Simulated data and relation between HRF-conv. GC and neural GC.

<p>A: Flowchart illustrating the process from neural to HRF-convolved neural to fMRI data. B: An example of the process in A where the neural time series (blue) was generated using AR(1) model and was convolved with a canonical HRF to yield the HRF-convolved neural time series (red), which, after down-sampling to TR = 2 s and addition of 20% white noise (SNR = 5), became the fMRI time series (Green). C: GC for HRF-convolved neural time series as a monotonically increasing function of neural GC where the slope of fitted linear trend is close to 1. D: HRF-conv. GC in the opposite direction is zero (unidirectional coupling).