Adaptive Multimodal AI for Early Detection of Cognitive Decline Using Real-Time Health Data and Wearables in Aged Care

Abstract

The increasing number of elderly individuals worldwide makes early cognitive decline detection vital for handling conditions such as Alzheimer's disease. Traditional diagnostic approaches, which rely on neuropsychological testing in clinical settings, often lack scalability and accessibility and are unsuitable for continuous monitoring. The research examines how adaptive multimodal Artificial Intelligence (AI) systems, combined with wearable health technologies, can facilitate the immediate detection of cognitive impairment in aged care environments. Wearable devices, such as accelerometers and heart rate monitors, produce detailed physiological and behavioral data streams that combine with electronic health records and self-reports to generate comprehensive multimodal datasets. Machine learning algorithms, combined with deep learning frameworks such as recurrent and convolutional neural networks, analyze these datasets to identify patterns that indicate mild cognitive impairment. The predictive systems identify early gait irregularities and changes in heart rate variability and sleep patterns, enabling proactive management of dementia.