Optimizing short-channel regression in fNIRS: an empirical evaluation with ecological audiovisual stimuli
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Significance
Functional Near-Infrared Spectroscopy (fNIRS) is increasingly favored for its portability and suitability for ecological paradigms, yet methodological standardization remains a challenge regarding the optimal use of short-separation channels (SC) to remove systemic physiological noise.
Aim
We aim to evaluate and compare methods of SC regression as implemented in the most widely used fNIRS analysis toolboxes, with the goal of reaching a consensus on best practices for incorporating SC into generalized linear model (GLM)-based analyses. Specifically, we compared ten SC regression strategies addressing SC selectivity, dimensionality reduction strategies, and SC availability.
Approach
16 healthy adults passively listened and watched ecological auditory, visual, and audiovisual stimuli while occipital and bilateral auditory cortices were recorded.
Results
Oxygenated hemoglobin signals (HbO) processed without SC regression produced uninterpretable results in the context of the present study. Non-selective SC regression methods that pooled all available SC signals consistently outperformed anatomically or functionally restricted approaches. Orthogonalization further enhanced performance by reducing redundancy and capturing shared systemic variance, improving detection of stimulus-specific cortical responses and contrast sensitivity. Deoxygenated hemoglobin signals (HbR), while less sensitive to systemic artifacts than HbO, benefited most, similarly to HbO, from the pooled, orthogonalized SC signal approach.
Conclusion
Overall, our findings highlight the essential role of SC regression in recovering physiologically meaningful signals in fNIRS and recommend including all available SC channels within the GLM, coupled with orthogonalization techniques, as a generalizable best practice for denoising across hardware configurations.
