HRfunc: a tool for modeling hemodynamic response variability in fNIRS

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Neural activation in functional near-infrared spectroscopy (fNIRS) signals is obscured by the hemodynamic response function (HRF).
• Variability in HRF complicates accurate recovery of neural activity from fNIRS data.
• Improved modeling of HRF is crucial for advancing fNIRS analysis.

Purpose of the Study:

• To introduce HRfunc, a novel Python tool for estimating local HRF distributions and neural activity from fNIRS data via deconvolution.
• To develop an efficient data structure (tree and hash table hybrid) for spatial and contextual identification of relevant HRFs.
• To facilitate the estimation of event-based HRFs and neural activity in fNIRS studies.

Main Methods:

• The HRfunc tool was validated using two analyses on hemoglobin and estimated neural activity.
• A general linear model (GLM) analysis was performed on a child executive function task dataset (n=79).
• A neural synchrony analysis assessed wavelet coherence between child-parent dyads (92 dyads).

Main Results:

• Estimated HRFs generally exhibited a canonical shape.
• Analysis revealed increased kurtosis and stable skew in estimated neural activity, with a decrease in signal-to-noise ratio.
• Emergent neural synchrony lateralization effects and consistent GLM outcomes were observed.

Conclusions:

• The HRfunc tool is effective for estimating event-based HRFs and neural activity in fNIRS research.
• The study supports the utility of HRfunc for enhancing the accuracy of fNIRS data interpretation.
• Establishing a collective HRF database will provide valuable resources across diverse research contexts.