The SERENADE project: Sensor-Based Explainable Detection of Cognitive Decline

Early detection matters: Mild Cognitive Impairment affects 12-18% of people over 60, and while it may progress to dementia, current clinical methods for predicting this transition are unreliable. SERENADE aims to improve early detection by monitoring behavioral changes in daily activities rather than relying on traditional clinical indicators.

Transparency is essential for clinical adoption: The paper identifies a critical gap in existing AI systems—current machine learning models for detecting cognitive decline "lack transparency, hindering clinicians' trust." SERENADE addresses this by using explainable AI methods, recognizing that doctors need to understand why an AI system flags a patient as at-risk before they can confidently use it in practice.

Non-intrusive monitoring through everyday technology: Rather than requiring patients to visit clinics repeatedly, SERENADE uses sensors in smart homes and wearable devices to continuously monitor Activities of Daily Living (ADLs). This approach allows for long-term observation (one year of data collection) while minimizing disruption to patients' lives.

Real-world validation is underway: The project involves 30 MCI patients living alone in a year-long study, representing a genuine effort to test these methods in realistic conditions rather than controlled laboratory settings, which is crucial for determining whether explainable AI can actually support clinical decision-making in practice.