Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
ISIS Clustering Functional Connectivity
1. Introduction
– Independent Component Analysis (ICA) is a
data-driven method to decompose functional
neuroimaging data into brain networks.
– The decomposed independent components
encompass a mix of true neural signal,
machine artifact, motion, and physiological
noise that are typically distinguished based on
visual inspection.
– Neurological disorders are beginning to be
understood based on aberrant brain structure
and function on the single network level.
– No methods exist for identifying patterns
across all networks to distinguish disorder.
Clustering of Functional Brain Networks to Classify Disorder
Preliminary Work
Vanessa Sochat
Stanford University
Objective
Identify patterns of brain networks for
diagnosis of neuropsychiatric disorder and
better definition of subtypes.
Rubin Lab
Research supported by NSF, SGF, Microsoft Research
vsochat@stanford.edu
Methods
1. Modify ICA algorithm to derive primary and
sub brain networks.
1. Develop functional network features that can
be used to:
Distinguish good vs. bad components
Distinguish network X from network Y
1. Use unsupervised clustering to group
networks
Use supervised clustering to identify and
remove noisy components
Assign groups to associated sub-networks.
Use patterns of sub-networks to distinguish
disorder
Neuroinformatics Landscape
Start with a question:
How does brain function differ between groups?
Collect functional and structural MRI data:
1
2
Preprocess data to overlay function on structure3
Segmentation
Registration
Normalization
Realign /
Reslice
Motion
Correction
Segmentation Smoothing Filtering
4 Higher Level Analysis to Answer Question
How does brain function differ between groups?
Statistical Test
5
Hypothesis
Driven
Data Driven Decomposition (ICA)
What are resting state functional networks?
More Statistics!
Signal Processing
What are features that describe these brain maps?
6 Can patterns of functional networks predict groups?
Machine Learning
A
B
Clinical Data
Data Repository
Decision Support
Biological
Significance
4
5
6
Pamino et. Al, 2012