BACKGROUND: Dual-task impairment severely limits functional recovery post-stroke. This diagnostic accuracy study aimed to develop an eye-tracking-based system for objectively quantifying ankle-cognitive integration deficits in stroke survivors. METHODS: This diagnostic accuracy study was conducted from January 2022 to October 2023 at Tsinghua Changgung Hospital in Beijing. A total of 20 healthy adults (mean age 53.15±6.26 years) participated in the study. In addition, 30 patients with a history of stroke (mean age 64.13±8.16 years, 8 females, disease duration 9.12±6.60 weeks) participated in a standardised dual-task evaluation. The novel system utilised 17 parameters, encompassing ankle kinematics (range of motion, velocity) and eye tracking (gaze duration, sweep latency), which were measured simultaneously during the cognitive motor task. Reliability was assessed by intragroup correlation coefficients (ICC), while criterion validity was assessed using 12 clinical evaluation metrics, including Spearman correlation with Montreal Cognitive Assessment (MOCA) scores and dual-task cost (DTC) percentages. RESULTS: The system demonstrated that 88.2% of the evaluation parameters exhibited high consistency, with 55.8% showing a moderate correlation with clinical benchmark scales (p<0.05). Notably, MOCA, DTC%, and TUG-subtraction task duration were identified as key indicators of dual-task ability (P<0.05), while the Self-Rating Anxiety Scale showed lower sensitivity. Furthermore, ankle motion parameters exhibited a strong correlation with balance and fall risk (P<0.05), effectively serving as predictors of motor function recovery and fall risk in stroke patients. CONCLUSIONS: This multimodal system reliably quantifies post-stroke dual-task deficits, with ankle kinematics and eye-tracking metrics serving as sensitive biomarkers for balance recovery and fall risk stratification. Findings advocate integrating objective dual-task metrics into neurorehabilitation protocols to optimize functional outcomes. (ChiCTR2300067640; URL: https://www.chictr.org.cn/showproj.html?proj=188211 ).

Human–machine interfaces based on force touch panels have attracted enormous attention due to the merits of the high human–machine interaction efficiency. Many studies have been devoted to diverse force touch technologies. Broad applications in terms of both actual use and research have been developed, such as 3D touch and force-based keystroke authentication. The fruitful results are based on the assumption that users’ touch habits remain unchanged over time; thus, a stationary customized force-sensing model can be built. However, for long-term use, users’ touch habits change due to time-drifting and specific events, causing a decrease in the performance of stationary force-sensing models. To address this issue, a rectified artificial neural network for long-term force sensing in piezoelectric touch panels is presented in this paper. With additional information on the touching time and the occurrence of specific events, the force level predictions were rectified, achieving an accuracy of 97.62% for a long-term data set. The proposed technique enables customized force sensing for long-term use and enhances the human–machine interactive efficiency.