BrainHack Speakers

Title: The Human Connectome Project’s Multi-modal Cortical Parcellation

Speaker: Matthew Glasser



Understanding the amazingly complex human cerebral cortex requires a map (or parcellation) of its major subdivisions, known as cortical areas. Making an accurate areal map has been a century-old objective in neuroscience. Using multi-modal magnetic resonance images from the Human Connectome Project (HCP) and an objective semi- automated neuroanatomical approach, we delineated 180 areas per hemisphere bounded by sharp changes in cortical architecture, function, connectivity, and/or topography in a precisely aligned group average of 210 healthy young adults. We characterized 97 new areas and 83 areas previously reported using post-mortem microscopy or other specialized study-specific approaches. To enable automated delineation and identification of these areas in new HCP subjects and in future studies, we trained a machine-learning classifier to recognize the multi-modal ‘fingerprint’ of each cortical area. This classifier detected the presence of 96.6% of the cortical areas in new subjects, replicated the group parcellation, and could correctly locate areas in individuals with atypical parcellations. The freely available parcellation and classifier will enable substantially improved neuroanatomical precision for studies of the structural and functional organization of human cerebral cortex and its variation across individuals and in development, aging, and disease.


Short Biography:

MattMatthew F. Glasser, PhD, is a medical student at Washington University in St. Louis, who did his PhD training with David Van Essen.  Dr. Glasser has over a decade of experience in brain imaging research with a focus on brain anatomy and brain imaging methods development, and has authored or coauthored 44 peer-reviewed articles.  He is best known for his work on reconstructing the arcuate fasciculus, the main connection between the brain’s language areas, developing novel or improved methods for mapping cortical areas, such as mapping the amount of neuronal insulation, called myelin, of the cortical grey matter based on clinical T1-weighted / T2-weighted MRI images, and for producing a new multi-modal map of the human cerebral cortex.  Dr. Glasser is pursuing clinical training in neuroradiology to be a physician-scientist neuroradiologist.



Title: Searching for multiple needles in multiple haystacks: Multimodality imaging and –omics studies of Alzheimer’s disease

Speaker: Andrew Saykin



Great advances have been made in MRI and PET neuroimaging technology. We can use these powerful tools to investigate the structural, functional and molecular bases of neurodegenerative disorders such as Alzheimer’s disease. At the same time, biomarker technologies such as genomics, proteomics and metabolomics provide unparalleled readouts on fundamental pathophysiological mechanisms and a range of other biological processes. The scope and volume of longitudinal data are increasing exponentially. Alzheimer’s disease is a prototype for the current challenge of large scale biomedical data. What kinds of questions should we ask? How can we find the truly informative features in imaging and “–omics” data? What tools are needed to take advantage of multi-layered longitudinal data? Hopefully the unprecedented data being generated coupled with thoughtful analytic strategies will lead to new insights into disease mechanisms, risk and protective factors, methods for early detection, and novel therapeutic targets.


Short Biography:



Andrew Saykin also leads the Genetics Core of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and is the founding Editor-in-Chief of Brain Imaging and Behavior, a Springer Nature journal.






Title: Computational Challenges and Opportunities in Network Neuroscience

Speaker: Olaf Sporns



Modern neuroscience is in the middle of a transformation, driven by the development of novel high-resolution brain mapping and recording technologies that deliver increasingly large and detailed “big neuroscience data”. Network science has emerged as one of the principal approaches to model and analyze neural systems, from individual neurons to circuits and systems spanning the whole brain [1,2].

A core theme of network neuroscience is the comprehensive mapping of anatomical and functional brain connectivity, also called connectomics. In this presentation I will review current themes and future directions of network neuroscience [3], including comparative studies of brain networks across different animal species, investigation of prominent network attributes in human brains, and use of computational models to map information flow and communication dynamics. I will argue that network neuroscience represents a promising theoretical framework for understanding the complex structure, operations and functioning of nervous systems.


[1] Sporns, O. (2011) Networks of the Brain. MIT Press.

[2] Sporns, O. (2014) Contributions and challenges for network models in cognitive neuroscience. Nature Neuroscience 17, 652-660.

[3] Bassett, D.S. and Sporns, O. (2017) Network neuroscience. Nature Neuroscience (in press).

Short Biography:

ospornsAfter receiving an undergraduate degree in biochemistry, Olaf Sporns earned a PhD in Neuroscience at Rockefeller University and then conducted postdoctoral work at The Neurosciences Institute in New York and San Diego. Currently he is the Robert H. Shaffer Chair and a Distinguished Professor in the Department of Psychological and Brain Sciences at Indiana University, and serves as co-director of the IU Network Science Institute. His main research area is theoretical and computational neuroscience, with a focus on complex brain networks. In addition to over 200 peer-reviewed publications he is the author of two books, “Networks of the Brain” and “Discovering the Human Connectome”. In 2016 he became the Founding Editor of “Network Neuroscience”, a journal published by MIT Press devoted to theintersection of brain and network sciences. Sporns was awarded a John Simon Guggenheim Memorial Fellowship in 2011 and was elected Fellow of the American Association for the Advancement of Science in 2013.



Title: Software tools motivated by analysis of fMRI data

Speaker: Kendrick Kay

As a visual neuroscientist who takes a data-driven and model-based approach to fMRI, I have created a number of software tools for fMRI analysis over the years. In this talk, I will give short teasers of these tools, which include: co-registration of brain volumes; pre-processing and fMRI data quality; computational modeling of fMRI time-series; GLMdenoise; and automated high-resolution cortical surface renderings.


Short biography:

2728338Kendrick Kay, Ph.D., is an Assistant Professor at the Center for Magnetic Resonance Research at the University of Minnesota. He uses experimental and modeling techniques to investigate how the visual system represents stimuli and makes perceptual judgments about these stimuli. He is an expert in functional magnetic resonance imaging methods, computational modeling, and data analysis methods.





Title: Microstructure Imaging of Complex White Matter Architecture from Diffusion MRI

Speaker: Hamza Farooq



Detailed tissue microstructure information (like axon diameter, density indices etc.) can be extracted from diffusion MRI (dMRI) data using multi-compartment biophysical models. These models can explain signals from dMRI in (i) intra and extra axonal spaces and (ii) along multiple fiber orientations, inside a voxel. However, estimation of those models’ parameters is an ill-posed problem because of their complex mathematical properties and the noise in the data.

Until recently, estimation/optimization techniques (like CAMINO and AMICO etc.) were limited to a single fiber orientation, although the vast majority of the brain white matter contains crossing fibers (at the current imaging resolution). We introduce a novel regression method, which we refer to as MIX (Microstructure Imaging in Crossing Fibers), that overcomes this barrier and enables the use of existing biophysical models to multiple fiber orientations. Additional attributes of MIX are: (i) improved accuracy for existing models like NODDI or ActiveAx, (ii) ease in initializing the parameter search, and (iii) ability to handle more complex/realistic models, which opens up the possibility to propose and test new models that were impossible to solve until now.


Short Biography:

humzaHamza Farooq is a Doctoral Candidate at Department of Electrical and Computer Engineering, University of Minnesota. He received his M.Sc. degree in Electrical Engineering from University of Minnesota in April 2016. His research interests include detection and estimation, optimization, medical imaging, diffusion MRI and geometrical methods for brain network analysis.