Exoskeletons are a fast-growing technology that enables multiple use-cases in clinical scenarios. They can be useful tools for the rehabilitation of patients with motor dysfunctions caused by neurological conditions, aging or trauma. Assistive exoskeletons modulate the torque exerted by the electrical motors moving their joints to allow the patients wearing them to achieve an intended movement, such as gait, correctly. Their effectiveness, therefore, requires accurate online control of such torques to complement those generated by the patient. Hereby we explored Deep Learning (DL) models to generate an online prediction of the gait phase, i.e., stance or swing, during assisted walking with a lower-limb exoskeleton based on surface electromyography (sEMG) data. We leveraged the lead of muscular activation with respect to the movement of the limbs to adjust the labeling based on joints kinematics. The cross-subject design allowed to generalize over subjects not considered for training A hyperparameter optimization algorithm was also implemented to further explore the capabilities of DL models of a reduced size. We simulated a use case scenario to assess whether online implementation of the proposed technique is feasible. We also proposed a new metric called trade-of score (TOS) for evaluating the cost-performance compromise of the optimized models which lead to identifying a DL model capable of classifying gait phases with an accuracy of about 95% while significantly reducing the number of parameters compared to the full architecture. Its mean computational time of less than 10 ms offers the opportunity for accurate, online exoskeleton control based on sEMG data.

Background. This study aimed to quantify differences in the internal training load (ITL) of an elite rink hockey (RH) team across days within and between three types of microcycles: pre-season, in-season regular, and in-season congested, to provide insights to optimise microcycle scheduling. Methods. One international-level male RH team comprising seven outfielders (29.6 ± 4.7 years; height, 178.9 ± 2.3 cm; body mass, 77.8 ± 5.7 kg) and one goalkeeper (32 years; height, 180.4 cm; body mass, 83.6 kg) was monitored for 21 microcycles. The ITL was assessed using the session rate of perceived exertion (sRPE) and quantified as time based on a triphasic classification commonly utilised in team sports: low-intensity training (LIT, 90% HRmax). Generalized estimating equations were used to examine differences across within-microcycle training days and between seasonal phases, with linear mixed models applied as sensitivity analyses. Results. Across all phases, significant day-to-day variations in ITL were observed within microcycles (all p < 0.001), with both subjective (sRPE) and objective (LIT–HIT) ITLs progressively decreasing as match days (MDs) approached, showing moderate-to-large population-averaged effects with 95% confidence intervals consistently not crossing zero. The pre-season exhibited the highest overall ITL (p < 0.001), characterised by a substantially greater sRPE and increased time spent across all intensity zones, with the largest magnitudes observed for LIT and MIT compared with the in-season phases. Conclusions. Findings suggest that an international-level RH team progressively reduced the ITL as MDs approached with the highest loads scheduled earlier within microcycles. Moreover, the pre-season had the highest ITLs. This ITL distribution may provide useful guidance for RH coaches and support staff in optimising microcycle planning.

Quantifying physical activity (PA) is clinically relevant in facioscapulohumeral muscular dystrophy (FSHD). Wearable accelerometers provide objective measu

Introduction/Aims: To investigate its potential role as a marker of disease severity in facioscapulohumeral dystrophy (FSHD), this study examined the association between whole‐body phase angle (PhA) and clinically assessed severity in FSHD patients. Methods: Patients were evaluated for PhA using bioelectrical impedance analysis ( BIA ) and for disease severity using the FSHD clinical score. In a subgroup, physical activity level ( PAL ) was assessed using a questionnaire. Pearson correlations were performed in the total cohort and PAL ‐assessed subgroup, stratified by sex and clinical phenotype (category A and non‐A) according to the Comprehensive Clinical Evaluation Form. Multivariate regression analyses were conducted adjusting for age, body mass index, and PAL . Due to the small sample, correlation was not performed in PAL ‐assessed non‐A patients, while multivariate regression was not performed in non‐A patients and PAL ‐assessed stratified subgroups. Results: Seventy‐two patients (mean age 39.94 ± 17.52 years; 27 females and 45 males; 58 category A and 14 non‐A) were enrolled. The PAL ‐assessed subgroup comprised 38 patients (mean age 41.39 ± 14.98 years; 16 females and 22 males; 31 category A and 7 non‐A). Pearson analyses revealed that PhA was inversely associated with clinical score in all groups ( r –0.63 to –0.84, p 0.02 to < 0.001). Multivariate regression confirmed PhA as the sole independent predictor across all groups ( β = –0.64 to –0.93, all p < 0.001). Discussion: These results support the potential role of PhA as a marker of disease severity in FSHD . Longitudinal studies are warranted to determine its sensitivity for monitoring disease progression over time.