AI Tool Predicts Psychosis in Alzheimer's Patients

Researchers at Northwell Health’s Feinstein Institutes for Medical Research have developed an AI-based tool that identifies a brain metabolic network capable of predicting psychosis in Alzheimer's disease (AD), announced in a press release. This discovery, published in Brain Communications, offers a potential breakthrough for earlier diagnosis and treatment development for individuals with AD.

The study, led by Dr. Jeremy L. Koppel and Dr. An Vo, utilized an AI-derived metabolic brain map to analyze FDG-PET scans of people with AD. The AI identified a unique pattern of brain activity, termed the Alzheimer's Disease Psychosis Network (ADPN), which acts as a biomarker for predicting psychosis. This pattern was particularly effective in distinguishing brain activity differences between those with and without psychosis.

Psychosis in AD, characterized by hallucinations and delusions, is linked to accelerated cognitive decline and increased caregiver burden. The ADPN offers a promising tool for identifying individuals at risk, potentially allowing for earlier and more targeted interventions. The stronger the ADPN 'fingerprint,' the more severe the cognitive decline observed in patients.