Suyash P. Awate Suyash P. Awate
Asha and Keshav Bhide Chair Professor
Computer Science and Engineering Department, Indian Institute of Technology (IIT) Bombay
Office: A-214, Kanwal Rekhi Building
Email: my-first-name
Research Publications Teaching Students CV Personal
Segmenting Functional MRI for Brain Connectivity
Wei Liu, Suyash P. Awate, Jeffrey S. Anderson, P. Thomas Fletcher
A functional network estimation method of resting-state fMRI using a hierarchical Markov random field
NeuroImage 100: 520-534 (2014)

Comparison of segmentation methods for group study of rs-fMRI.
Most methods use a one-way approach, either in a subject-group order (left) or a group-subject order (middle).
Our method (right) aims at a joint estimation of both levels of network maps, where group and subject maps help each other in a bidirectional flow.

Our MRF on a graph that includes the voxels of all subject maps as well as the group map.
The set of edges includes between-level links with weight α,and within-subject links with weight β.
The square box on the subject level and time courses repeats J times the nodes in the square, representing all the subjects.
Only the central voxels connection is shown for the between-level links, whereas in practice the links exist on all other voxels.
The BOLD signal variables are shaded, meaning they are set to the observed value.

The group level's mean functional networks estimated from all bootstrapped data by three segmentation methods.
The binary maps of each network are averaged over all bootstrap samples. The average intensity ranges from 0 to 1.
Related Works