TD‐P‐017: High‐Frequency Monitoring of Cognition, Mood and Behaviour Using Commercially Available Wearable Devices

Clinical and research assessments of cognition, mood and behaviour provide validated snapshots of functioning at specific time points. However, depression and mild cognitive impairment exhibit fluctuations which are challenging to quantify, but impact on quality of life and ability to perform daily tasks. The identification of these fluctuations has the potential to provide ecologically relevant outcomes for clinical research, complementing infrequent laboratory or in clinic measures. Sensor technology, embedded into wearable and mobile devices, is increasingly ubiquitous and sophisticated, and is capable of generating large volumes of multidimensional data. There is a growing evidence base linking these measurements to meaningful changes in symptoms and functional status. There are three major challenges in translating these data into clinically actionable information. The first is a problem of data reduction, combining and extracting features to produce meaningful metrics. The second is the validation of these metrics against existing tools. Third, and perhaps most fundamentally, the design of the device and software needs to be tailored to the user group in question to ensure tolerability, acceptability and compliance, and therefore generate meaningful data. We describe a novel method to address these three challenges. Data were collected from mobile and wearable devices, providing high-frequency measurement of autonomic variables and accelerometer data, and episodic sampling of behaviour and self-rated mood in healthy adults. Low-frequency, laboratory measurements of cognition, depression and anxiety were also collected for validation purposes, as were assessments of user experience. There was good compliance and acceptability of data collection, and consistency between average measures of cognition from episodic measurement and low-frequency laboratory measurement. Distinct temporal variability was found in cognitive assessment, physiological measurement and mood, indicating that this may be a suitable approach for tracking cognitive performance and mood over time. Our initial research has shown meaningful assessment of cognitive performance is possible using wearable devices and mobile phone technology, and that these data can be linked to monitoring of physiological and movement parameters, providing a rich picture of daily fluctuations in state. These data have implications for detection and monitoring of subtle changes due to a variety of conditions or interventions.