To evaluate the current scope and future potential of AI-based interventions in neurocognitive rehabilitation and elucidate the ethical challenges that need to be addressed to ensure their effective incorporation in patient care.

The significant healthcare burden of neurological conditions like multiple sclerosis (MS), stroke, traumatic brain injury (TBI), and dementia highlight the critical need for effective neurocognitive rehabilitation interventions. Rapid advancements in artificial intelligence (AI) promote novel solutions in this domain by providing personalized and adaptable cognitive deficit-specific interventions compared to their traditional non-specific counterparts.

A contemporary review of the literature following PRISMA guidelines was conducted in September 2024 to identify articles published between 2014-2024 that examine the utility of different AI-based interventions such as robotic-assisted therapies, virtual reality (VR), and machine learning (ML) models in cognitive rehabilitation. From an initial screening of 267 papers, 29 studies that met the inclusion criteria were analyzed for key insights, including improvements in various cognitive domains and the technical challenges associated with these emerging technologies.

The literature suggests that AI-enhanced rehabilitation shows promising improvements across cognitive domains, especially in telehealth settings. VR-driven interventions improve cognitive flexibility and behavioral outcomes by offering real-time feedback and optimizing dynamic rehabilitation, while ML-driven approaches can assist in predicting treatment efficacy. Robot-assisted therapies demonstrated significant improvements in both cognitive and motor domains. However, these improvements were inconsistent across studies due to a lack of standardization, smaller sample sizes, and shorter intervention durations.

AI-based interventions seemingly offer effective personalized neurocognitive rehabilitation solutions that can improve patient outcomes. Challenges around limited standardization, generalizability, and integration in clinical settings, as well as ethical implications, including patient privacy and data security, equitable access, and variability in results across diverse population groups, necessitate further investigation before large-scale implementation.