Short-channel functional near-infrared spectroscopy regressions improve when source-detector separation is reduced

In functional near-infrared spectroscopy (fNIRS) of human cerebral hemodynamics, dedicated surface-sensitive recording channels are useful for regressing out background hemodynamics and isolating activation-specific responses. A wide variety of source-detector separations have been utilized for this purpose. Here, we report a direct comparison of regression performance between two extremes of the reported range, 13 and 6 mm. Measurements of visual stimulation response (flickering radial checkerboard) were obtained from nine adults using a standard commercial source-detector grid with 13-mm diagonals, into which three extra detector fibers were placed to provide 6-mm channels at certain locations. When the NIRS recordings (17 total trials) were processed, the contrast-to-noise ratio was significantly higher with 6-mm regression channels than with 13 mm. The advantage could be due in part to the undesired sensing of brain activity by the 13-mm channels. We suggest that shorter distances be considered for optimal removal of superficial hemodynamics in NIRS signals from the adult brain.