HLHuey-Wen Liang

Huey-Wen Liang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Huey-Wen Liang has a total of 164 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Association of Daily Activity Regularity With Functional Recovery in Patients With Stroke and Spinal Cord Injury: A Prospective Actigraphy Study

    by Yu-Hsuan Lin, Chen Lin, Chueh-Hung Wu, Hsiang-Chih Chang, Huey-Wen Liang, Shao-Yu Chen, Kuan-Lin Sung, Wei-Chen Hsu

    Objective: To determine whether the stability of daily physical activity, quantified by interdaily stability (IS), is associated with activities of daily living (ADL) recovery during inpatient neurorehabilitation, independent of overall activity level, we examined the association between physical activity stability and ADL recovery.Methods: This prospective study recruited 161 patients with stroke or spinal cord injury (mean age: 62.4±14.8 years; 42% female) from two hospitals. Physical activity was assessed using actigraphy for seven days after admission. IS represented activity regularity, and M10 the total activity volume during the most active 10-hour period. ADL was evaluated using the Barthel Index (BI) (baseline BI=32.5±22.9; discharge BI=61.1±28.6). Multiple linear regression analyses examined IS, M10, and ADL recovery, adjusting for baseline BI, age, sex, hospital stay, days from onset to rehabilitation admission, and diagnosis group.Results: Higher IS during the first week after admission was significantly associated with better ADL recovery at discharge after controlling for all covariates (B=30.86, p=0.005). In contrast, total physical activity (M10) was not significantly associated with ADL recovery (B=-0.11, p=0.313). Younger age, lower baseline BI, longer length of stay, and shorter onset-to-admission time were also associated with greater ADL recovery.Conclusion: Greater activity stability was associated with greater ADL improvement in patients with stroke, but not in those with spinal cord injury. IS may be a clinically relevant behavioral marker during inpatient rehabilitation, although its causal role remains to be established.

  • Open Access

    Validation of the Taiwan Chinese Version of the Assistive Technology Usability Questionnaire for People With Neurological Diseases for Wearable Robotic Exoskeletons: Usability Study

    by Hander Wang, Jian-Jia Huang, Yu-Cheng Pei, Huey-Wen Liang

    Background: Assessing usability is important given the increasing use of technology in rehabilitation. While wearable robotic exoskeletons (WREs) are commonly incorporated into neurological rehabilitation in both clinical and community settings, there is a need to examine the applicability of different usability instruments in this context. The Assistive Technology Usability Questionnaire for People with Neurological Diseases (NATU Quest) was developed to evaluate the usability of assistive technology in individuals with neurological conditions. However, its reliability and validity have been established only for assistive devices such as wheelchairs and canes, restricting its generalizability. Objective: This study aimed to cross-culturally adapt a Taiwan Chinese version of the NATU Quest and to preliminarily evaluate its applicability, reliability, and validity for assessing WRE use among patients who had a stroke as an expanding application context. Methods: Cross-cultural adaptation was achieved through a structured translation from Spanish to Taiwan Chinese, followed by content validation by an expert panel of 10 raters. Following the adaptation process, psychometric testing of the NATU Quest was administered to 20 patients who had a stroke following the completion of a comprehensive course of robot-assisted gait training facilitated by a WRE. The internal consistency of the NATU Quest was evaluated using the Cronbach α coefficient. The reliability of the test-retest procedure was evaluated using the intraclass correlation coefficient (ICC). Convergent validity was ascertained by examining the correlation between the NATU Quest scores and the System Usability Scale (SUS). Results: The culturally adapted NATU Quest showed acceptable content validity, with item-level content validity index (CVI) values ranging from 0.80 to 1.00 and scale-level CVI values of 0.93 for clarity and 0.97 for relevance. Cronbach α was 0.943, indicating high internal consistency. Of the 20 participants who completed all 6 rehabilitation sessions, 13 participants completed repeat testing. Preliminary test-retest reliability was acceptable (ICC=0.823; P <.001; SE of measurement=0.31). NATU Quest scores were strongly correlated with SUS scores (Spearman ρ=0.818; P <.001), providing preliminary support for convergent validity. Conclusions: The adapted NATU Quest demonstrated promising preliminary psychometric properties when applied to WRE use among individuals who had a stroke. Further studies with larger sample sizes are warranted to examine additional psychometric properties such as construct validity and responsiveness.

