NeuroAIreh@b is an AI-based methodology for personalized and adaptive neurorehabilitation. It integrates neuropsychological assessment (NPA) with computational modeling to create a cognitive profile (ACP) that guides personalized training. The methodology uses virtual reality (VR) simulations of daily living activities to enhance ecological validity and efficacy. A clinical study with stroke patients using a tablet-based intervention demonstrated the feasibility of NeuroAIreh@b. The system uses AI to optimize neurorehabilitation prescription by dynamically adjusting training tasks based on patient performance. It addresses challenges in traditional rehabilitation, such as lack of adaptability, resource intensity, and limited engagement. NeuroAIreh@b combines expertise from neuropsychology, computer science, game design, and AI for health to create a multidisciplinary approach. The system includes a framework for patient profiling, training selection, session adaptation, and system calibration. It uses ML algorithms to aggregate NPA data, define cognitive domains, and adjust task difficulty. The methodology aims to improve cognitive function in patients with acquired brain injury, dementia, and stroke by providing adaptive, personalized training. The system has potential for future large-scale randomized controlled trials to validate its efficacy. NeuroAIreh@b addresses limitations of existing AI-driven platforms, such as limited transfer effect, reduced engagement, and lack of clinical supervision. It uses a dynamic profile framework based on belief revision theory to adapt training tasks and ensure optimal patient outcomes. The system is designed to be flexible, scalable, and user-friendly, with a focus on ecological validity and clinical relevance.NeuroAIreh@b is an AI-based methodology for personalized and adaptive neurorehabilitation. It integrates neuropsychological assessment (NPA) with computational modeling to create a cognitive profile (ACP) that guides personalized training. The methodology uses virtual reality (VR) simulations of daily living activities to enhance ecological validity and efficacy. A clinical study with stroke patients using a tablet-based intervention demonstrated the feasibility of NeuroAIreh@b. The system uses AI to optimize neurorehabilitation prescription by dynamically adjusting training tasks based on patient performance. It addresses challenges in traditional rehabilitation, such as lack of adaptability, resource intensity, and limited engagement. NeuroAIreh@b combines expertise from neuropsychology, computer science, game design, and AI for health to create a multidisciplinary approach. The system includes a framework for patient profiling, training selection, session adaptation, and system calibration. It uses ML algorithms to aggregate NPA data, define cognitive domains, and adjust task difficulty. The methodology aims to improve cognitive function in patients with acquired brain injury, dementia, and stroke by providing adaptive, personalized training. The system has potential for future large-scale randomized controlled trials to validate its efficacy. NeuroAIreh@b addresses limitations of existing AI-driven platforms, such as limited transfer effect, reduced engagement, and lack of clinical supervision. It uses a dynamic profile framework based on belief revision theory to adapt training tasks and ensure optimal patient outcomes. The system is designed to be flexible, scalable, and user-friendly, with a focus on ecological validity and clinical relevance.