Multimodal Neuroimaging and Artificial Intelligence (AI) Models in Old Age Psychiatry
老年精神疾病的多模态神经影像和人工智能模型
报告人：Li Su, PhD, Principle Investigator
剑桥大学精神病学系
Department of Psychiatry, University of Cambridge, UK
时间：2019年3月1日 下午2-4点
地点：北京师范大学京师学堂大楼第三会议室（地下一层）
报告人简介：
Dr. Su now is a Principle Investigator leading the Computational Psychiatry Laboratory and NeuroTechnology Group in the Institute for Neuroscience at the University of Cambridge. He is also an affiliated member of the EPSRC Centre for Mathematical Imaging in Healthcare at Cambridge and has led the neuroimaging research in PREVENT-dementia, a multi-centre international consortium. He has a distinguished track record in the development of innovative methods of imaging analysis and their applications to psychiatric and late-life degenerative disorders, and is a renowned expert in MR, PET, EEG and MEG as well as Artificial Intelligence (AI) based computational modelling. He has authored several high impact papers that have been widely cited in the field. He was awarded the International College of Geriatric Psycho-neuropharmacology Junior Investigator Award (2015) and International Psychogeriatric Association Junior Research Award in Psychogeriatrics (2016).
报告摘要：
Neural degeneration and dysfunction leading to dementia may begin decades before symptoms and cognitive decline present. Hence, the recognition of early imaging markers of dementia may allow more effective intervention or even prevent dementia from happening. In addition, brain inspired Artificial Intelligence (AI) is beginning to achieve if not surpass human-level performance. Here, I will present current and recent neuroimaging research in seeking for potential early neuroimaging biomarkers for different types of dementia, such as Alzheimer's disease and Lewy body dementia. I combine multimodal imaging (MEG, EEG, MRI and PET) with genetics and AI based computational modelling to reveal neural mechanisms underlying dementia, and how risk factors (e.g. family history of dementia, APOE and MAPT genotype) affect brain structure, functions and behaviours across lifespan. Multimodal imaging and genetics have the potential to identify early changes in brain structure and functions at both regional and molecular levels. However, characterizing their relationships and neural mechanisms requires formal modelling such as computational psychiatry approaches.
