Roger Gassert

Introduction: In recent years, the growing demand for cost-effective rehabilitation after stroke has motivated the development of novel technologies to support high-intensity rehabilitation. However, available knowledge on the economic benefits of stroke technology-assisted rehabilitation (TAR) is relatively scarce and fragmented. Therefore, this scoping review aimed to review evidence on the economic benefits of TAR in laboratory, clinical, and home settings, as well as the methodologies used to evaluate these outcomes. Methods: A systematic search of four electronic bibliographic databases and two additional grey literature/ economic evidence sources (until 03 June 2026) identified 6469 records. Following predefined inclusion and exclusion criteria, 44 records were considered for further analysis. Additionally, 21 records were identified through a citation search. Results: In total, 65 articles reporting TAR-related costs were included, but only a few of these studies reported cost analyses (mostly cost comparisons). Firstly, of the 38 studies conducted in laboratory settings, only one reported a formal TAR-related cost analysis. Secondly, out of 16 studies reporting costs, 11 studies performed cost analysis for TAR in clinical settings. Of these, the majority (six) estimated TAR to be equally effective or even superior (one) and cheaper than conventional therapy (CT) in the clinic, but only when TAR enabled the therapist to treat several stroke survivors in parallel. This was captured using the relative therapist-to-patient ratio (rT/P), indicating that staffing and supervision models were more important than device cost. Lastly, seven of 13 studies conducted TAR-related cost analysis in a home setting. Three studies suggested that TAR may be economically beneficial compared with CT in the clinic. Conclusion: Empirical findings suggest that TAR may generate cost savings compared to CT in the clinic, particularly when rT/P is < 0.5. However, the current evidence is sparse, with limited comparability and generalisability due to heterogeneous TAR modalities, comparators and cost-calculation methods. Clinical trial number: Not applicable.

Study Objectives: Approximately 4.7 million subjects worldwide suffer from narcolepsy (type 1 and 2) and idiopathic hypersomnia and are impaired in their sleep and wakefulness behavior. However, the role of the circadian rhythm in their sleep-wake behavior remains largely unknown. With this study, we investigated the influence of the internal clock on vigilance and sleep in central disorders of hypersomnolence. Methods: We implemented a nap protocol with 10 cycles of 80 minutes sleep and 160 minutes wake time and compared various sleep and wake parameters (vigilance, subjective sleepiness, sleep efficiency, stages of sleep, slow wave activity) of patients with idiopathic hypersomnia (N=12, ∅ 25.8 ± 4.1 years, 10 females) and narcolepsy type 1 (N=12, ∅ 25.0 ± 4.4 years, 8 females) with a healthy control group (N=12, ∅ 26.8 ± 4.7 years, 6 females) using a linear mixed-model analysis. Results: Our protocol successfully disentangled the circadian rhythm from homeostatic sleep pressure and revealed an intact circadian melatonin pattern as assessed by dim light melatonin onset and offset in both patient groups. Patients with narcolepsy showed high sleep efficiencies of ∅ 94.3 ± 5.3% over all naps (group effect p < 0.001) and responded positively to the short sleep episodes with an increase in vigilance in the late afternoon (PVT speed @wake seven 2.5 ± 1.0 vs @wake ten 3.9 ± 0.6, p < 0.001). Patients with idiopathic hypersomnia showed an increased subjective sleepiness (group effect p < 0.001), yet practically no statistically significant differences in sleep parameters compared to healthy controls. Surprisingly, we observed a high number of SOREMPs in the healthy control group under low sleep pressure - a finding not previously reported in literature. Conclusion: These findings argue against circadian disruption as a primary mechanism in idiopathic hypersomnia and point toward other underlying causes, such as altered sleep homeostasis or neurochemical dysregulation. Plain Language Summary: This study is the first to test a nap schedule in people with narcolepsy type 1 and idiopathic hypersomnia that filters out the influence of the internal clock. Both conditions lack the brain’s ability to regulate sleep and wakefulness effectively. Surprisingly, we found that both groups had a normal daily rhythm, which contradicts some earlier studies. This suggests that idiopathic hypersomnia may not be caused by a problem with the internal clock. We examined various sleep-related measures and how they change over the day. Our findings help us better understand these conditions and show how well-timed naps can support treatment. For patients, the results offer practical advice on how to use their rhythm to stay more alert during the day.

