Background: Recent advancements in wearable sensor technology have shown the feasibility of remote physical therapy at home. In particular, the current COVID-19 pandemic has revealed the need and opportunity of internet-based wearable technology in future health care systems. Previous research has shown the feasibility of human activity recognition technologies for monitoring rehabilitation activities in home environments; however, few comprehensive studies ranging from development to clinical evaluation exist.Objective: This study aimed to (1) develop a home-based rehabilitation (HBR) system that can recognize and record the type and frequency of rehabilitation exercises conducted by the user using a smartwatch and smartphone app equipped with a machine learning (ML) algorithm and (2) evaluate the efficacy of the home-based rehabilitation system through a prospective comparative study with chronic stroke survivors.Methods: The HBR system involves an off-the-shelf smartwatch, a smartphone, and custom-developed apps. A convolutional neural network was used to train the ML algorithm for detecting home exercises. To determine the most accurate way for detecting the type of home exercise, we compared accuracy results with the data sets of personal or total data and accelerometer, gyroscope, or accelerometer combined with gyroscope data. From March 2018 to February 2019, we conducted a clinical study with two groups of stroke survivors. In total, 17 and 6 participants were enrolled for statistical analysis in the HBR group and control group, respectively. To measure clinical outcomes, we performed the Wolf Motor Function Test (WMFT), Fugl-Meyer Assessment of Upper Extremity, grip power test, Beck Depression Inventory, and range of motion (ROM) assessment of the shoulder joint at 0, 6, and 12 months, and at a follow-up assessment 6 weeks after retrieving the HBR system.Results: The ML model created with personal data involving accelerometer combined with gyroscope data (5590/5601, 99.80%) was the most accurate compared with accelerometer (5496/5601, 98.13%) or gyroscope data (5381/5601, 96.07%). In the comparative study, the drop-out rates in the control and HBR groups were 40% (4/10) and 22% (5/22) at 12 weeks and 100% (10/10) and 45% (10/22) at 18 weeks, respectively. The HBR group (n=17) showed a significant improvement in the mean WMFT score (P=.02) and ROM of flexion (P=.004) and internal rotation (P=.001). The control group (n=6) showed a significant change only in shoulder internal rotation (P=.03).Conclusions: This study found that a home care system using a commercial smartwatch and ML model can facilitate participation in home training and improve the functional score of the WMFT and shoulder ROM of flexion and internal rotation in the treatment of patients with chronic stroke. This strategy can possibly be a cost-effective tool for the home care treatment of stroke survivors in the future.Trial Registration: Clinical Research Information Service KCT0004818; https://tinyurl.com/y92w978t

To propose an artificial intelligence (AI)-based decision-making rule in modified Ashworth scale (MAS) that draws maximum agreement from multiple human raters and to analyze how various biomechanical parameters affect scores in MAS. Prospective observational study. Two university hospitals. Hemiplegic adults with elbow flexor spasticity due to acquired brain injury (N=34). Not applicable. Twenty-eight rehabilitation doctors and occupational therapists examined MAS of elbow flexors in 34 subjects with hemiplegia due to acquired brain injury while the MAS score and biomechanical data (ie, joint motion and resistance) were collected. Nine biomechanical parameters that quantify spastic response described by the joint motion and resistance were calculated. An AI algorithm (or artificial neural network) was trained to predict the MAS score from the parameters. Afterwards, the contribution of each parameter for determining MAS scores was analyzed. The trained AI agreed with the human raters for the majority (82.2%, Cohen's kappa=0.743) of data. The MAS scores chosen by the AI and human raters showed a strong correlation (correlation coefficient=0.825). Each biomechanical parameter contributed differently to the different MAS scores. Overall, angle of catch, maximum stretching speed, and maximum resistance were the most relevant parameters that affected the AI decision. AI can successfully learn clinical assessment of spasticity with good agreement with multiple human raters. In addition, we could analyze which factors of spastic response are considered important by the human raters in assessing spasticity by observing how AI learns the expert decision. It should be noted that few data were collected for MAS3; the results and analysis related to MAS3 therefore have limited supporting evidence.

