Bio:

Dr. Juan Helen ZHOU is an Associate Professor at the Centre for Sleep and Cognition and Director of the Centre for Translational MR Research at the Yong Loo Lin School of Medicine, National University of Singapore. She is also affiliated with the NUS Medicine Healthy Longevity & Human Potential Translational Research Program, the Department of Electrical and Computer Engineering at the School of Design and Engineering, NUS.

Her research focuses on selective brain network-based vulnerability in aging and neuropsychiatric disorders, leveraging multimodal neuroimaging and machine learning approaches. She is widely recognized for her pioneering work on brain connectomics, particularly in understanding network-based breakdown in aging, neurodegenerative, and cerebrovascular disorders. More recently, her contributions to brain decoding and brain foundation models using deep learning have gained significant attention.

Helen earned her bachelor’s and Ph.D. in Computer Science at Nanyang Technological University, Singapore. She completed her postdoctoral fellowship at the Memory and Aging Center, Department of Neurology, University of California, San Francisco (UCSF). She also worked in the Computational Biology Program at the Singapore-MIT Alliance and the Department of Child and Adolescent Psychiatry, New York University.

Helen has served as a Council Member and Program Committee member of the Organization for Human Brain Mapping. She is on the advisory board of Cell Reports Medicine and serves as an editor for eLife, Human Brain Mapping, and the Journal of Alzheimer’s Disease. Additionally, she contributes to the organization committees of multiple international conferences and is a member of ISMRM and ISTAART. She served as Treasurer of Society for Neuroscience Singapore chapter and board member of ISMRM Singapore Chapter. Helen’s research has been supported by various funding bodies in Singapore, the Royal Society (UK), and the NIH (USA).

Abstract: Integrating Brain Imaging and AI: Applications in Neurological Disorders

Advances in brain imaging and AI provide an unprecedented opportunity to explore the human mind and develop new approaches for treating neurological disorders. Each neurodegenerative disorder affects distinct large-scale brain networks. This talk will focus on brain network phenotypes in neurological disorders such as Alzheimer’s and cerebrovascular disease. Specifically, how these network phenotypes relate to pathology, help identify at-risk groups, and predict cognitive decline. Our recent work on AI-driven models for brain decoding and interpretable brain foundation models with efficient adaptation strategies will be discussed. Moving forward, integrating AI with brain imaging paves the way for improved early diagnosis and treatment strategies for neuropsychiatric disorders.

Bio:

Dr. Dimitri Van De Ville received his Ph.D. degree in computer science engineering from Ghent University, Belgium, in 2002. He was a post-doctoral fellow (2002-2005) at the lab of Prof. Michael Unser at the Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland, before becoming responsible for the Signal Processing Unit at the University Hospital of Geneva, Switzerland, as part of the Centre d’Imagerie Biomédicale (CIBM). In 2009, he received a Swiss National Science Foundation professorship and since 2015 became a Professor of Bioengineering at the EPFL, jointly affiliated with the University of Geneva, Switzerland. His main research interest is in computational neuroimaging to advance cognitive and clinical neurosciences. His methods toolbox includes wavelets, sparsity, deconvolution, graph signal processing. He was a recipient of the Pfizer Research Award 2012, the NARSAD Independent Investigator Award 2014, the Leenaards Foundation Award 2016, and IEEE EMBS Technical Achievement Award 2024. He was elevated to Fellow of the IEEE in 2020 and the EURASIP in 2023.

Dr. Van De Ville has served as a Senior Member of the Editorial Board for the IEEE Signal Processing Magazine since 2021, as a Handling Editor for the new journal Imaging Neuroscience since 2023, and as an Editor for the SIAM Journal on Imaging Science since 2018. Before, he served for the IEEE Transactions on Signal Processing, IEEE Transactions on Image Processing, and IEEE Signal Processing Letters. He was the Chair of the Bio Imaging and Signal Processing (BISP) TC of the IEEE Signal Processing Society (2012-2013) and the Founding Chair of the EURASIP Biomedical Image & Signal Analytics SAT (2016-2018). He is Co-Chair of the biennial Wavelets & Sparsity series conferences, together with Y. Lu and M. Papadakis.

