Wearable multimodal sensors for the detection of behavioral and psychological symptoms of dementia using personalized machine learning models

TLDR

Behavioral and psychological symptoms of dementia (BPSD) indicate distress or unmet needs and pose risks to patients and caregivers, yet their variable expression hampers the development of reliable digital biomarkers. This study aimed to create personalized machine learning models that detect individual BPSD patterns using wearable multimodal sensors. Wristbands recorded motion, blood volume pulse, electrodermal activity, and skin temperature while staff annotated rare agitation or aggression events, and 1‑minute interval models classified symptom presence and type. The personalized models achieved a median AUC of 0.87 (range 0.64–0.95), with sensor feature importance varying by individual and behavior type, underscoring the need for personalization in digital phenotyping of BPSD.

Abstract

Behavioral and psychological symptoms of dementia (BPSD) signal distress or unmet needs and present a risk to people with dementia and their caregivers. Variability in the expression of these symptoms is a barrier to the performance of digital biomarkers. The aim of this study was to use wearable multimodal sensors to develop personalized machine learning models capable of detecting individual patterns of BPSD.Older adults with dementia and BPSD (n = 17) on a dementia care unit wore a wristband during waking hours for up to 8 weeks. The wristband captured motion (accelerometer) and physiological indicators (blood volume pulse, electrodermal activity, and skin temperature). Agitation or aggression events were tracked, and research staff reviewed videos to precisely annotate the sensor data. Personalized machine learning models were developed using 1-minute intervals and classifying the presence of behavioral symptoms, and behavioral symptoms by type (motor agitation, verbal aggression, or physical aggression).Behavioral events were rare, representing 3.4% of the total data. Personalized models classified behavioral symptoms with a median area under the receiver operating curve (AUC) of 0.87 (range 0.64-0.95). The relative importance of the different sensor features to the predictive models varied both by individual and behavior type.Patterns of sensor data associated with BPSD are highly individualized, and future studies of the digital phenotyping of these behaviors would benefit from personalization.