EPSRC funded PhD position: A musical bio-feedback loop to facilitate sleep
Applications are invited for a funded PhD position that will investigate the use of music to facilitate sleep. The project is a collaboration between the Department of Music and Department of Automatic Controls and Systems Engineering of The University of Sheffield, and the Sheffield based company Braintrain2020. Findings from the PhD project will contribute to the development of SleepCogni, hardware and software to assist sleep. You can apply using the provided link. Shortlisted candidates will be asked to submit a full CV, letter of motivation, and project proposal before the interview date.
- Supervisors: Dr Renee Timmers, Dr George Panoutsos, Dr Maan van de Werken
- Application deadline: Thursday 27 June 2019
- Interview date: Thursday 11 July 2019
- Starting date: 1 October 2019
- Duration of funding: 4 years
- Contact: firstname.lastname@example.org
- Application submission: https://forms.gle/QU9mPejXjwxyS8jt8
Whilst music is often used as a self-medication to promote sleep, little is known about how music promotes sleep and how its effectiveness can be increased. In this project, we investigate how music can be employed to facilitate physiological and attentional changes required to make the transition from an awake state to sleep. This is done by first investigating adjustments of musical characteristics to match physiological and attentional characteristics. Subsequently, this matching process is used to set up a bio-feedback loop that provides cues and direction for the sleep induction process. These processes will first be trialled with healthy volunteers and secondly with a group of insomniacs.
Hosting department and unit: https://www.sheffield.ac.uk/music/research/psychology
Music comes in a great variety of shapes and forms, and indeed it may analogously serve a variety of purposes. Depending on such features as tempo, pitch range, spectral centroid, intensity, pulse salience and melodic and harmonic complexity, music may support activation and motivation for a high intensity sports event, or music may support relaxation and sleep induction. Processes central to the activating and relaxational nature of music engagement include bodily and neurophysiological entrainment, as well as positive enjoyment of the music, and a willingness to use music for a particular purpose.
Capitalising on the ability of listeners to entrain with music, and for music to subsequently influence physiology, attention, and relaxation of listeners, we propose the development of a system that sets up a musical biofeedback loop that is used to facilitate sleep induction: attentional and physiological measures are taken of participants to assess their state of arousal/sleepiness. Musical characteristics are adjusted to match the arousal state of the participant. Finally, a feedback loop is established as listeners receive feedback on their state of arousal, providing cues for stages towards sleep and positive reinforcement for relaxation, when participants learn to adjust physiology to accomplish a sonic result.
This project is realised in collaboration with researchers of SleepCogni and will use a handheld device developed by the company that measures skin temperature, reaction time, heart rate and skin conductance. It builds on a pilot study that showed promising results for the use of music for sleep with gradually changing characteristics when compared to using no music or music with constant characteristics. Music specially composed for the purpose of the pilot was used that had properties characteristic of music used for sleep induction. This set of characteristics was defined through an analysis of sleep music play lists. The pilot was conducted with healthy volunteers and without any reference to normal duration of sleep induction of participants. The central thesis of SleepCogni is that lighting and music/sound can inform the participant about his or her wakeful state, and provide reinforcing feedback on the transition from awake to asleep. The proposed project will offer a central contribution to this thesis by accomplishing the following innovations:
- Experiment 1: Modelling and testing of relationship between musical parameters and sleepy-awake judgments and relaxation judgments of participants
- Experiment 2: Modelling and testing of physiology and attentional states of participants with respect to awake and asleep states (in collaboration with SleepCogni partners)
- Experiment 3: Modelling and testing relationships between musical parameters and induced physiological-attentional states.
- Experiment 4: Modelling and testing a musical biofeedback loop to facilitate sleep onset and quality (in collaboration with SleepCogni partners)
This project sets up a human-computer interface based on physiological and attentional measures. It will use music synthesis tools for music production and machine learning to model relationships between music and physiological and attentional states. Fellow researchers from the SleepCogni project will lead Experiment 2, and will assist the researcher in programming and evaluating the success of a feedback loop in Experiment 4.
Training and development plan
The researcher will be supervised by three supervisors with different disciplinary expertise: Dr Renee Timmers, expert in psychology of music, including experimental psychological quantitative research methods, physiological and emotional responses to music, and modelling of musical features and music perception. Dr George Panoutsos, expert in computational intelligence and data mining, and collaborator with Sleep Cogni. And Dr Maan van de Werken, expert in chronobiology and the biology of sleep, lead researcher for SleepCogni. This team will facilitate access to training within the respective departments: Music technology and music psychological research methods at the Department of Music, machine learning methods in the Department of Automatic Control and Systems Engineers (ACSE), and measurements of sleep and physiological changes related to sleep at Sleep Cogni, Braintrain2020. Aligned projects in the Department of ACSE will ensure that the student receives the necessary training for using existing Machine Learning (ML) tools to develop a modelling framework, without the need to be an expert in ML. The student will also have access to ACSE's training programme, including introductory, as well as advanced (if needed) taught modules and seminars in data-driven modelling and machine learning.
A phasing of training and development is planned that will allow the project to become increasingly sophisticated, whilst allowing the PG researcher to collect meaningful data from the first year onwards: Year 1 (Experiment 1): experimental psychological testing, physiological data measurement, musical synthesis (expertise available in Music). Year 2 (Experiment 2): computational modelling techniques for physiological and attentional data and relationship to sleep (expertise obtained in placement at SleepCogni and classes in ACSE). Year 2 (Experiment 3): similar expertise needed as for Experiment 1. Year 3 (Experiment 4): programming for real-time feedback, measures of sleep onset and quality. Training at SleepCogni and ACSE.
Training and development will further be assured through regular participation in reading groups and lab-meetings as the researcher will be part of the research centre Music, Mind, Machine, and participation and presentation at conferences, a summer school and relevant workshops.
Applicants are expected to have a good undergraduate and postgraduate degree in an area relevant for the PhD position, ideally combining music, psychology and computer science. Candidates may have an educational background in music technology, artificial intelligence, experimental psychology with a computational element in it, or computer science with a focus on modelling human behaviour, cognition and/or music.
Personal statements and CVs should show experience and knowledge of computational modelling, human psychology and behaviour, and music. Prior experience with physiological measures is an optional benefit. Further requirements include demonstration of the ability to work effectively independently as well as in the context of a team, and good awareness of research ethics and data management. A clear motivation for the project’s objectives is expected.