A Care Home Deployment of a Dynamic Lighting and Sensing System to Support Sleep Cycles for Dementia

One therapeutic approach to supporting well-being in dementia is through the use of lighting optimised for the circadian rhythm, otherwise known as daylight-simulating lighting. However, existing studies are commonly short-term (sub 12 months) and the datasets which report metrics of the lighting exposure and spectra are not systematically reported. Consequently, the relationship between lighting and dementia is not well understood. As dementia by nature is heterogeneous (and by extension as are their circadian rhythms), the need to tailor lighting based on an individual's demands is paramount. This paper presents a novel IoT architecture which facilitates the data-driven actuation of daylight-simulating lighting based on the observations of an individual's circadian-related sleep metrics. A trial of the technology in a care home is presented with environmentally-deployed radar data collected from 10 participants over 4 months. The preliminary results found that the median total time in bed at night increased by 45 minutes (±4.87 minutes) and 27 minutes (±3.24 minutes) from baseline to weeks 5-10 and weeks 10-16 respectively after dynamic lighting exposure. Additionally, the disturbance counts at night decreased by 2.70 from a group perspective, but exhibited rather extreme heterogeneity upon individual analyses. This highlights the need to tailor the timing of dynamic lighting on an individual basis as informed by sleep-wake data collected over time. Future work aims to implement the tailored actuation; providing optimum dynamic lighting exposure for supporting both circadian rhythms and wellbeing at a human-centred level.