Figure 4.

Schematic illustration of (A) the ROIs for the fMRI data extraction and (B) the channel layout and corresponding fNIRS ROIs approximated by means of the fOLD toolbox⁵ for the fNIRS data extraction. In addition, on the right side of (B) the channel frequency across subjects after channel pruning is visualized. For each subject the channel positions and the individual anatomical images were co-registered and channels were labeled as covering one of the fMRI ROIs or not. (C) Illustrates the frequency at which channels were labeled as covering an fMRI ROI. For easier comparison channels are overlaid on the fNIRS ROIs. The fNIRS ROIs and the corresponding channels are separately illustrated because some channels were assigned to more than one ROI (e.g., channel 19). The lower part of the figure illustrates all steps taken to extract data types (D) fMRI PEAK, (E) fNIRS $\mathrm{\Delta }[HbO]$ and $\mathrm{\Delta }[HbR]$, (F) fNIRS $\mathrm{\Delta }[HbO]$ and $\mathrm{\Delta }[HbR]$ LABELED and (H) fMRI CHANLOCS. Boxes representing the final data type are highlighted with solid or dashed black lines. Data types (D,E) were derived directly from the functional data sets. For data types (F,H) co-registration of MRI and NIRS data was required. The large grey box (G) highlights the coregistration steps. In the figure, arrows between and within the smaller boxes illustrate the work flow. Arrows pointing from A to D and from B to E indicate that the areas from which functional data were extracted were restricted by a-priori ROI definitions. The arrow pointing from F to C indicates that (C) shows a result of co-registration and of labelling fNIRS channels based on individual anatomy.