Investigator:
Erik Page, Blue Iris.

MassAITC Cohort: Year 2 (AD/ADRD)

Project Accomplishments: The Blue Iris Labs pilot study investigated how personal light exposure affects sleep quality using a new wearable light sensor called the Speck. Conducted from mid-2023 to early 2025, the study involved 31 participants, 27 of whom activated the device and 15 who provided high-quality paired light and sleep data. The Speck collected detailed light exposure data, while sleep metrics were gathered via Apple Watch. The study confirmed that bright light in the morning and low light at night were associated with better sleep, aligning with existing research. Despite some technical and compliance challenges, the Speck performed well in real-world conditions, and participants generally adhered to the protocol.

The study also explored individual differences in light-sleep relationships and began developing models to explain these variations using demographic and behavioral data. The findings support the feasibility of using the Speck in long-term, real-world studies and suggest it could be a valuable tool for monitoring and improving sleep health, especially in aging populations. The pilot laid the groundwork for future research, including studies with Alzheimer’s and dementia populations, and highlighted the need for standardized sleep tracking tools and more diverse participant recruitment. The success of this study is expected to accelerate the commercial launch of the Speck and its companion app, Speck Connect. 

Initial Proposal Abstract: Light is the primary regulator of human circadian rhythms. Through our prior NIH-funded research, we have developed and commercialized the first wearable light exposure devices that gather all physical properties of light thought to impact circadian health. Sleep quality is a critical measure of circadian health. Sleep trackers are in mass commercial use and are improving in accuracy and precision. Here, we seek to identify correlations between what are arguably the most important environmental inputs (light exposure) and health outcomes (sleep) in circadian science.

We propose gathering data using commercially available devices (i.e., our light exposure tracker and leading commercial sleep trackers) and machine learning techniques to identify the light exposure parameters and patterns that correlate most strongly with measured sleep metrics. We propose to collect light exposure and sleep metric data from at least 15 (non-AD/ADRD) individuals for at least 50 days per individual using methods that could scale commercially.

This pilot test is primarily intended to validate the methodology related to correlating light exposure and sleep metric data rather than to directly estimating the nature or magnitude of these correlations. These data will be analyzed for correlations between light exposure inputs and sleep quality outputs. Specifically, we propose to identify how the intensity, spectrum, and timing of light exposure of individuals correlate with their sleep duration, sleep phases, and sleep scores, as reported by various commercial sleep trackers.

We further propose to investigate how these correlations vary between individuals, seeking to identify potential parameters that influence interpersonal variations, such as demographic differences and differences in personal light sensitivity. If the methods proposed in this pilot are validated, the study results are expected to inform the design of a larger clinical trial that includes participants who have Alzheimer’s disease (AD) or related dementia (ADRD), with the aim of identifying specific light exposure patterns associated with specific, quantifiable improvements in sleep metrics in these populations.