このページは自動翻訳されたものであり、翻訳の正確性は保証されていません。を参照してください。 英語版 ソーステキスト用。

Muscle Synergy-Inspired Functional Electrical Stimulation for Rehabilitating Hand Motor Functions Post-Stroke

2026年11月13日 更新者:Cheung Chi Kwan Vincent、Chinese University of Hong Kong

The goal of this clinical trial is to learn whether muscle-synergy-based functional electrical stimulation works to improve upper limb motor function in stroke subjects. The main questions it aims to answer are:

Does muscle-synergy inspired FES strategy help participant get better motor recovery?

Participants will:

Receive FES or sham stimulation to the affected arm for 30 minutes a day, for 5 sessions.

Continue their usual rehabilitation programme throughout the study. Record clinical assessments and high-density EMGs before and after the treatment.

調査の概要

状態

募集

条件

介入・治療

詳細な説明

Unilateral stroke survivors (N=50) will be recruited and randomly assigned to one of two groups, which will receive usual rehabilitation care only, and usual care plus functional electrical stimulation (FES) training, respectively. Each subject will be evaluated before (T0), after (T1) intervention. At T0, clinical scores (Fugl-Meyer assessment, Brunnstrom stage), video recordings of hand movement, and high-density EMGs will be applied on the patients' forearm to assess the motor functions. At T1, clinical scores, video recordings, and high-density EMGs will be assessed again for characterizing post-rehab functional gain and recovery. Treatments will last ≥5 training sessions depending on the inpatient stay days. All T0 and T1 data will be analyzed offline with muscle synergy analysis and motor unit decomposition analysis.

Clinical scores. Clinical assessments will be conducted at both T0 and T1 to characterize upper extremity motor impairment and evaluate changes in motor recovery between the two assessment time points. Two complementary clinical measures will be used: (1) the Fugl-Meyer Assessment for Upper Extremity (FMA-UE) and (2) the Brunnstrom Stages of Motor Recovery.

The FMA-UE will be used to quantify motor impairment of the affected upper extremity. It assesses reflex activity, voluntary movements within and outside synergistic patterns, wrist and hand motor function, and coordination and movement speed. The motor score ranges from 0 to 66, with higher scores indicating less severe motor impairment. This assessment provides a structured characterization of residual motor abilities and difficulties in selective movement control and is widely used in clinical stroke rehabilitation. Changes in FMA-UE scores between T0 and T1 will be used to quantify changes in upper extremity motor impairment.

The Brunnstrom Stages will be used to characterize the stage of motor recovery following stroke. This assessment describes recovery patterns ranging from flaccidity and limited voluntary movement to the emergence of synergistic movements and, subsequently, more selective and coordinated voluntary control. It complements the FMA-UE by providing an ordinal classification of motor recovery, with particular emphasis on the presence of abnormal synergies and the ability to perform movements independently of these patterns. Comparisons between T0 and T1 will indicate whether participants have progressed to a more advanced recovery stage.

The two assessments are expected to take approximately 30 minutes in total at each time point. Participants' basic demographic characteristics, comorbidities, and other relevant medical conditions will be retrieved from their medical records to characterize the study population and provide clinical context for interpreting the assessment results.

Hand Kinematics recording. Hand kinematic recording. Hand kinematics will be assessed at T0 and T1 using a markerless, multi-camera recording system to quantify extension performance and selective finger control under FES-on and FES-off conditions. The emphasis on extension is motivated by the prognostic relevance of early voluntary finger extension for subsequent upper limb recovery after stroke (Nijland et al., 2010). Extension performance and finger individuation will be evaluated separately, as strength and selective finger control represent distinguishable dimensions of post-stroke hand recovery (Xu et al., 2017). Previous studies have demonstrated the feasibility of video-based markerless assessment of finger movements and upper limb reaching in individuals with stroke (Choudhury et al., 2025; Suwabe et al., 2024).