Marker-based Optical Motion Capture (OMC) paired with biomechanical modeling is currently considered the most precise and accurate method for measuring human movement kinematics. However, combining differentiable biomechanical modeling with Markerless Motion Capture (MMC) offers a promising approach to motion capture in clinical settings, requiring only minimal equipment, such as webcams, and minimal effort for data collection. This study compares key kinematic outcomes from biomechanically modeled MMC and OMC data in 15 individuals with stroke performing the drinking task, a functional task recommended for assessing upper limb movement quality. We observed a high level of agreement in kinematic trajectories between MMC and OMC, as indicated by high correlations (median r > 0.95 for the majority of kinematic trajectories) and median RMSE (root mean squared error) values ranging from 2∘-5∘ for joint angles, 0.04 m/s for end-effector velocity, and 6mm for trunk displacement. Trial-to-trial biases between OMC and MMC were consistent within participant sessions, with interquartile ranges of bias around 1-3∘ for joint angles, 0.01m/s in end-effector velocity, and approximately 3mm for trunk displacement. Our findings indicate that our MMC for arm tracking is approaching the accuracy of marker-based methods, supporting its potential for use in clinical settings. MMC could provide valuable insights into movement rehabilitation after stroke, potentially enhancing the effectiveness of rehabilitation strategies.

BACKGROUND Stroke affects over 12 million people annually, with up to 80% experiencing gait impairments three months post-event. While rehabilitation can improve mobility, outcomes often vary due to limited personalization of conventional therapy. Real-time gait biofeedback, defined as the delivery of immediate sensory information about motor performance to guide movement, shows promise for enhancing motor learning. However, current biofeedback designs are typically developed without end-user input, potentially reducing end-users’ understanding, enjoyment, and engagement. OBJECTIVE The objective of this study was to investigate the user experience of people with stroke involved in co-designing personalized real-time gait biofeedback and its associated virtual environment for rehabilitation. A mixed-methods approach was used to identify preferred biofeedback features and assess the impact of personalization on task load and motivation. METHODS Ten post-stroke participants completed multiple sessions categorized as kick-off, discovery, or consolidation, each lasting approximately 1 hour. After an initial assessment of gait deficits, an iterative co-design process was used to progressively refine the biofeedback and virtual environment based on individual input. During these sessions, participants walked in the CAREN while receiving real-time biofeedback based on a therapist-selected gait parameter and collaborated with the experimenter to modify feedback modalities and design. Qualitative data were collected via unstructured interviews to determine the preferred biofeedback features, while quantitative data were collected through the raw NASA-Task Load Index and Intrinsic Motivation Inventory. RESULTS Each participant developed a unique biofeedback and virtual environment setup. Despite this, common preferences emerged, including positive feedback, avatars that mirrored participant’s movements, and multimodal visual-auditory hit/error indicators. Age-related differences in motivation were observed across sessions, with participants over 60 showing greater motivation as personalization evolved. When examining the percent change in scores between the first discovery and last consolidation sessions, no significant changes were found in task load (t₈=–1.13, adjusted P=.366), value/usefulness (z=21.00, adjusted P=.858), or effort/importance (t₉=1.29, adjusted P=.366). However, a marginally significant 13% increase in perceived competence (z=0.00, adjusted P=.058) suggested participants felt more capable of achieving the target parameter value with personalized feedback. Success/error counters led to minor individual-level changes in enjoyment, effort, and competence, though no single feature significantly influenced overall outcomes. CONCLUSIONS This study highlights the value of co-design in developing personalized biofeedback for gait rehabilitation, revealing consistent preferences for features such as positive feedback, avatars, and multimodal cues. Age-related differences in motivation also emerged, suggesting that personalization may be particularly beneficial for older adults. CLINICALTRIAL Ethical approval was obtained from the ETH Zurich Ethics Commission (EK-2022-N-173).

Marker-based Optical Motion Capture (OMC) paired with biomechanical modeling is currently considered the most precise and accurate method for measuring human movement kinematics. However, combining differentiable biomechanical modeling with Markerless Motion Capture (MMC) offers a promising approach to motion capture in clinical settings, requiring only minimal equipment, such as synchronized webcams, and minimal effort for data collection. This study compares key kinematic outcomes from biomechanically modeled MMC and OMC data in 15 stroke patients performing the drinking task, a functional task recommended for assessing upper limb movement quality. We observed a high level of agreement in kinematic trajectories between MMC and OMC, as indicated by high correlations (median r above 0.95 for the majority of kinematic trajectories) and median RMSE values ranging from 2-5 degrees for joint angles, 0.04 m/s for end-effector velocity, and 6 mm for trunk displacement. Trial-to-trial biases between OMC and MMC were consistent within participant sessions, with interquartile ranges of bias around 1-3 degrees for joint angles, 0.01 m/s in end-effector velocity, and approximately 3mm for trunk displacement. Our findings indicate that our MMC for arm tracking is approaching the accuracy of marker-based methods, supporting its potential for use in clinical settings. MMC could provide valuable insights into movement rehabilitation in stroke patients, potentially enhancing the effectiveness of rehabilitation strategies.