AI in Pervasive Well-Being and Healthy Ageing

In contrast to the field of adult neuroimage, which has well validated tools and widely accepted best practices for data acquisition, processing, and analysis, the field of early childhood neuroimaging is still nascent. The rapid brain growth and the profound changes in brain anatomy and physiology (ie. myelination and synaptogenesis) occurring during early childhood affect tissue properties, and consequently the MRI signal intensity and the resulting image contrasts. Thereby, young child’s brain present significant and unique challenges for data processing and analysis that cannot be addressed using adult’s tools which the algorithmic parameters and assumptions (eg. tissue priors, brain size, contrast differences, motion artifacts) are not tuned for the anatomy and physiology of a young child's brain and their rapid change.

Artificial intelligence techniques have progressed exponentially in recent years in all fields including neuroimaging. Can future advances help address the inherent challenges of imaging the uniqueness of young child’s brain?

The UN Decade of Healthy Ageing describes wellbeing as a complex interaction between our physical, sensory, vitality and psychological capacities and the environment we live in. Places, spaces, relationships and communities have a big impact on how we develop and maintain functional ability in later life. Supporting older people to remain active and engaged in communities and taking action to tackle social isolation and loneliness are fundamental to enabling older people to thrive in later life.   

People are living longer but the prevalence of multiple chronic conditions and frailty increases with ageing. Frailty affects up to half of the population aged over 85 and costs UK healthcare systems £5.8billion per year. However frailty is not an inevitable part of ageing, and putting in place measures to slow its onset or progression should be a priority for every health and care system. Prevention and reversal of frailty enables people to live independently for longer and helps to reduce the increasing demand for emergency care and long-term support. But our current health and care services are largely reactive and don't prioritise investment in the preventative or proactive care that can reduce escalation of dependency and avoid associated costs.

This session will outline policy recommendations from the European Joint Action on Frailty and describe the seven system Touchpoints in the British Geriatrics Society Blueprint for preventing and managing frailty.  Participants will explore opportunities for academics, industry and health, social care and housing providers to work together to improve population health through targeted AI and digital solutions at each of these Touchpoints. A collaborative approach, valuing the contribution of experts by experience, is required to design, test and scale innovative integrated solutions for proactive and personalised care in order to enable wellbeing in later life and support older adults to remain independent at home for longer.

Automating the prediction and classification of neurological disorders lead to a wide range of benefits, including time and cost saving for both the clinicians and patients/carers, reducing the risk of human error and subjectivity in the assessment, as well as supporting remote assessment to minimize the geographical barriers for providing timely healthcare support.

In this talk, a wide range of our team's ongoing research on human motion analysis for predicting and classifying neurological disorders will be presented. In particular, our methods are mostly focused on using machine learning techniques for analysing skeletal motion data extracted from vision-based sensors. We further enhanced the interpretability of the proposed models with intuitive visualization.

Motivation is often a major obstacle to adherence to an exercise plan. For instance, poor adherence to rehabilitation programs has been observed, despite the importance they can have, notably in the treatment of cardiovascular conditions. In this work, we explore how nudging through personalized incentives (small monetary rewards spread over the day) can improve adherence to a daily exercise goal in participants.

For personalizing the incentives, we are using the active inference framework that allows us to deal with the trade-off between learning about how the policy concerning the incentives affects the participant's behavior and helping them meet their objective.

My research aims to develop and implement accurate and efficient algorithms for MRI analysis for research and clinical neuroscience. The talk will discuss the challenges of MRI image analysis, such as noise, variability, and the need for manual segmentation. I will then describe how deep learning techniques can overcome these challenges and provide more accurate and reliable results. Specifically, I will present my proposed deep learning architectures and the results of experiments conducted on a large dataset of MRI scans.

Finally, I will present potential applications of deep learning for MRI analysis in the context of general well-being and healthy ageing. The proposed algorithms have the potential to improve early diagnosis and treatment of age-related neurodegenerative diseases, ultimately leading to better health outcomes and quality of life for ageing populations.

I will provide an indicative overview of digital technologies and how these have been successfully implemented to address a range of challenging problems in healthcare, in particular to facilitate diagnosis, monitoring of symptoms, and assessment of interventions during rehabilitation. I will outline some of the promising developments that are happening in this space, key challenges in translating research outputs to mainstream technologies, and give my view on how future home-based healthcare might look like.

Interindividual differences across the life-course, including lifestyle, play a significant role in healthy ageing. Not everyone who has amyloid pathology or brain atrophy, who smokes or has midlife hypertension will develop dementia, although these are all risk factors. By identifying those at high risk early in the disease process they can be targeted for treatment. However, there are currently no accurate and objective methods to determine an individual's personal risk of developing dementia.

AI represents a novel and powerful tool for the detection of neuropathological burden and therefore dementia risk.

 I am a Consultant in Older Peoples Medicine in NHS Lanarkshire.  As a recently appointed and first cohort of Innovation Fellows in Scotland I will take you through my innovation journey.  This fellowship focuses on the area of falls prevention in older people who are at high risk of falls or who have already fallen.  I will share this problems statement and horizon scanning that has been completed to date.  In addition will cover the innovation structure and how that supports the quadruple helix within NHS Scotland and Scottish Government. 

The benefits of remote monitoring technology as an enabler for virtual wards are currently well-publicised, and by applying a joined-up approach across health and social care it can offer even greater value, enabling proactive and preventative care, which in turn will reduce hospital admissions, and better support ongoing wellbeing.

The presentation will discuss how remote monitoring solutions can be tailored specifically to meet a service or individual’s needs and applied across both health and social care settings, for ongoing wellbeing management and to monitor for increasing risk escalations. Learn about the technology, as we share how our unique platform uses real-time data (collected from a wide variety of devices) and predictive models, to turn data into risk alerts and actionable insights for clinicians, practitioners, and carers.