O3D

Open Diffusion Data and Derivatives



What is O3D

Open Diffusion Data and Derivatives (O3D) is an online repository of data to support the analysis of brain connectivity.

O3D Repository

You can download a rich dataset composed of dMRI obtained from healthy, cognitively normal human individuals and derivatives in a BIDS-standard data structure for further analysis.

O3D Pipeline

You can reproduce, or generate your own O3D data derivatives through O3D online pipeline by uploading your data to our online workflow submission system.



Who can use it?


Neuroscientists

Neuroscientists interested in developing algorithms for community detection and network science on brain network data, without having to process the raw diffusion data into connectivity matrices. Our data set provides a jump-start mechanism by making available unthresholded raw brain connection matrices built in multiple individuals using different tracking methods.

Computer Scientists

Computer scientists are interested in implementing algorithms to perform white matter fascicles registration or clustering (REF). Our data set provides multiple sets of human white matter tracts segmented in several subjects with repeated measures and estimates made using multiple tractography methods.

Students

Investigators studying white matter, neuroanatomy, as well as software developers for continuous testing (Dipy, Travis) or MRI physicists could use the data Derivatives as reference given current state-of-the-art methods. Importantly, the Derivatives dataset will be of help for students and beginning neuroimaging trainees.




Datasources

Stanford datasets

STN
Ddata collected in four subjects (age 37-39) at the Stanford Center for Cognitive and Neurobiological Imaging with a 3T General Electric Discovery 750 (General Electric Healthcare), using a 32-channel head coil (Nova Medical).

- http://purl.stanford.edu/rt034xr8593
- http://purl.stanford.edu/ng782rw8378

Human Connectome Projects datasets

HCP3T
Data collected in four subjects at the University of Washington St. Louis as part of the Human Connectome Project, with the Siemens 3T "Connectome."

HCP7T
Data collected in four subjects part of the Human Connectome 7-Tesla (7T) dataset were used. Data were collected the University of Minnesota Siemens 7T scanner.

- http://www.humanconnectome.org/data



How is data processed?

White matter fascicles tracking was performed using MRtrix 0.2.12. White- and gray-matter tissues were segmented using the T1-weighted MRI images associated to each individual brain, and then resampled at the resolution of the dMRI data. Only voxels identified primarily as white-matter tissue were used to perform tracking.

Tract evaluation

We used the LiFE (Linear Fascicle Evaluation) method to optimize each fascicle group generated with tractography. The LiFE method identifies fascicles that successfully contribute to predicting the measured dMRI signal. It has been shown that only a small percentage of the total number of fascicles generated through a single tractography method is supported by the properties of given data set. Because of this we removed all fascicles making no significant contribution to explaining the diffusion measurements. The number of streamlines retained in these optimized fascicles groups ranged about 80,000 (STN96) to 160,000 (HCP90). Only fascicles groups processed using LiFE were used for all subsequent analyses.

Tract segmentation

Whole brain fascicles groups comprising only the tracts successfully contributing to predicting the measured dMRI signal as evaluated using LiFE were segmented to generated following major human white matter fascicle.