Abstract: Brain Networks & Graph Signal Processing: A Perfect Fit

State-of-the-art neuroimaging, such as magnetic resonance imaging (MRI), provides unprecedented opportunities to non-invasively measure human brain structure (anatomy) and function (physiology). To fully exploit the rich spatiotemporal structure of these data and gain insights into brain function in health and disorder, novel signal processing and modeling approaches are needed, instilled by domain knowledge from neuroscience and instrumentation. I will highlight one research axis that tailors and applies graph signal processing to neuroimaging data. This approach integrates a brain graph (i.e., the structural connectome determined by diffusion-weighted MRI and tractography) and graph signals (i.e., the spatial activity patterns obtained by fMRI). The latter are decomposed onto a graph harmonic basis defined through the eigendecomposition of the graph Laplacian. Spectral filtering operations are then designed to separate brain activity into its structurally aligned and liberal parts, respectively, which allows quantifying of how strongly the underlying structure shapes function. The structure-function strength throughout the brain uncovers a behaviorally relevant spatial gradient from uni- to transmodal regions, which is also informative about task conditions or identifying individuals. Statistical assessment can be performed using a surrogate data approach that preserves smoothness, as measured by the Laplacian. Finally, I will indicate some recent trends and challenges in the field to surpass current temporal and spatial limitations of fMRI and thus advance analysis and modeling techniques.

Bio:
Dr. Jayanthi Sivaswamy is a Professor at the International Institute of Information Technology (IIIT) Hyderabad, where she holds the Raj Reddy Chair and leads significant research initiatives in image processing, medical image computing, and computer vision. Her academic journey began with a Bachelor's degree in Electrical Engineering from the Rochester Institute of Technology, followed by a Master's and PhD in Electrical Engineering from Syracuse University. Dr. Sivaswamy's research focuses primarily on medical image analysis, particularly in developing computer-aided diagnostic tools and enhancing medical education through technological innovations.

She has played a pivotal role in several high-impact research projects, including the creation of the Indian Brain Atlas, which provides crucial insights into brain structure and aging in the Indian population. Her work also includes the development of automated screening techniques for diseases such as glaucoma and diabetic retinopathy, as well as creating virtual reality tools for medical education. Dr. Sivaswamy has published extensively in top-tier journals and conferences, contributing to advancements in medical imaging and its applications.

In addition to her research, she has held leadership roles at IIIT Hyderabad, including Dean of Academics and Chair of the Electronics and Communications Department. Dr. Sivaswamy has also championed diversity and inclusion, implementing initiatives to enhance student representation across gender and regional lines.

Abstract: Towards understanding normative aging of the human brain in a population

In this talk, I will present the challenges encountered and solutions developed while studying brain aging in the Indian population. The study began by creating a brain atlas for a young age group to capture structural variability before the onset of aging ; the atlas laid the groundwork for understanding population-specific differences in brain structure. Next, aging-related changes in the brain were analysed and compared with other populations to identify population-driven differences in brain aging. A cross-sectional study was designed to simplify data collection, considering practical aspects. Data was carefully curated and efforts were made to minimize demographic variability given the small-scale of the study. As the study relied on cross-sectional data, customized techniques were developed to help analyse aging-related changes which was also used to compare aging trends across different populations. I conclude with insights gained at various levels in the study including data collection and data curation.

Bio:

Shaoshi Zhang is a research scientist at the Centre for Sleep and Cognition, National University of Singapore. His research focuses on computational modeling, neurodevelopment, neurodegeneration, and the application of machine learning techniques to neuroimaging. In this talk, he will explore how sample size and scan duration impact fMRI brain-wide association studies and discuss their implications for future study designs.

Abstract: Longer scans boost prediction and cut costs in brain-wide association studies

A pervasive dilemma in brain-wide association studies (BWAS) is whether to prioritize functional MRI (fMRI) scan time or sample size. We derive a theoretical model showing that individual-level phenotypic prediction accuracy increases with sample size and total scan duration (sample size × scan time per participant). The model explains empirical prediction accuracies extremely well across 76 phenotypes from nine resting-fMRI and task-fMRI datasets (R2 = 0.89), spanning a wide range of scanners, acquisitions, racial groups, disorders and ages. For scans ≤20 mins, prediction accuracy increases linearly with the logarithm of total scan duration, suggesting interchangeability of sample size and scan time. However, sample size is ultimately more important than scan time in determining prediction accuracy. Nevertheless, when accounting for overhead costs associated with each participant (e.g., recruitment costs), to boost prediction accuracy, longer scans can yield substantial cost savings over larger sample size. To achieve high prediction performance, 10-min scans are highly cost inefficient. In most scenarios, the optimal scan time is ≥20 mins. On average, 30-min scans are the most cost effective, yielding 22% cost savings over 10-min scans. Overshooting is cheaper than undershooting the optimal scan time, so we recommend aiming for ≥30 mins. Compared with resting-state whole-brain BWAS, the most cost-effective scan time is shorter for task-fMRI and longer for subcortical-cortical BWAS. Standard power calculations maximize sample size at the expense of scan time. Our study demonstrates that optimizing both sample size and scan time can boost prediction power while cutting costs.