Impaired motor control after stroke commonly reduces walking speed and increases gait variability. Increased variability can lower measurement reliability and obscure true longitudinal changes following rehabilitation. The authors sought to separate the influence of walking speed from stroke-related neuromotor variability by quantifying between-session reliability and minimal detectable change (MDC) for gait biomechanics in people post-stroke and in neurotypical controls walking at matched speeds.
The study collected treadmill gait data on two separate days. The post-stroke cohort comprised N=15 participants who walked at their self-selected speed. The control cohort included N=13 age- and sex-matched neurotypical individuals who walked at their self-selected speed and at a speed matched to a post-stroke participant. The two-day repeated-measures design enabled computation of between-session intra-class correlation coefficients (ICC) and MDC values for multiple gait variables.
The investigators computed reliability statistics for three classes of gait measures: spatiotemporal variables, bilateral joint ranges of motion (ROM), and bilateral peak propulsive and peak vertical ground reaction forces (GRF). Analyses were performed separately for paretic and non-paretic limbs where applicable.
The study reports intra-class correlation coefficients (ICC) to indicate the consistency of measures across sessions and minimal detectable change (MDC) to indicate the smallest change likely to reflect a true difference beyond measurement error. MDC values were expressed relative to the magnitude of the gait variable (for example, percent of ROM) to facilitate clinical interpretation.
Spatiotemporal measures demonstrated excellent reliability across groups, with ICC values in the range of 0.813 to 0.988. Despite high ICCs, the MDC magnitudes were larger in the post-stroke group compared with the speed-matched neurotypical controls, indicating greater between-session variability in absolute change thresholds for people post-stroke.
Joint ROM reliability varied from poor to excellent depending on the joint and group, with ICC values spanning 0.362 to 0.960. The MDC for joint ROM in the post-stroke group ranged between 27% and 53% of the gait ROM, whereas in neurotypical controls MDCs ranged from 11% to 42% of ROM. Additionally, ROM MDCs were higher in the non-paretic extremity than in the paretic extremity within the post-stroke cohort.
Peak propulsive and peak vertical GRF measures showed good to excellent reliability, with ICCs between 0.778 and 0.980 across groups. However, MDC values for peak GRFs were greater in the post-stroke participants than in speed-matched controls, indicating larger absolute thresholds needed to conclude a real change in force measures post-stroke.
Because neurotypical controls walked both at their preferred speed and at speeds matched to post-stroke participants, the design isolated the influence of reduced walking speed from stroke-related neuromotor impairments. The pattern of larger MDC values in the post-stroke group—despite comparable walking speeds—suggests that stroke-related variability contributes to reduced measurement precision beyond the effects of walking speed alone. In short, impaired neuromotor control after stroke increases the magnitude of change required to be confident that an observed difference is not measurement noise.
The authors present quantitative MDC benchmarks for spatiotemporal variables, joint ROM, and peak GRFs in post-stroke and speed-matched neurotypical participants. These benchmarks can inform interpretation of longitudinal gait changes after clinical interventions by indicating the minimum between-session change likely to represent a true improvement or decline rather than measurement error. Clinicians and researchers should consider these higher MDC thresholds when designing studies or evaluating patient progress post-stroke.
The study received ethical approval from the Institutional Review Board of Chapman University (IRB #23-57 for post-stroke participants; IRB #23-12 for controls) and all participants provided consent. The authors declared no competing interests. Data produced in the study are available upon reasonable request to the authors. Funding sources reported include R03HD107630 and R03TR004248.