Background: Rehabilitation in cerebral palsy (CP) seeks to harness neuroplasticity to improve movement, including walking, yet cortical activation underlying gait is not well understood. Methods: We used electroencephalography (EEG) to compare motor related cortical activity, measured by mu rhythm, during quiet standing and treadmill walking in 10 children with unilateral CP and 10 age- and sex-matched children with typical development (TD). Peak mu band frequency, mu rhythm desynchronization (MRD), and gait related intra- and inter-hemispheric coherence were examined. Results: MRD during walking was observed bilaterally over motor cortex in both cohorts but peak mu band frequency showing MRD was significantly lower in CP compared to TD. Coherence during quiet standing between motor and frontal regions was significantly higher in the non-dominant compared to dominant hemisphere in CP with no hemispheric differences in TD. Conclusions: EEG-based measures should be further investigated as clinical biomarkers for atypical motor development and to assess rehabilitation effectiveness.

Children with cerebral palsy typically exhibit reduced complexity of muscle coordination patterns during walking; however, the specific patterns that characterize their gait abnormalities are still not well documented. This study aimed to identify the specific repertoire of muscle coordination patterns in children with CP during walking compared to same-aged peers without CP and their relationships to gait performance. To identify muscle coordination patterns, we extracted muscle synergies from 10 children with CP and 10 age-matched typically developing children (TD). K-mean clustering and discriminant analyses of all extracted synergies were used to group similar synergies. Then, weight-averaged z-scores were quantified for each cluster to determine their group-specific level. In this cohort, 10 of the 17 distinct clusters were largely CP-specific while six clusters were seen mainly in TD, and one was non-specific. CP-specific clusters generally showed merging of two TD synergies, excessive antagonist co-activation, decreased muscle activation compared to TD, and complex or atypical pattern. Significant correlations were found between weight-averaged z-scores and step length asymmetry, cadence asymmetry, self-selected treadmill speed and AP-COM displacement of the pelvis such that greater CP-specificity of muscle synergies was related to poorer performance, thus indicating that CP-specific synergies can influence motor dysfunction.

Despite external environmental changes in walking, such as manipulating gait speed, previous studies have shown that the underlying muscle synergy structures (synergy weights or vectors) rarely vary. The purpose of this study is to examine if external environmental changes to the walking task influence muscle synergies in children with cerebral palsy (CP) and/or typical development (TD). To identify muscle synergies, we extracted muscle synergies from eight children with CP and eight age-matched TD in three treadmill walking conditions, e.g., baseline (adjusted to individual comfortable walking speed), variable speed (VS), or restricted foot width (RW). Then, we grouped similar muscle synergies using k-mean clustering and discriminant analyses from all datasets of individual synergy structures. Proportion tests exhibited six clusters of muscle synergies predominantly arising from children with CP and four clusters from children with TD. Also, the proportion of muscle synergies was significantly different in four of the CP-preferred clusters across conditions. Specifically, the proportion of the baseline condition was significantly different from VS and RW conditions in one and two clusters, respectively. The proportion was significantly different between VS and RW conditions in four clusters. Cadence and step lengths differed across conditions but not groups which makes the group differences in proportion even more notable. In contrast, step width, while significantly lower in CP, did not differ across conditions. Our findings demonstrate that muscle synergies in children with CP are more sensitive to changes in the external walking environment than in typically developing children.

Objective . The development of non-invasive, quickly reversible techniques for controlling undesired muscle force production (e.g. spasticity) could expand rehabilitation approaches in those with pathology by increasing the type and intensity of exercises that can be performed. High-frequency alternating current (HFAC) has been previously established as a viable method for blocking neural conduction in peripheral nerves. However, clinical application of HFAC for nerve conduction block is limited due to the invasiveness of surgical procedures and the painful onset response. This study aimed to examine the use of transcutaneous HFAC (tHFAC) at various stimulation frequencies to address these shortfalls. Approach . Ten individuals participated in the study. Surface electrodes were utilized to apply tHFAC (0.5–12 kHz) to the median and ulnar nerves. Individual pain threshold was determined by gradual increase of stimulation amplitude. Subjects then performed a force-matching task by producing grip forces up to the maximal voluntary contraction level with and without application of tHFAC below the pain threshold. Main results . Pain threshold current amplitude increased linearly with stimulation frequency. Statistical analysis showed that both stimulation frequency and charge injected per phase had significant effects ( p < 0.05) on grip force reduction. At the group level, application of tHFAC below pain threshold reduced grip force by a maximum of 40.7% ± 8.1%. Baseline grip force trials interspersed between tHFAC trials showed consistent grip force, indicating that fatigue was not a factor in force reduction. Significance . Our results demonstrate the effectiveness of tHFAC at reducing muscle force when applied below the pain threshold, suggesting its potential clinical viability. Future studies are necessary to further elucidate the mechanism of force reduction before clinical application.