Abstract
Although ElectroEncephaloGram (EEG) signals allow subjects suffering from neuromuscular disorders to interface their brains with the cyber-physical world, occupational therapy can be enhanced with the introduction of further modalities better assist the disabled person. In this paper, we propose an in-home occupational therapy environment, which leverages a rich set of occupational therapy-related activity recognition modalities, namely, EEG signals to understand brain activity, ElectroMyoGram (EMG) signals for muscle activity, gesture-tracking sensors for forward and inverse kinematics activities, and smart home appliance control sensors. To support a wide variety of disabled people's in-home occupational therapy, we have incorporated both selective attention and motor imagery processes for mapping a mental command with that of an occupational therapy-related command within a serious game environment. To attain higher accuracy and to avoid a higher number of false positives, a subject is first recommended to use a selective attention-based serious game in which a digital avatar of the subject acting as a model therapist will guide the therapy session. Once familiar with the generation of proper motor imagery, an advanced user can use self-paced motor imagery signals to perform occupational therapy activities within the serious game environment. The occupational therapy consists of a serious game environment in which smart home appliances are mapped with therapeutic activities through forward and inverse kinematics. The therapy data has been secured through blockchain and off-chain-based distributed repositories. The test results show the viability of using the framework in a clinical environment.