Affect detection in rehabilitation using wearable robotics, multiple biosensors and serious gaming: a concept using the NeuroSuitUp platform

Background: Neural rehabilitation and rehabilitation robotics explore their predominant focus on restoration of sensorimotor function, attempting to recruit and promote adaptive neural plasticity while inhibiting maladaptive plasticity. Although the role of emotions and affective states are considered crucial to the learning process, their implications on rehabilitation process have not been sufficiently studied.

Methods: The NeuroSuitUp platform, at full deployment, consists of a non-rigid wearable robotics modality (jacket and gloves), featuring autonomous on-board computing, and a live visualization avatar-based serious game created in Unity engine. Incorporated sensors include: a) inertial measurement units (IMUs) at joints of upper-body kinematic chain, b) finger bend and pressure sensors, c) electromyography (EMG) sensors at major muscle groups, d) electrocardiography (ECG) and heart rate (HR) monitoring, e) skin conductance sensor, f) wireless electroencephalography (EEG). Hardware also includes surface electrical muscle stimulation (EMS) devices for movement assistance and mixed reality (MR) presentation of the game. Hybrid control scheme will allow either biofeedback (EMG/IMUs) or neurofeedback (EEG/IMUs) as input to control the game and to modulate received assistance from EMS.

Conclusion: While motor control and assistance is the main focus of the platform aiming to promote dormant neuroplasticity in chronic neurological disability, biosignal sensors (EEG, EMG, ECG, HR) will be used to track affect states of users in real-time, assessing emotional impact, learning gain, as well as engagement and motivation. Rehabilitation experience could therefore be individualized not only according to disability status but also tailored real-time to user’s interest, attention and effort.