  • Smartphone-Based Telerehabilitation for Low Back Pain: A Randomized Controlled Trial on Pain, Disability, Fear-Avoidance, and Trunk Muscle Endurance

    by Cheng-Yen Liao, Bor-Shing Lin, Huey-Wen Liang, Po-Yao Wang, Tzu-Jung Huang, Si-Huei Lee

    Low back pain (LBP) is a widespread global health problem associated with a significant economic burden. Compared with traditional treatment, telerehabilitation may enhance exercise adherence and execution. This randomized controlled trial evaluated the efficacy of a smartphone-based mHealth platform for telerehabilitation (SMPT) in individuals with LBP. Participants were randomly assigned to either the SMPT group (n = 23) or a control group (n = 19) for a 12-week home-based program. Outcomes included pain assessed by the visual analogue scale (VAS), disability by the Roland–Morris disability questionnaire (RMDQ), fear-avoidance beliefs by the fear-avoidance beliefs questionnaire (FABQ), and trunk endurance tests. After the 12-week intervention, the SMPT group showed significantly lower postintervention VAS scores compared with the control group (19.0 [11.5–31.5] vs. 27.0 [20.0–45.5]; p = 0.036; r = 0.35). A greater median reduction in pain intensity was observed in the SMPT group (−29.0 [−44.5 to −17.5]) than in the control group (−11.0 [−33.5 to −4.0]) (p = 0.044; Hodges–Lehmann Δ = 15.0 [95% CI −60.1, 64.3]). Within the SMPT group, disability and fear-avoidance outcomes also improved significantly, with RMDQ scores decreasing from 7.0 [5.0–9.5] to 4.0 [3.0–6.5] (p = 0.013) and FABQ total scores from 34.0 [27.5–41.5] to 27.0 [19.5–34.5] (p = 0.005). These results suggest that the SMPT is effective in reducing pain, disability, and fear-avoidance beliefs, while improving trunk endurance. The SMPT may serve as a practical adjunct to conventional rehabilitation in patients with LBP.

  • Open Access

    Predicting Fall Risk in Community-Dwelling Older Adults Using a Fine-Tuned Quantized Large Language Model

    by Shahab S. Band, Fatemeh Asghari Hampa, Faezeh Gholamrezaie, Hsin-Shui Chen, Kai-Chieh Chang, Huey-Wen Liang

    Computerized posturography has been employed to quantify an individual’s intrinsic balance control under varying stances, thereby presenting the potential to support autonomous and ambulatory fall risk assessment when integrated with machine learning (ML) techniques. However, the superiority of posturography-based approaches over conventional methods such as questionnaires or physical performance tests remain insufficiently documented. In this study, we compared the predictive performance of various combinations of input data and introduced a novel ML approach that incorporates a Large Language Model (LLM) to enhance prediction while enabling feature-based, summarized explanations to improve the transparency of the predictions. We followed 206 community-dwelling older adults over a 6-month period to monitor fall events. At baseline, all participants completed a survey capturing demographic information, self-reported questionnaires, various physical performance tests, and four standing tasks assessed via tracker-based posturography. The predictive validity of these data in distinguishing fallers from non-fallers was evaluated using traditional ML models, and an LLM enhanced with Quantized Low-Rank Adaptation (QLoRA). The 6-month fall incidence was 16.9%. Traditional ML models achieved an area under the curve (AUC) ranging from 0.54 to 0.71 using different combinations of questionnaire responses, physical performance data, and posturographic parameters. Notably, a higher AUC (0.88) and accuracy (0.86) were achieved by applying the LLM with QLoRA to posturographic parameters alone. In conclusion, this study contributes to a deeper understanding of the relationship between postural control and fall risk, and demonstrates the potential of LLMs to improve predictive accuracy while minimizing the need for labor-intensive expert annotation.

  • Open Access

    A Review on AI Approaches in Elderly fall Monitoring Systems: Taxonomies, Challenges, and Open Issues

    by Shahab S.BAnd, Meghdad Biyari, Chung-Chian Hsu, Tommaso Ruga, Ester Zumpano, Luciano Caroprese, Eugenio Vocaturo, Iman Dehzangi, Bharat S. Rawal, Huey-Wen Liang

    Early fall detection among the elderly can dramatically reduce the risk of major problems, including disabling fractures and psychological issues, with falls potentially leading to severe consequences, including death. Artificial Intelligence has emerged as a valuable tool, leading to the development of various automatic fall detection systems for elderly people. The main objective of this paper is to analyze the state of the art in Elderly Fall Detection Systems while maintaining a focus on the explainability of results. A comprehensive analysis of the literature will be conducted, examining works that employ both standard Artificial Intelligence techniques and eXplainable Artificial Intelligence (XAI) approaches. This work aims to provide a solid foundation for researchers in this domain by offering a multi-level analysis: examining proposed solutions (both explainable and non-explainable), analyzing various data types utilized, and exploring relevant medical aspects closely related to the domain. The literature review follows standard PRISMA guidelines for systematic reviews. Major academic platforms were leveraged to extract relevant papers and generate the dataset. We categorized Elderly Fall Detection System applications based on different explainable and non-explainable artificial intelligence techniques. Various evaluation metrics were established to assess the techniques used for prediction, detection, recognition, classification, and identification tasks. The paper concludes with an in-depth discussion of current developments and potential future directions, emphasizing the importance of trustworthy artificial intelligence systems that combine high operating accuracy, transparency, rapid response times, and reliability.


  • Open Access

    Impact of actigraphy-based circadian rest-activity rhythms on functional outcomes in post-stroke rehabilitation

    by Kuan-Lin Sung, Yu-Hsuan Lin, Chen Lin, Chueh-Hung Wu, Hsiang-Chih Chang, Huey-Wen Liang, Shao-Yu Chen, Wei-Chen Hsu