Computational approaches for mapping the human connectome

Lecture / Panel
For NYU Community

Speaker: R. Cameron Craddock, PhD, Director, Computational Neuroimaging Lab, Center for Biomedical Imaging and Neuromodulation, Nathan S. Kline Institute for Psychiatric Research, Orangeburg, NY, Director of Imaging, Center for the Developing Brain, Child Mind Institute, New York, NY

The human connectome is a graph representation of the brain’s functional and structural architectures. In this graph, nodes represent brain areas and edges represent either anatomical connections (fibers) or synchrony between physiological measures of brain activity that occurred at those areas. This latter method has been called “functional connectivity” based on the assumption that the measured synchrony is a result of the implicated brain areas working together to perform some function. Estimating the functional interactions between brain areas and mapping them to corresponding inter-individual variability (i.e. age, sex, IQ, disease state, disease severity) are “big data” problems that require very large numbers of high quality datasets and the tools and resources to efficiently analyze them. In this talk, I will discuss some of the computational challenges inherent in brain connectivity research and discuss the tools and algorithms developed in the Computational Neuroimaging Lab (CNL) for addressing them. These include classifier and feature selection for identifying functional connectivity based biomarkers of disease, unsupervised identification of brain areas to be used as graph nodes, multivariate regression to decode ongoing brain activity, and brain computer interfaces for probing network dynamics. Additionally I will discuss CNL’s ongoing open science initiatives that include developing software pipelines for performing high capacity data processing and analysis, as well as, aggregating and sharing large-scale neuroimaging datasets. The overarching goal of these initiatives is to make tools and data accessible to a wider audience of data science researchers so as to accelerate the pace of discovery in neuroimaging neuroscience.


R. Cameron Craddock, PhD, is a computer engineer who combines an extensive knowledge of MRI acquisition and analysis methods with computational sciences to research the impact of development and mental health disorders on brain function. Dr. Craddock earned his Bachelor’s in Computer Engineering from Georgia Tech and worked for two years in the telecommunications industry before returning the Georgia Tech for his graduate education. His PhD dissertation, "Support Vector Classification Analysis of Resting State Functional Connectivity fMRI," was completed under the advisement of Dr. Helen Mayberg and Dr. Xiaoping Hu. Dr. Craddock joined the Child Mind Institute and Nathan S. Kline Institute for Psychiatric research in 2012 after completing post-doctoral fellowships at the Baylor College of Medicine and Virginia Tech Carillion Research Institute. Dr. Craddock is a strong advocate of open science. He is the co-founder of The Neuro Bureau, a collaborative initiative that supports open neuroscience and advocates for the free sharing of data, methods, and ideas. He is also an organizer of international and regional Brainhack events that bring together brain enthusiasts to work in collaboration on neuroscience projects. He is a recipient of a NARSAD Young Investigator Award from the Brain and Behavior Research Fund and a NIMH Biobehavioral Research Award for Innovative New Scientists (BRAINS R01) for his research that uses real-time fMRI neuro-feedback to investigate the role of network dysregulation in mental health disorders. Dr. Craddock is a reviewer for many respected journals in the field, and the recipient of several honors and awards, including the first place poster award for functional imaging at the 19th scientific meeting of the International Society for Magnetic Resonance in Medicine in 2011.

For more information, please contact:  Prof. Guido Gerig.