Directional connectivity of resting state human fMRI data using cascaded ICA-PDC analysis
Minna J Silfverhuth1, Jukka Remes, Tuomo Starck
1Department of Computer Science and Engineering, University of Oulu, Linnanmaa, Finland. Minna.Silfverhuth@oulu.fi
Background:
Directional connectivity measures, such as partial directed coherence (PDC), give us means to explore effective connectivity in the human brain. By utilizing independent component analysis (ICA), the original data-set reduction was performed for further PDC analysis.
Purpose:
To test this cascaded ICA-PDC approach in causality studies of human functional magnetic resonance imaging (fMRI) data.
Material And Methods:
Resting state group data was imaged from 55 subjects using a 1.5 T scanner (TR 1800 ms, 250 volumes). Temporal concatenation group ICA in a probabilistic ICA and further repeatability runs (n = 200) were overtaken. The reduced data-set included the time series presentation of the following nine ICA components: secondary somatosensory cortex, inferior temporal gyrus, intracalcarine cortex, primary auditory cortex, amygdala, putamen and the frontal medial cortex, posterior cingulate cortex and precuneus, comprising the default mode network components. Re-normalized PDC (rPDC) values were computed to determine directional connectivity at the group level at each frequency.
Results:
The integrative role was suggested for precuneus while the role of major divergence region may be proposed to primary auditory cortex and amygdala.
Conclusion:
This study demonstrates the potential of the cascaded ICA-PDC approach in directional connectivity studies of human fMRI.
