Smartphones and Wearables as a Method for Understanding Symptom Mechanisms

Abstract

While psychological treatments have been shown to be effective in treating psychiatric disorders, the mechanism of their therapeutic effect is less well understood. An improved mechanistic understanding of psychiatric disorders and their treatments would enable refinement of existing interventions, and more targeted intervention and the development of new treatments. A major limitation in understanding the mechanism of effect in psychological treatments has been the challenge of capturing what happens outside of the clinical setting. The development of new digital technologies such as smartphones and wearables enables much more inter-session data to be collected. The rapid evolution of smartphones and wearable technologies, combined with the ubiquity of mobile networks means that is possible for patients to provide regular, longitudinal, and high-resolution data. This allows a previously inaccessible and untapped stream of a specific patient's behaviours, moods, activities, and thoughts to be quantified. Monitoring through such technologies may be of therapeutic value, improving self-awareness and promoting mentalization. Smartphones and wearable technologies can also be used to delivered therapies remotely. Digital technologies enable new insights to be gained into the lived experience of mental disorder enabling current treatments to be refined and personalised, as well as generating new targets for future treatment development. In this article we discuss how such technologies are improving our understanding of psychiatric disorder, informing psychological treatments before considering the future potential of such technologies. We will also consider the challenges and ethical concerns of such approaches.

AI in patient flow: applications of artificial intelligence to improve patient flow in NHS acute mental health inpatient units

“…Smartphone-based data such as app usage, spatial trajectories (GPS), physical mobility patterns (accelerometer), and audio samples (microphone) were collected to develop ‘precise and temporally dynamic disease phenotypes and markers to diagnose and treat psychiatric and other illnesses’ [ 107 ]. Griffin & Saunders [ 108 ] investigated the use of digital phenotyping in psychiatric treatments, using naturalistic data from smartphone and wearable devices, using physiological metrics such as heart rate, respiratory rate, and sleeping patterns in conjunction with behavioural data. AI analysis of such an array of longitudinal input data could enable personal relapse risk modelling and subsequently, the ability to target early interventions in the community to most effectively prevent avoidable admissions, with an additional offering of self-monitoring functionality for the patient [ 108 ].…”

“…Griffin & Saunders [ 108 ] investigated the use of digital phenotyping in psychiatric treatments, using naturalistic data from smartphone and wearable devices, using physiological metrics such as heart rate, respiratory rate, and sleeping patterns in conjunction with behavioural data. AI analysis of such an array of longitudinal input data could enable personal relapse risk modelling and subsequently, the ability to target early interventions in the community to most effectively prevent avoidable admissions, with an additional offering of self-monitoring functionality for the patient [ 108 ]. Integrating multiple sources of data presents a processing challenge, and such vast quantities of data require ML to analyse, correlate, and create baselines for different illnesses [ 109 ].…”

Digital Phenotyping in Livestock Farming

“…High standards have been implemented to increase the validity of questionnaire-based scoring systems in human trials, but animals are unable to answer such questions, and their caretakers are unlikely to have the time to fill out these papers on larger farms [23]. Algorithms can be created for the systematic scoring of farm animals based on digitally collected phenotypic traits to create a more modern, efficient, and accurate scoring system [24].…”

Digital Phenotyping in Livestock Farming

“…High standards have been implemented to increase the validity of questionnairebased scoring systems in human trials, but animals are unable to answer such questions, and their caretakers are unlikely to have the time to fill out these papers on larger farms [20]. Algorithms can be created for the systematic scoring of farm animals based on digitally collected phenotypic traits to create a more modern, efficient, and accurate scoring system [21].…”