Social factors and motivation are key factors for recovery in stroke patients (Glass, Matchar, Belyea, & Feussner, 1993). The goal of this study is to enhance accessibility and evaluate the effects of including social interaction in a virtual reality (VR) -based system for stroke rehabilitation. We hypothesize that a multiplayer competitive context will have a positive effect on the involvement of the patients in the therapy and thus on the rehabilitation process. We test this hypothesis using the Rehabilitation Gaming System (RGS), an ICT virtual reality tool for upper extremities motor rehabilitation. First, we implemented and evaluated a new interface based on a low-cost key-glove. Then, we developed a dedicated RGS scenario where the player has to match pairs of cards from a stack of playing cards. This task trains cognitive (memory) and motor tasks (grasping and reaching). Eight stroke patients participated in two sessions lasting 20 min, one using a single-player VR environment and another using a multiplayer version of the same game. A usability test showed that participants interact with the system much faster when using the new key-glove–based interface ( p = .02) in comparison to a mouse and keyboard. In addition, our results showed that upper limb exercises performed by the patients in multiplayer mode reached wider elbow flexion/extension movements than the ones performed during the single-player game session ( p = .04). Considering that the presence of spasticity is very common in patients affected by an ictus and that it causes an ongoing level of contraction, these results suggest that the patients affected displayed more effort in reaching if engaged in a social task. Our study shows that accessibility and social engagement in multiplayer environments positively affects the patients' performance and enjoyment during the task. Although the long-term impact of this enhanced motivation needs to be further assessed, our results do suggest that the inclusion of social factors such as multiplayer capabilities is an important factor for the rehabilitation process in VR-based therapy and might have an impact on both performance and mood of stroke patients.

Virtual and mixed reality environments (VMRE) often imply full-body human-computer interaction scenarios. We used a public multimodal mixed reality installation, the Synthetic Oracle, and a between-groups design to study the effects of implicit (e.g., passively walking) or explicit (e.g., pointing) interaction modes on the users' emotional and engagement experiences, and we assessed it using questionnaires. Additionally, real-time arm motion data was used to categorize the user behavior and to provide interaction possibilities for the explicit interaction group. The results show that the online behavior classification corresponded well to the users' interaction mode. In addition, contrary to the explicit interaction, the engagement ratings from implicit users were positively correlated with a valence but were uncorrelated with arousal ratings. Interestingly, arousal levels were correlated with different behaviors displayed by the visitors depending on the interaction mode. Hence, this study confirms that the activity level and behavior of users modulates their experience, and that in turn, the interaction mode modulates their behavior. Thus, these results show the importance of the selected interaction mode when designing users' experiences in VMRE.