Participants will be seated comfortably, with the affected forearm supported in pronation, the palm facing downward, and the hand initially in a relaxed resting posture. The support will be arranged to allow unrestricted finger and wrist extension. The initial wrist and finger positions will be documented before each trial, and the forearm position and support configuration will be reproduced across conditions and assessment sessions. Following a standardized verbal cue, participants will perform five tasks: (1) simultaneous extension of all five digits to open the hand; (2) selective thumb extension; (3) selective index finger extension; (4) simultaneous extension of the middle, ring, and little fingers; and (5) wrist extension. Participants will attempt to maintain the achieved position for 5 s before returning to the resting posture. During tasks 2-4, they will be instructed to keep the non-target digits as still as possible without external restraint. During finger tasks, participants will also be instructed to minimize wrist movement and forearm rotation. Each task will be repeated five times per condition, with rest periods between trials to limit fatigue. Condition order will be counterbalanced across participants.

Three USB cameras will be positioned at complementary angles around the hand, with each camera configured to acquire video at 1280 × 720 pixels, in YUYV format, at 60 frames per second. OBS Studio will control recording initiation and termination, and the three camera feeds will be recorded within a single scene using a fixed layout that preserves the native resolution of each view. Camera positions, focus, and lighting will remain unchanged during calibration and task recording. Intrinsic and extrinsic camera parameters will be estimated using a calibration board of known dimensions. Inter-camera timing offsets and temporal stability will be checked using a visual timing signal visible in all views, because simultaneous recording within an OBS scene does not establish hardware synchronization. Stimulation onset and offset will be documented relative to the video timeline.

Offline processing will combine DeepLabCut for two-dimensional landmark detection with Anipose for three-dimensional reconstruction (Karashchuk et al., 2021; Mathis et al., 2018). Composite videos will be separated into individual camera views using fixed crop coordinates, with identical processing applied to calibration recordings. A DeepLabCut model will be trained or fine-tuned using annotated frames covering different participants, hand postures, camera views, and stimulation conditions. Landmarks on the forearm, wrist, hand, and digits will be tracked, and temporally aligned detections will be triangulated using Anipose. Joint angles and digit movements will be calculated relative to anatomical segment coordinate systems to reduce the influence of whole-hand translation and rotation. Tracking accuracy will be evaluated using held-out annotated frames, reprojection errors, and visual inspection. Filtering, short-gap interpolation, and exclusion criteria will be specified before outcome analysis.

Kinematic outcomes will include joint extension excursion, peak extension velocity, maximum continuous holding duration, and unintended movement of non-target digits. Extension task success will be defined as reaching the prespecified task-specific joint-angle targets within the allowed response period and maintaining them continuously for 5 s without prohibited assistance or compensatory movement. For tasks 2-4, selective extension success will additionally require non-target digit movements to remain within predefined tolerances throughout the movement and holding phases. Target angles, response periods, and movement tolerances will be established through pilot testing, considering measurement error, passive range of motion, and physiological finger coupling, and will be fixed before the main study (Lang and Schieber, 2004). Task success rates will be calculated separately for each task and condition as the number of successful trials divided by the number of assessable trials, multiplied by 100%. Unsuccessful movement attempts will remain in the denominator, whereas technically unassessable trials will be documented separately. FES-on versus FES-off comparisons will characterize immediate stimulation-assisted performance, while changes in FES-off performance between T0 and T1 will characterize changes in voluntary motor performance.

High-density EMGs recording. High-density surface electromyography (HD-sEMG) will be recorded at T0 and T1 to characterize forearm muscle activity during voluntary hand and wrist movements without FES. HD-sEMG enables the assessment of the spatial distribution of muscle electrical activity across the skin surface (Gallina et al., 2022). Signals will be acquired using a TMSi recording system at a sampling rate of 2048 Hz, synchronized with the three-camera kinematic recording system.

Two customized flexible electrode arrays will be positioned around the middle third of the affected forearm. Each array will contain 64 recording channels arranged into eight longitudinal electrode columns (threads), with eight electrodes per thread and a longitudinal interelectrode spacing of 1 cm. Channels 1-64 will cover the palmar aspect of the forearm, while channels 65-128 will cover the dorsal aspect, providing 128 recording channels distributed across 16 threads around the forearm circumference. The spacing between adjacent threads will vary according to forearm circumference. A measuring tape will guide placement to distribute the threads as evenly as possible, and the position of the first thread will be marked on the skin to guide consistent array alignment across assessments.

