Resting-State Functional Connectivity Analysis at 3T

Within each surface-projected resting-state run, vertex-wise Pearson’s product-moment correlations were computed to generate a correlation matrix. Correlation matrices were r-to-z transformed before averaging across runs within each participant. We defined DN-A and DN-B in the 3T data using the same process as described in the 7T data, including seed-based and data-driven clustering approaches. The MS-HBM parcellation method was again applied to this dataset to define networks,¹⁶³ which is supposed to improve stability in the network estimation procedure. In each individual, we used a k value between 14–20 and selected the solution (k = 14) that best matched the networks observed by the seed-based approach.