Mapping Trajectories of Gait Recovery in Clinical Stroke Rehabilitation

Neurorehabilitation and neural repair, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 14, 2025

Background How gait changes during the early stages of stoke rehabilitation, and which patient characteristics are associated with these is still largely unknown. Objective he first objective was to describe in stroke rehabilitation. Secondly, we determined how various were rate change over time. Methods Participants measured every 3 weeks The assessment consisted an inertial measurement unit (IMU) based 2-minute walk test (2MWT), IMU-based balance tests, standard clinical tests. In 2MWT, participants equipped IMUs, from speed, variability, asymmetry, smoothness calculated. examined admission discharge at individual level. effect on features time assessed growth models. Results A total 81 Trajectories 72 analyzed. On basis, speed increased 32 trajectories. Only a few trajectories exhibited significant rehabilitation period. models revealed increase decrease variability smoothness. Berg Balance Scale onset (negatively) rates smoothness, respectively. Conclusion We found substantial gait-feature outcomes their progression individuals after studied had limited associations

Language: Английский

---

Evaluation of Upper Body and Lower Limbs Kinematics through an IMU-Based Medical System: A Comparative Study with the Optoelectronic System

Sensors, Journal Year: 2023, Volume and Issue: 23(13), P. 6156 - 6156

Published: July 5, 2023

In recent years, the use of inertial-based systems has been applied to remote rehabilitation, opening new perspectives for outpatient assessment. this study, we assessed accuracy and concurrent validity angular measurements provided by an device rehabilitation with respect state-of-the-art system motion tracking. Data were simultaneously collected two across a set exercises trunk lower limbs, performed 21 healthy participants. Additionally, sensitivity inertial measurement unit (IMU)-based its malpositioning was assessed. Root mean square error (RMSE) used explore differences in outputs terms range (ROM), their agreement via Pearson’s correlation coefficient (PCC) Lin’s concordance (CCC). The results showed that IMU-based able assess upper-body lower-limb kinematics general than 5° moderately biased mispositioning. Although does not seem be suitable analysis requiring high level detail, findings study support application programs unsupervised settings, providing reliable data remotely monitor progress pathway change patient’s motor function.

Language: Английский

---

Latent Space Representation of Adversarial AutoEncoder for Human Activity Recognition: Application to a low-cost commercial force plate and inertial measurement units

Smart Health, Journal Year: 2025, Volume and Issue: unknown, P. 100537 - 100537

Published: Jan. 1, 2025

Language: Английский

---

An open-source, externally validated neural network algorithm to recognize daily-life gait of older adults based on the lower-back sensor

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 5, 2025

Abstract Background Accurate gait recognition from daily physical activities is a critical first step for further fall risk assessment and rehabilitation monitoring based on inertial sensors. However, most openly available models are healthy young adults ambulating in structured conditions. Objective This study aimed to develop an open-source externally validated algorithm daily-life of older acceleration angular velocity, as well data only, explore the effect use augmentation model training. Methods A convolutional neural network was trained recognition. The training lower-back sensor 20 (mean age 76 years old), with annotated synchronized activity labels semi-structured randomly split into training, validation, testing datasets by participants, multiple times using these different splits. six channels (accelerations velocities) three only) under conditions without augmentation, respectively. External validation evaluated collected 47 stroke survivors 72.3 old) balance walking tests. Results For dataset, median accuracy ranged 94 % 98 %, precision 63 85 sensitivity 95 97 F1-score 90 specificity %. external 100 99.9 71 83 Conclusions Based lower-back-worn data, we provide accurate, open-source, adults, one six-axis input another three-axis data. Besides, found when model, especially helpful only.

Language: Английский

---

Assessment of ambulation functions through kinematic analysis in individuals with stroke: a systematic review

European Journal of Physical and Rehabilitation Medicine, Journal Year: 2025, Volume and Issue: 61(1)

Published: Feb. 1, 2025

Although kinematic assessments for stroke-induced lower limb impairments offer a promising alternative to conventional scale evaluations, interpreting high-dimensional data remains challenging due numerous metrics reported in past studies. This study aimed provide an exhaustive overview of existing studies using kinematics assess the gait individuals with stroke, along examining their clinimetric properties future clinical applications. A systematic search was conducted across PubMed (08/2024), Scopus Web Science CINAHL EMBASE and IEEE (08/2024). We included articles that recruited over 18 years old stroke utilized motion capture technologies evaluate kinematics. Similar were consolidated analysis, COSMIN Risk Bias Checklist used methodological quality investigating metrics. Convergent validity evaluated by association Fugl-Meyer limbs walking speed. Moreover, GRADE approach rate evidence. total 383 classified into 10 categories. Seven on metric reliability rated high quality. Metrics satisfactory spatiotemporal, spatial metrics, data-driven score. Six assessed convergent validity. The dynamic index, angular component coefficient correspondence (ACC), change cadence, stride length, hip range showed Among 13 studies, 12 as moderate evidence approach. There are significant variations measurements high-quality evaluating scarce. For more standardized evidence-based assessment, further research validating these assessments' is essential.

