Edge-Ready Explainable AI for Early Detection of Neurocognitive Decline

The timely identification of neurocognitive disorders such as Alzheimer's and dementia is very critical for proper treatment and management. The current approaches involve regular clinical assessments and central computing that make continuous monitoring impossible and create risks in terms of privacy issues. The XAI approach proposed within this chapter is aimed at detecting neurocognitive impairments in time, with edge computing being the execution of computations and inference on the local device close to the source of data. This solution utilizes advanced ML and DL algorithms that allow real-time processing at the same time, having low computational power. Explainable methods are used to help understand the key determinants for the cognitive disorders. Performance evaluation by common metrics shows the efficiency of the suggested approach.