During recording, participants will be seated with the affected forearm supported in pronation, the palm facing downward, and the hand initially in a relaxed resting posture. Participants will perform five tasks in a fixed sequence: (1) simultaneous extension of all five digits; (2) selective thumb extension; (3) selective index finger extension; (4) simultaneous extension of the middle, ring, and little fingers; and (5) wrist extension. For each task, participants will attempt to maintain the achieved position for 5 s before returning to the resting posture. During selective digit tasks, participants will be instructed to minimize movement of the non-target digits without external restraint. The five-task sequence will be repeated eight times, yielding eight trials per task and 40 trials per assessment session. This predictable sequence is intended to reduce cognitive demands and distribute repetitions of each task across the recording session rather than requiring consecutive repetitions of the same movement. EMG and video recordings will include unsuccessful movement attempts as well as successfully completed trials.

研究の種類

介入

入学 (推定)

50

段階

  • 適用できない

連絡先と場所

このセクションには、調査を実施する担当者の連絡先の詳細と、この調査が実施されている場所に関する情報が記載されています。

研究連絡先

研究場所

    • Guangdong
      • Guangzhou、Guangdong、中国
        • 募集
        • Guangdong Work Injury Rehabilitation Hospital
        • コンタクト:

参加基準

研究者は、適格基準と呼ばれる特定の説明に適合する人を探します。これらの基準のいくつかの例は、人の一般的な健康状態または以前の治療です。

適格基準

就学可能な年齢

18年~80年 (大人、高齢者)

健康ボランティアの受け入れ

はい

受講資格のある性別

全て

説明

Inclusion Criteria:

  • Unilateral stroke
  • Normal Cognition

Exclusion Criteria:

  • Bilateral stroke
  • Upper limb injury

研究計画

このセクションでは、研究がどのように設計され、研究が何を測定しているかなど、研究計画の詳細を提供します。

研究はどのように設計されていますか?

デザインの詳細

  • 主な目的:処理
  • 割り当て:ランダム化
  • 介入モデル:並列代入
  • マスキング:独身

武器と介入

参加者グループ / アーム
介入・治療
実験的:Apply personalized functional electrical stimulation based on high-density EMG assessment

Apply personalized functional electrical stimulation to upper limb of stroke patients

偽コンパレータ:Apply functional electrical stimulation that targets dorsal forearm

Apply personalized functional electrical stimulation to upper limb of stroke patients

この研究は何を測定していますか?

主要な結果の測定

結果測定
メジャーの説明
時間枠
Fugl-Meyer assessment - upper extremity
時間枠:Two weeks

Two-stage of assessment including pre-assessment and post-assessment

Two weeks
High-density EMG assessment
時間枠:two weeks

High-density EMG assessment

two weeks

協力者と研究者

ここでは、この調査に関係する人々や組織を見つけることができます。

スポンサー

協力者

捜査官

  • 主任研究者:Vincent Cheung, Doctor of Philosophy、The Chinese University of Hong Kong, Gerald Choa Neuroscience Institute

研究記録日

これらの日付は、ClinicalTrials.gov への研究記録と要約結果の提出の進捗状況を追跡します。研究記録と報告された結果は、国立医学図書館 (NLM) によって審査され、公開 Web サイトに掲載される前に、特定の品質管理基準を満たしていることが確認されます。

主要日程の研究

研究開始 (推定)

2026年11月16日

一次修了 (推定)

2027年12月30日

研究の完了 (推定)

2028年4月15日

試験登録日

最初に提出

2026年11月13日

QC基準を満たした最初の提出物

2026年11月13日

最初の投稿 (実際)

2026年11月18日

学習記録の更新

投稿された最後の更新 (実際)

2026年11月18日

QC基準を満たした最後の更新が送信されました

2026年11月13日

最終確認日

2026年11月1日

詳しくは

本研究に関する用語

その他の研究ID番号

  • FES_HDEMG_UE_V1
  • GHP/166/23SZ (その他の識別子:Innovation and Technology Commission, Hong Kong)

個々の参加者データ (IPD) の計画

個々の参加者データ (IPD) を共有する予定はありますか?

未定

医薬品およびデバイス情報、研究文書

米国FDA規制医薬品の研究

いいえ

米国FDA規制機器製品の研究

いいえ

この情報は、Web サイト clinicaltrials.gov から変更なしで直接取得したものです。研究の詳細を変更、削除、または更新するリクエストがある場合は、register@clinicaltrials.gov。 までご連絡ください。 clinicaltrials.gov に変更が加えられるとすぐに、ウェブサイトでも自動的に更新されます。