Language: Английский

---

Berg Balance Scale Scoring System for Balance Evaluation by Leveraging Attention-Based Deep Learning with Wearable IMU Sensors

Bioengineering, Journal Year: 2025, Volume and Issue: 12(4), P. 395 - 395

Published: April 7, 2025

Balance assessment is crucial for health monitoring and rehabilitation evaluation of neurological diseases like Parkinson’s disease (PD) stroke. The Berg Scale (BBS) a widely used clinical tool balance evaluation. However, its dependence on trained therapists subjective, time-consuming assessments limits scalability. Current researchers have proposed several automated systems. they suffer from difficulty in use settings the need feature engineering. rapid advancement wearable inertial measurement units (IMUs) provides an objective motion analysis that suitable environments. Thus, to address limitations manual scoring complexities capturing gait features, we BBS system using attention-based deep learning algorithm with IMU data, integrating convolutional neural networks (CNNs) spatial extraction, bidirectional long short-term memory (Bi-LSTM) temporal modeling, attention mechanisms emphasize informative features. Validated 20 healthy subjects (young elderly) patients (PD stroke), achieved mean absolute error (MAE) 1.1627 root squared (RMSE) 1.5333. Requiring only 5 min walking this approach provided efficient, solution assist healthcare physicians as well their own monitoring. key included: limited generalizability severely impaired who were unable walk independently, inability predict score individual tasks.

Language: Английский

---

Predicting community walking after stroke

Frontiers in Stroke, Journal Year: 2025, Volume and Issue: 4

Published: April 9, 2025

Introduction A key element of personalized stroke rehabilitation is early prediction an individual's potential to walk in the community. Objective We aim determine predictive value patient characteristics, clinical test results, and Inertial Measurement Units (IMU) based balance, gait daily-life measures, measured at admission discharge rehabilitation, for community walking 6 months after stroke. Methods Data were collected from people during post The assessment consisted IMU-based 2-min (2MWT), three balance tests, measurement daily life, several standard including Berg Balance Scale, Barthel Index, Functional Ambulation Categories, Motricity Index (MI), Trunk Control Test (TCT). At 6-months, life was with IMU two consecutive days. From this measurement, features calculated, namely strides per day, average maximum speed. assessed gait, tests characteristics predicting measures univariate ordinary least squares regression. Subsequently, significant predictors included a multivariate Results Thirty-five individuals included. Ordinary regression analysis indicated that age, day had step count months. For speed months, 2MWT speed, TCT, MI baseline predictors. Multivariate outcomes more than discharge, adjusted R 2 values models 0.60, 0.42, 0.53, respectively. Conclusions Age, trunk stability (TCT), affected leg strength 6-months Future research larger sample size required refine these findings.

Language: Английский

---

Improving the prediction of functional recovery in older adults with stroke in geriatric rehabilitation using an inertial measurement unit combined with the Utrecht Scale for Evaluation of Rehabilitation (USER)

Journal of Rehabilitation Medicine – Clinical Communications, Journal Year: 2025, Volume and Issue: 8, P. jrmcc43129 - jrmcc43129

Published: May 7, 2025

Background: Prediction of functional recovery in older adults recovering from stroke is typically based on observational scales, such as the Utrecht Scale for Evaluation Rehabilitation (USER). Objectively measuring postural sway using inertial measurement devices (IMU) may complement or improve conventional approaches. The aim this study was to evaluate whether integrating an IMU with USER data enhances accuracy predicting at discharge. Methods: This prospective cohort includ-ed (≥ 65 years) stroke. Postural assessed during 2 different balance conditions and analysed principal component analysis (PCA). Using 3 regression models, percentage explained variance compared assess predictive performance vs IMU. Results: 71 patients included had a mean age 78 (SD 7.6) median time since 16 days (IQR 19–60). Of patients, 12 (16.9%) were unable perform condition due insufficient balance. 35 features displaying reliability both conditions, selected PCA. Incorporation components final model increased which only USER-mobility admission used predict delta-USER discharge (R2 = 0.61 0.30). Conclusions: Sitting standing measured by improves prediction alone.

Language: Английский

---

Balance Assessment Using a Handheld Smartphone with Principal Component Analysis for Anatomical Calibration

Sensors, Journal Year: 2024, Volume and Issue: 24(17), P. 5467 - 5467

Published: Aug. 23, 2024

Most balance assessment studies using inertial measurement units (IMUs) in smartphones use a body strap and assume the alignment of smartphone with anatomical axes. To replace need for strap, we have used an method that employs calibration maneuver Principal Component Analysis (PCA) so can be held by user comfortable position. The objectives this study were to determine if correlations existed between angular velocity scores derived from handheld PCA functional vs. placed assumed alignment, analyze acceleration score differences across poses increasing difficulty. exhibited moderately strongly correlated (r = 0.487-0.983,

Language: Английский

---

Walk Longer! Using Wearable Inertial Sensors to Uncover Which Gait Aspects Should Be Treated to Increase Walking Endurance in People with Multiple Sclerosis

Sensors, Journal Year: 2024, Volume and Issue: 24(22), P. 7284 - 7284

Published: Nov. 14, 2024

Reduced walking endurance is common in people with multiple sclerosis (PwMS), leading to reduced social participation and increased fall risk. This highlights the importance of identifying which gait aspects should be mostly targeted by rehabilitation maintain/increase this population. A total 56 PwMS 24 healthy subjects (HSs) executed 6 min walk test (6 MWT), a clinical measure endurance, wearing three inertial sensors (IMUs) on their shanks lower back. Five IMU-based digital metrics descriptive different domains, i.e., double support duration, trunk sway, regularity, symmetry, local dynamic instability, were computed. All demonstrated moderate-high ability discriminate between HSs (AUC: 0.79-0.91) able detect differences at minimal (PwMS

Language: Английский

---