# RadioMe: an automated home-based radio, music playlist, and diary reminder system: Report on recruitment, music compilation, and listening, and preliminary testing of heart rate activated music

## TL;DR

RadioMe is a system designed to help people with dementia at home by using music to manage symptoms, but early testing showed it needs significant improvements.

## Contribution

The paper introduces RadioMe, a novel home-based system that uses heart rate to trigger music for managing dementia symptoms.

## Key findings

- The music compilation process allowed for personalized playlists for 25 participants with dementia.
- Participants found the automated music activation system too complex and not sensitive enough.
- Data collection during mild neuropsychiatric symptoms is needed to refine the system's algorithm.

## Abstract

One of the leading reasons for early admission to a care home in dementia is the escalation of neuropsychiatric symptoms (NPS), and music listening can help regulate these symptoms.

RadioMe was a project designed for people living at home with dementia to build a system to help them maintain the highest quality of life there for as long as possible, with three functional components: 1. Streaming radio; 2. Providing pre-recorded spoken diary reminders; 3. interrupting the radio with pre-compiled playlists when a wrist-worn heart rate (HR) monitor detects stress. This article reports on the first two stages of this three-stage project: 1. recruitment, the music compilation process, responses of participants when listening, collection of daily agitation HR and behavioural data, and 2. preliminary testing of HR-activated music.

In stage 1, a playlist compilation process was co-designed with a lived experience group; HR and behavioural data were collected by participants when agitated to refine the algorithm used for automated music activation; 15 home visits were conducted to compile and test the playlists, collecting video, HR, and autobiographical data in each session to inform on playlist suitability for NPS management. Stage 2 involved installing systems to test automated playlist activation, and informal feedback was gathered on system functionality and user experience.

The music compilation process enabled the creation of bespoke playlists. Sessional HR and video data had limited utility in supporting the suitability of music for NPS management. The methodology for collecting agitation data from participants failed, and the algorithm was not refined. Researchers compiled playlists with 25 people living with dementia, with a mean age of 73.8 years (n = 12 men, 13 women). Ten participants had systems installed to test automated music activation. They found it too complex; system calibration was not sensitive enough, music played at random times, and it became repetitive. The system needs extensive refinement to simplify its operation. The activation of the music needs to be better calibrated. A feasible and effective method of gathering data from participants in their homes is required to refine the algorithm, which must include HR/biodata during milder NPS events, as participants reported these to be more in line with their symptoms.