The goal of neurorehabilitation, a sophisticated medical procedure, is to facilitate the recovery from a nervous system damage and to compensate for any functional alterations occurring from stroke, meningitis, brain or spinal trauma, or even from hereditary illnesses. Current neurorehabilitation models mostly rely on prolonged hospital visits and regular therapy sessions necessitating frequent physical interactions between rehab specialists and patients. Assistive devices were developed to maintain or enhance an individual's functionality and independence while undergoing rehabilitation, hence enhancing their well-being and supporting specialist's procedures. Modern rehabilitation equipment is adaptable, portable, and integrated with clinical intelligence, coaching, and remote patient monitoring features. The development of computational learning methods and the creation of systems that can learn things automatically through machine learning (ML) make assistive technology smarter. ML enhances assistive technology biomonitoring systems, human–machine interfaces, physical support, and other assistive equipment ultimately aiding neurorehabilitation. This is akin to using ML techniques to classify human posture and movement to evaluate how well the intervention is responding to the movement quality. Without configuring software applications periodically for them to learn, ML generates judgments based on collected experiences via the successful resolution of prior challenges, leading them to become more intelligent and predictive. The chapter provides details on a range of ML computational techniques such as the spiking neural network, common spatial pattern, support vector machine, and others, used in neuro rehab assistive technologies, such as hearing aids, gesture assistants, spectacles, wheelchairs, communication aids, text-to-speech systems, and prostheses, for people who have difficulty speaking, typing, writing, remembering, pointing, seeing, hearing, learning, and walking.
Elsevier, Computational Intelligence and Deep Learning Methods for Neuro-rehabilitation Applications, 2024, Pages 121-148