Multi-scale markers of circadian rhythm changes for monitoring of mental health

Lead Research Organisation: University of Oxford
Department Name: Engineering Science


Almost one quarter of adults currently experience some form of mental health disorder in the UK, costing the healthcare system an estimated £77 billion each year. However, there exists very little objective or real-time monitoring of sufferers of mental health issues. This pilot project will investigate the development of a novel data fusion framework that will be suitable for combining many observations of a patient's behaviour to allow accurate mental health monitoring in any environment. Recent studies have shown that certain types of physical behaviour, daily cycles (circadian rhythms) and social networking activity can be indicative of an individual's state of mental health. However, recording the necessary data to make a diagnosis is difficult, both due to the nature of the health issues and because of the instrumentation needed. Recent developments in commercially available equipment (including smart phones) mean that we now have the opportunity to cheaply and routinely record human behaviour as well as daily patterns of physiology (such as sleep and cardiac activity). By then applying advanced pattern recognition and data fusion techniques, we intend to provide daily feedback of mental well-being to both the patient and care providers. This could facilitate early interventions in deteriorating individuals, thereby lowering costs of health care and reducing the severity of the illness. We also intend to begin to answer the more fundamental question about how circadian rhythms change as mental health deteriorates.
The developed of a user-friendly and user-controlled monitoring system, together with a suite of suitable algorithms, will be an important step towards a larger integration of the ever increasing multi-dimensional biometric data we are beginning to collect. This includes signals such as location, body temperature, speech patterns and social interaction behaviours. The potential to fuse data from many different sensors, and many different algorithm, will provide a platform for intelligible interpretation of the vast quantities of data that are beginning to confront researchers in biomedical applications. It will also help to improve the accuracy of monitoring systems and provide the doctor with more objective assessments of patient behaviour, which could lead to more accurate and timely diagnoses.

Planned Impact

To maximise the impact of this project we will work closely with our clinical collaborators (Prof. Guy Goodwin at the Warneford Hospital, Oxford, UK), who specialize in mental health disorders. We also have begun to collaborate with Prof Russell Foster and his Circadian and Visual Neuroscience group in the Department of Ophthalmology at the University of Oxford, to help understand the mechanistic pathways involved in changes in mental health. This grant will help to consolidate our collaboration with these groups and provide a dedicated full-time postdoctoral researcher to work closely with all three teams. The Centre for Doctoral Training in Healthcare Innovation, of which the PI is the Associate Director, will also provide a catchment pool from which to recruit high quality, fully funded students to work with our team.

We will engage the wider clinical community and the NHS about the impact of these technologies in improving health care provision and reducing costs in the NHS and how they can be widely integrated into clinical workflow. We are currently in discussion with Prof. Digby Quested and Prof. John Geddes (at the Department of Psychiatry in Oxford) concerning clinical trials and mood monitoring programmes such as 'True Colours'. The research detailed in this proposal will have direct applicability to such systems and could be integrated into conventional clinical pathways through our clinical partners at the Warneford Hospital.

The dissemination of the key research findings will be to leading international journals such as IEEE Transactions on Biomedical Engineering, Annals of Biomedical Engineering, IOP Physiological Measurement, and Journal of the Royal Society Interface, and to international conferences such as the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP) and IEEE Engineering in Medicine and Biology.

The research is likely to have direct commercial applicability on a broad level, ranging from close commercial partners to the sleep and health monitoring industry in general. There has been significant direct industrial interest in this research. Potential partners, who have already shared data and signed confidentiality agreements, include Proteus Biomed, who manufacture monitoring patches and edible digital tags for medication, which we intend to incorporate into our monitoring process. We will continue to work with them and explore the potential for commercialisation of the developed algorithms. The University of Oxford has a dedicated technology transfer company, ISIS Innovation, which has a long experience in the commercialisation of intellectual property including patenting and licensing of research findings, and the foundation of spin-out companies. We will involve ISIS Innovation in developing an appropriate IP protection strategy for the algorithms developed. The results of the studies may also have an impact on government policies to do with management of mental health, since the proposed technology will allow us to assess the response of an individual to a particular drug or therapy.

Finally, the proposed research complements, but could also augment the recently funded Wellcome Trust 'Sleep and Circadian Neuroscience Institute' at the University of Oxford. We expect that the proposed research will translate to the ongoing research activities in this Institute, and further strengthen the efforts. Our long term funding goal at the end of this 1-year grant is to extend the findings of this work, drawing on additional funding and resources from the BRC, Wellcome Trust and commercial partners.

We expect to be able to deliver a prototype into the NHS to begin trials by the end of the 1-year grant and begin commercial partnerships in the following year, with an expectation of scaling to a randomised clinical trial over the following three years.


