AHA: Augmented Human Assistance

Funder

Organizational Unit

Publications

Interactive hybrid approach to combine machine and human intelligence for personalized rehabilitation assessment

Publication . Lee, Min Hun; Siewiorek, Daniel P.; Smailagic, Asim; Bernardino, Alexandre; Bermúdez i Badia, Sergi

Automated assessment of rehabilitation exercises using machine learning has a potential to improve current rehabilitation practices. However, it is challenging to completely replicate therapist’s deci sion making on the assessment of patients with various physical conditions. This paper describes an interactive machine learning approach that iteratively integrates a data-driven model with ex pert’s knowledge to assess the quality of rehabilitation exercises. Among a large set of kinematic features of the exercise motions, our approach identifies the most salient features for assessment using reinforcement learning and generates a user-specific analysis to elicit feature relevance from a therapist for personalized rehabilita tion assessment. While accommodating therapist’s feedback on fea ture relevance, our approach can tune a generic assessment model into a personalized model. Specifically, our approach improves performance to predict assessment from 0.8279 to 0.9116 average F1-scores of three upper-limb rehabilitation exercises (𝑝 < 0.01). Our work demonstrates that machine learning models with feature selection can generate kinematic feature-based analysis as expla nations on predictions of a model to elicit expert’s knowledge of assessment, and how machine learning models can augment with expert’s knowledge for personalized rehabilitation assessment.

2020Conference object

Open access

Coaching or gaming? Implications of strategy choice for home based stroke rehabilitation

Publication . Cameirão, Mónica S.; Smailagic, Asim; Miao, Guangyao; Siewiorek, Dan P.

Background: The enduring aging of the world population and prospective increase of age-related chronic diseases urge the implementation of new models for healthcare delivery. One strategy relies on ICT (Information and Communications Technology) home-based solutions allowing clients to pursue their treatments without institutionalization. Stroke survivors are a particular population that could strongly benefit from such solutions, but is not yet clear what the best approach is for bringing forth an adequate and sustainable usage of home-based rehabilitation systems. Here we explore two possible approaches: coaching and gaming. Methods: We performed trials with 20 healthy participants and 5 chronic stroke survivors to study and compare execution of an elbow flexion and extension task when performed within a coaching mode that provides encouragement or within a gaming mode. For each mode we analyzed compliance, arm movement kinematics and task scores. In addition, we assessed the usability and acceptance of the proposed modes through a customized self-report questionnaire. Results: In the healthy participants sample, 13/20 preferred the gaming mode and rated it as being significantly more fun (p < .05), but the feedback delivered by the coaching mode was subjectively perceived as being more useful (p < .01). In addition, the activity level (number of repetitions and total movement of the end effector) was significantly higher (p <.001) during coaching. However, the quality of movements was superior in gaming with a trend towards shorter movement duration (p=.074), significantly shorter travel distance (p <.001), higher movement efficiency (p <.001) and higher performance scores (p <.001). Stroke survivors also showed a trend towards higher activity levels in coaching, but with more movement quality during gaming. Finally, both training modes showed overall high acceptance. Conclusions: Gaming led to higher enjoyment and increased quality in movement execution in healthy participants. However, we observed that game mechanics strongly determined user behavior and limited activity levels. In contrast, coaching generated higher activity levels. Hence, the purpose of treatment and profile of end-users has to be considered when deciding on the most adequate approach for home based stroke rehabilitation.

2016Journal article

Open access

A virtual reality bus ride as an ecologically valid assessment of balance: a feasibility study

Publication . Gonçalves, A.; Montoya, M. F.; Llorens, R.; Bermúdez i Badia, S.

Balance disorders can have substantial adverse implications on the performance of daily activities and lead to an increased risk of falls, which often have severe negative consequences for older adults. Quantitative assessment through computer ized force plate-based posturography enables objective assessment of postural control but could not successfully represent specifc abilities required during daily activities. The use of virtual reality (VR) could improve the representative design of functional activities and increase the ecological validity of posturographic tests, which would enhance the transferability of results to the real world. In this work, we investigate the feasibility of a simulated bus ride experienced in a surround-screen VR system to assess balance with increased ecological validity. Participants were frst evaluated with a posturography test and then with the VR-based bus ride test, while the reactions of their centre of pressure were registered. Lastly, participants provided self-reported measures of the elicited sense of presence during the test. A total of 16 healthy young adults completed the study. Results showed that the simulation could elicit signifcant medial–lateral excursions of the centre of pressure in response to variations in the optical fow. Furthermore, these responses’ amplitude negatively correlated with the participants’ posturography excursions when fxating a target. Although the sense of presence was moderate, likely due to the passive nature of the test, the results support the feasibility of our proposed paradigm, based in the context of a meaningful daily living activity, in assessing balance control components.

2021Journal article

Open access

Closing the loop in exergaming - Health benefits of biocybernetic adaptation in senior adults

Publication . Muñoz, John Edison; Cameirão, M.; Bermúdez i Badia, S.; Gouveia, E. Rubio

Exergames help senior players to get physically active by promoting fun and enjoyment while exercising. However, most exergames are not designed to produce recommended levels of exercise that elicit adequate physical responses for optimal training in the aged population. In this project, we developed physiological computing technologies to overcome this issue by making real-time adaptations in a custom exergame based on recommendations for targeted heart rate (HR) levels. This biocybernetic adaptation was evaluated against conventional cardiorespiratory training in a group of active senior adults through a floor-projected exergame and a smartwatch to record HR data. Results showed that the physiologically-augmented exergame leads players to exert around 40% more time in the recommended HR levels, compared to the conventional training, avoiding over exercising and maintaining good enjoyment levels. Finally, we made available our biocybernetic adaptation software tool to enable the creation of physiological adaptive videogames, permitting the replication of our study.

2018Conference object

Open access

Measured and perceived physical responses in multidimensional fitness training through exergames in older adults

Publication . Munoz, J. E.; Gonçalves, A.; Gouveia, E. R.; Cameirão, M. S.; Bermúdez i Badia, S.

Exergames have been used to increase physical activity levels to produce health benefits in older adults. However, only a small number of studies have quantified the physical activity levels produced by custom-made Exergames and their capacity to elicit recommended levels of exercise. This study investigates the effectiveness of custom-made Exergames, designed for multidimensional fitness training, in eliciting recommended levels of exercise. We rely on both objective (accelerometry) and subjective (perceived exertion) information collected in two different modalities of exercise, consisting of 40- minutes sessions: Exergaming and conventional training (Control). A between-subjects analysis was done involving two groups of active older adults (n=33). Participants in the Control Between condition performed physical activity in conventional group fitness training, while the intervention group used individualized Exergaming as training modality. In addition, a sub-group of the Exergaming participants also performed a conventional training session (Control-Within), which enabled a within-subjects comparison. Results show that participants spent significantly more time in moderate-to-vigorous intensities during Exergaming, interestingly, perceiving significantly lower exertion levels. The between-subjects analysis only presented statistically significant differences for the perceived exertion scale. This study helps to unveil the impact of custom-made Exergames in physical activity levels during training when compared to conventional training for the older adult population.

2018Conference object

Open access