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

Alex Street^1*Paul William George Fernie^1*Jörg Christfried Fachner^1*Patrizia Di Campli San Vito²Nicolas Farina³Ming-Hung Hsu¹Leonardo Muller¹Stephen Brewster²Sube Banerjee⁴Alex Kirke³Hari Shaji³Paulo Itaborai³Eduardo Reck Miranda³

• ¹Anglia Ruskin University, Cambridge, United Kingdom
• ²University of Glasgow, Glasgow, United Kingdom
• ³University of Plymouth, Plymouth, United Kingdom
• ⁴University of Nottingham, Nottingham, United Kingdom

Abstract (323) Background: One of the reasons for early care home admissions in dementia is escalating neuropsychiatric symptoms (NPS) and music listening can help regulate these. Aims: RadioMe was a project for people living at home with dementia with the objective of building a system to help them maintain best 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 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, music compilation process and responses of participants when listening, collection of daily agitation HR and behavioural data, and 2. preliminary testing of HR activated music. Methods: In stage 1: a playlist compilation procedure 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 delivered 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: Systems were installed to test the automated playlist activation and informal feedback gathered on system function/user experience. Findings: The music compilation procedure enabled bespoke playlists. Sessional HR and video data had limited utility in supporting suitability of music for NPS management. The methodology for participants to collect agitation data failed and the algorithm was not refined. Researchers compiled playlists with 25 people living with dementia, mean age 73.8 (n=12 male, 13 female). 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 became repetitive. The system needs extensive refinement to simplify operation. Activation of the music needs to be better calibrated. A feasible, 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, which participants reported to be more in-line with their symptoms.