neurdylab / deep-physio-recon

MIT License
4 stars 4 forks source link

R3C3: network based statistical analysis #8

Open cately opened 6 months ago

cately commented 6 months ago

R3 C3: The observations about the differences between the seed-based FC and connectome analyses are written in a very qualitative manner. Provide a more quantitative analysis that demonstrates the similarity between the maps and matrices? For example, a paired t-test network based statistics analysis between the matrices?

Lines in ref to the above comment. "We observe that both the ROI-based (Figure 2d - Single Scan) and seed-based (Figure 2e - Single Scan) correlations are altered with the removal of physiological signals, and that comparable results are obtained using the measured and decoded physiological waveforms, as shown for an example scan."

cately commented 6 months ago

voxelwise 3d t-tests over the pairs of maps, threshold via FDR, make sure to test both directions (pos, neg correlations)

rgbayrak commented 4 months ago

fisher z-tranform before t-test.