Abstract: Wearable devices, allowing the monitoring of a wide range of physiological parameters, are having a transformative impact on health and wellbeing applications. We are now able to collect longitudinal data in a way never previously possible. At the same time, current wearables have a wide range of open opportunities for making them even better, from improving accuracy, to improving comfort, to embedding real-time machine learning bases analyses. This talk will provide an overview of these different topics. It will start with a high level overview and roadmap for future wearable hardware, with particular emphasis on advances being enabled by emerging flexible electronics. I will then present two examples of our work on data science and machine learning applied to wearable data. Example 1 will focus on the analysis of the 100,000 accelerometer data records in the UK Biobank, investigating how activity patterns vary across this large cohort and how we can use this to inform the design for future wearables. Example 2 will focus on real-time machine learning applied to wearable brain (EEG) data. I will overview a deep autoencoder based approach for removing artefacts that operates using neural network delegates on a smartphone based platform to substantially out-perform previous artefact removal approaches for real-time application.
Biography: Alex Casson is Professor of Biomedical Engineering at the University of Manchester. He is a specialist in non-invasive bioelectronic interfaces: the design and application of wearable sensors, and skin-conformal flexible sensors, for human body monitoring and data analysis from highly artefact prone naturalistic situations. This work is highly multi-disciplinary, spanning ultra-low power sensing, signal processing and machine learning in power constrained rich environments, and real-time data analysis towards closed loop systems for remote monitoring and digital therapeutics. He has research experience in:
- Manufacturing using 3D printing, screen printing, and inkjet printing.
- Ultra low power microelectronic circuit and system design at the discrete and fully custom microchip levels.
- Sensor signal processing and machine learning for power and time constrained motion artefact rich environments.
- Using bespoke and off-the-shelf wearable devices in a wide range of environments.
Professor Casson’s ultra low power sensors work is mainly for health and wellness applications, with a strong background in brain interfacing (EEG and transcranial current stimulation) and heart monitoring. Applications focus on both mental health situations including chronic pain, sleep disorders, and autism, and physical health/rehabilitation applications including diabetic foot ulceration, and chronic kidney disease. He has particular interests in closed loop systems: those which are tailored to the individual by personalised manufacturing via printing; and tailored to the individual by adjusting non-invasive stimulation (light, sound, electrical current) using data driven responses/outputs from real-time signal processing.
Professor Casson has cross-disciplinary appointments to support this work. He is a Professor in the School of Engineering at the University of Manchester; Visiting Reader (2022-2024) in School of Medicine at the University of Leeds; and Honorary Reader (2022-2024) in the Medical Physics Department at Northern Care Alliance NHS Foundation Trust. He is a Future Leader in Innovation, Enterprise and Research (FLIER) for the Academy of Medical Sciences (2022-2024), Bioelectronics technology platform lead for the Henry Royce Institute, and a Fellow of the Alan Turing Institute (2021-2023). Professor Casson is currently a Senior Member of the IEEE and Fellow of the Higher Education Academy. He is a past chair of the IET Healthcare Technologies Network, and the Biomedical Engineering joint steering group.
Professor Casson gained his undergraduate degree from the University of Oxford where he read Engineering Science specialising in Electronic Engineering (MEng). He completed his PhD from Imperial College London, winning the prize for best doctoral thesis in electrical and electronic engineering. Professor Casson worked as a research associate and research fellow at Imperial College until 2013 when he joined the faculty at the University of Manchester.