10 25 50
Description This grant sought to discover whether it was possible to identify differences between patients with significant mental health conditions and those without, using simple wearable technology (such as a wrist worn movement/activity band, or a mobile phone). We found that we were able to differentiate between schizophrenics and control patients with a 85% accuracy using movement, and a 100% accuracy using movement and a wearable heart rate sensor. This is the first time that mental health has been quantified in such a way and opens up a new path in digital mental health. We are continuing the project with funding from the Wellcome Trust and are expanding the analysis to include social networking behavior and other physiology.

New collaborations with LifeQ and Medibio Inc have been established, which will enable us to translate the research into clinical practice much more rapidly, including the licensing of one patent.

We also have a significant publications:
Tsanas A1, Saunders KE, Bilderbeck AC, Palmius N, Osipov M, Clifford GD, Goodwin G?, De Vos M.Daily longitudinal self-monitoring of mood variability in bipolar disorder and borderline personality disorder. J Affect Disord. 2016 Nov 15;205:225-233. doi: 10.1016/j.jad.2016.06.065. Epub 2016 Jul 2.
E Reinertsen, M Osipov, C Liu, JM Kane, G Petrides, GD Clifford, Continuous assessment of schizophrenia using heart rate and accelerometer data, Physiological Measurement, 2017 (Awarded IOP Select Distinction);
M Osipov, Y Behzadi, JM Kane, G Petrides, GD Clifford, Objective identification and analysis of physiological and behavioral signs of schizophrenia, Journal of Mental Health 24 (5), 276-282, 2015
Osipov M., Towards automated symptoms assessment in mental health, DPhil thesis (Supervised by Clifford, GD), Department of Engineering Science, University of Oxford, 2016.
N Palmius, M Osipov, AC Bilderbeck, GM Goodwin, K Saunders, A Tsanas, GD Clifford, A multi-sensor monitoring system for objective mental health management in resource constrained environments, pp4, 2014/1/1, IET Digital Library
A Roebuck, V Monasterio, E Gederi, M Osipov, J Behar, A Malhotra, GD Clifford, A review of signals used in sleep analysis, Physiological measurement 35 (1), R1, 2013
M Osipov, K Wulff, RG Foster, GD Clifford, Multiscale Entropy of physical activity as an objective measure of activity disorganization in a context of schizophrenia, Informática Salud 2013
O Carr, M de Vos, KEA Saunders, Heart rate variability in bipolar disorder and borderline personality disorder: a clinical review, Evidence-based mental health 21 (1), 23-30 2018 .
O Carr, KEA Saunders, A Tsanas, AC Bilderbeck, N Palmius, JR Geddes, Variability in phase and amplitude of diurnal rhythms is related to variation of mood in bipolar and borderline personality disorder, Scientific Reports 8 (1), 1649 2018
N Palmius, A Tsanas, KEA Saunders, AC Bilderbeck, JR Geddes, Detecting bipolar depression from geographic location data, IEEE Transactions on Biomedical Engineering 64 (8), 1761-1771, 2017
Exploitation Route We expect the rich data to provide many more results over the next few years and are creating the world's richest data set on mental health, which is being adopted as a standard in Oxford's Psychiatric Research and hopefully in clinical practice.
In the medium term we expect companies like Proteus Digital Health to provide monitoring of psychiatric patients and our algorithms as severity of illness assessments to identify rapid deterioration or inappropriate medication / low compliance.

During the project we developed a substantial suite of open source software to run on smart phones, and have made that publicly available either directly on Google Play (AMoSS) or via google code. We expect other researchers to freely use this software to extend our results and apply it to other patient health issues.
Sectors Digital/Communication/Information Technologies (including Software),Education,Healthcare,Pharmaceuticals and Medical Biotechnology

Description The findings have now been used to provide additional information for an ongoing and scaled study at the department of Psychiatry in Oxford 'the AMoSS study' and are informing how and what instrumentation patients will toleration. Psychiatric patients have reported an increased feeling of wellbeing and satisfaction, as well as a deeper understanding of their own condition, and how they can try to manage it through behavioural changes. Collaboration with industry has led to a long term partnership (with Proteus Biomed), with potential impacts on the UK economy. Software from the project has been posted open source and other groups around the country have started to investigate the usage of the software. This saves time and money on other research projects, and potentially leads to quicker gains and faster impacts on patient populations.
First Year Of Impact 2014
Sector Digital/Communication/Information Technologies (including Software),Healthcare,Pharmaceuticals and Medical Biotechnology
Impact Types Cultural,Societal