Using our proposed pipeline, we found close agreement with all three modalities as presented in Fig. 3. The activations for all three modalities were localized to the primary somatosensory (postcentral) gyrus and primary motor (precentral) gyrus. Although the fNIRS and MEG spatial resolution is generally coarser than this, the additional information added by the cortical surface used in the image reconstruction process adds spatial information to constrain the images. Thus, with the addition of this structural information, we found that the overall spatial resolution of the fNIRS images was comparable to the BOLD fMRI. We found that for shallower regions of the brain, there was only a 9- and 16-mm displacement between the BOLD and fNIRS activations [center-of-mass; and Hb, respectively)] and and correlation between the two maps at a depth of 20 mm. This error was similar to that observed between the MEG and BOLD signal (10.4 mm) with a similar correlation (). To our knowledge, this is the first study to use the pulsed-pair task with fNIRS. However, our results are similar to the work by Stevenson et al.34 that used a similar pulsed-pair median nerve task during independent sessions of fMRI and MEG in seven subjects. They reported good spatial agreement between the two modalities. In their study, the displacement of the average BOLD and MEG (dipole) locations was found to be , which is similar to the 25-mm displacement found in this study for the whole head analysis, but greater than the 7.1-mm displacement found when only the superficial layers of the cortex were considered in the analysis. The relatively large inaccuracy of the signals in MEG and fNIRS for depths beyond 30 mm suggests that it is prohibitively difficult to image deep brain sources, such as subcortical regions, with these methods. Previously, this falloff of signal with depth had been modeled using simulated data,46–48 but the current work provides an empirical examination of this issue. One caveat of the current study is that it is done in the context of median nerve stimulation, which provides a relative context for comparing localization.