Deep reinforced cognitive analytics algorithm (DRCAM): An advanced method to early detection of cognitive skill impairment using deep learning and reinforcement learning

A Deep Reinforced Cognitive Analytics Model (DRCAM) has been proposed in this work, integrating multimodal learning and reinforcement-based interventions for enhanced cognitive impairment diagnosis and management. This study proposes a novel approach combing Multimodal Transformers (MMT) for fusion features, namely, neuroimaging data, wearable sensors, neuropsychological test scores, and text pro-appraisals. A CNN-LSTM hybrid model is used for mapping spatial and temporal dependencies, and, on the other hand, a Deep Q-Network (DQN) improves while instructing how to perform proper cognitive training. Long-term cognitive state predictions are made by a Temporal Convolution Network (TCN). The MMT model achieves a classification accuracy of 90-92 %. Improvement in accuracy and intervention with discussable efficacy and potential for explanation is seen when benchmarked against conventional cognition.•Proposed the Deep Reinforced Cognitive Analytics Algorithm (DRCAM) for multimodal data.•The proposed model outperforms traditional models in cognitive skill impairment detection.•Demonstrated scalability for diverse healthcare datasets.