. Cerebellum (London, England). https://doi.org/10.1007/s12311-026-02107-x