University of Southern California
Mrs T.H. Chan Division of Occupational Science and Occupational Therapy
Neural Plasticity and Neurorehabilitation Laboratory | Research

About this image

Big Data Neuroimaging

Big Data Neuroimaging

Big data approaches can help us to predict outcomes from stroke and how people respond to treatments with greater reliability. We have several initiatives in this area:

ENIGMA Stroke Recovery

ENIGMA Stroke Recovery: The ENIGMA Stroke Recovery working group is using a ‘big data’ neuroimaging approach to better understand how changes the brain relate to motor recovery after stroke. We are analyzing thousands of brain MRIs of individuals after stroke collected from around the world. Combining large amounts of data from sites worldwide gives us a better ability to find reliable patterns of brain changes that are related to recovery. The long-term goal is to use these patterns we uncover to predict and prescribe personalized treatments for individuals after stroke.

ENIGMA Homepage; ENIGMA Stroke Recovery Website

Members and Collaborators:
Neda Jahanshad, Paul Thompson (Imaging Genetics Center, USC); Steve Cramer (UCI); ENIGMA & PGC; ENIGMA & CHARGE; ENIGMA & ILAE

National Institutes of Health/National Center for Medical Rehabilitation Research K12 and K01


Analyzing brain MRIs of people after stroke takes a lot of time and work, in part because we have to map out the brain damage (aka, lesions), which can be a different size, shape, and location for each person. Usually we do this by hand, but a single MRI lesion can take hours. We, along with other research groups, are working on ​​developing computer algorithms to do this automatically, but the algorithms need a lot of examples of properly traced lesions to work. We developed ATLAS (Anatomical Tracings of Lesions After Stroke), which is a large database of about 315 manually traced lesions, to share with the research community. Researchers will be able to download and use ATLAS to test and train their own algorithms and build better, faster ways to analyze post-stroke brains. Funded by the NIH-funded Center for Large Data Research and Data Sharing in Rehabilitation. Link to ATLAS data archive forthcoming.

-Paper in Scientific Data
Full citation: Liew, S.-L.*, Anglin, J.M*., Banks, N.W., Sondag, M., Ito, K.L., Kim, H., Chan, J., Ito, J, Jung, C., Khoshab, N., Lefebvre, S., Nakamura, W., Saldana, D., Schmiesing, A., Tran, C., Vo, D., Ard, T., Heydari, P., Kim, B., Aziz-Zadeh, L., Cramer, S.C., Liu, J., Soekadar, S., Nordvik, J.-E., Westlye, L.T., Wang, J., Winstein, C.J., Yu, C., Ai, L., Koo, B., Craddock, R.C., Milham, M., Lakich, M., Pienta, A., & Stroud, A. (2018) A large, open source dataset of stroke anatomical brain images and manual lesion segmentations. Scientific Data, 5, 180011. doi:10.1038/sdata.2018.11​  *denotes equal contributions

-Download the ICPSR dataset (N=304 raw MRIs with lesion segmentations)

-Download the FCP INDI dataset (N=229 normalized MRIs with lesion segmentation)

-Post any issues with the ATLAS dataset on our ATLAS GitHub page.

-Members and Collaborators:
USC NPNL Members: Julia Anglin, Nicolas Banks, Kaori Ito, Jennifer Chan, Joyce Ito, Connie Jung, Stephanie Lefebvre, William Nakamura, David Saldana, Allie Schmiesing, Catherine Tran, & Danny Vo
Collaborators: Matt Sondag, Panthea Heydari, Bokkyu Kim, Hosung Kim, Lisa Aziz-Zadeh, Carolee Winstein, Tyler Ard (USC); Steven Cramer (UCI); Jingchun Liu, Junping Wang, Chunshi Yu (Tianjin Medical University General Hospital); Surjo Soekadar (University of Tuebingen); Jan-Egil Nordvik (Sunnaas Rehabilitation Hospital HT); Lars Westlye (Oslo University Hospital); Lei Ai, Bonhwang Koo, Cameron Craddock, Michael Miham (Child Mind Institute); Matthew Lakich (University of Texas Medical Branch); Amy Pienta, Allison Stroud (University of Michigan)

NIH Center for Large Data Research and Data Sharing in Rehabilitation (CLDR) Pilot Grant
NIH National Center for Medical Rehabilitation Research K01 Grant



Based on our work in ENIGMA Stroke Recovery, we believe that bringing together data from researchers worldwide will be much more effective if we are collecting similar measures of recovery and function. We developed SCORE (Stroke Comprehensive Outcomes for Rehabilitation Excellence) to provide short, medium, and long batteries of tests that we plan to use in prospective data collection, and invite other researchers to join us.


Members and Collaborators:
Michael Borich (Emory U), Keith Lohse (University of Utah)


We have also developed open-source software to support analyses of large-scale neuroimaging dataset.

This is an online game to crowd source quality control of lesion segmentation masks, using a format similar to Tinder but for stroke brains: Play Braindr-les Now!

This is work based on Dr. Anisha Keshavan’s excellent Braindr platform (Braindr Github). Lesion masks were generated by MRIvis and visualqc tools by Dr. Pradeep Raamana (MRIvis and visualqc Github).

Pipeline for analyzing lesions after stroke: This is a toolbox that provides good visual quality control, white matter correction for lesions, lesion visualization, normalization, and lesion overlap calculations: GitHub

SRQL (precursor to PALS):
Toolbox that provides capability for lesion analyses including white matter denoising, lesion visualization, normalization, and more: RIO; GitHub

ENIGMA Wrapper Scripts:
Three simple scripts designed to facilitate ENIMGA analyses: gigascience; GitHub

Laterality Index for FSL:
GIGA; GitHub

Brainhacks (Neuroscience Hackathons)

Brainhacks are a fun collaborative environment for working on neuroimaging problems and developing innovative new software. We are happy to support brainhacks, e.g,. through

Cameron Craddock (Child Mind Institute)

NIH BD2L Centers Coordination Center Hackathon Grant