The influence of extra-cerebral vasculature on the efficacy of the short-separation regression approach applied to fNIRS data analysis

Functional near-infrared spectroscopy (fNIRS) is a non-invasive, portable, and inexpensive optical neuroimaging technique with respectable spatial and temporal resolution. fNIRS is, however, susceptible to extra-cerebral physiological noise, potentially compromising its sensitivity to detect task-related brain activation. Previous studies have addressed this issue by correcting fNIRS signals with additional information recorded exclusively from extra-cerebral regions using short-distance channels (SDCs). This method, termed short-separation regression (SSR), can improve fNIRS-signal quality and the sensitivity to detect task-related brain activation. However, it is unclear whether the efficacy of SSR depends on factors such as the presence of blood vessels in the channel’s vicinity. Here, we combined anatomical, functional and angiographic magnetic resonance imaging data with continuous-wave fNIRS data to quantify the impact of SSR on the fNIRS-signal quality and sensitivity and investigated how vascular proximity/density contributes to SSR efficacy. Our investigation verifies that SSR improves fNIRS-signal quality and the sensitivity to detect task-related brain activation considerably and shows that signals obtained via SDCs are affected by close vascular structures. The present study extends our understanding of the relationship between vasculature features, the fNIRS signal quality, and methods (e.g., SSR) designed to increase fNIRS applicability to accurately detect brain activity.