Optimal hemodynamic response model for functional near-infrared spectroscopy.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Functional near-infrared spectroscopy (fNIRS) is a non-invasive brain imaging technique measuring cortical activity via near-infrared light.
• Cortical hemodynamic responses (HR) exhibit variability across brain regions and trial repetitions, posing challenges for accurate functional mapping.
• Existing models often fail to capture these dynamic variations in hemodynamic responses.

Purpose of the Study:

• To develop and validate a robust hemodynamic response (HR) model capable of estimating variations in cortical activity measured by fNIRS.
• To improve the accuracy of functional brain maps by accounting for inter-subject variability in HR and physiological noise.

Main Methods:

• A novel HR model was developed, incorporating a canonical hemodynamic response function (cHRF) modeled by two Gamma functions with adjustable parameters.
• The model accounts for HR, baseline, and unknown physiological noise components, formulated as an objective function with constraints on 12 parameters.
• An iterative optimization algorithm was employed to estimate unknown parameters, and statistical analysis (t-value, p-value) verified significance.

Main Results:

• The developed HR model successfully estimated inter-subject variations in hemodynamic responses and physiological noise.
• Accuracy was validated using 10 real and 15 simulated datasets, demonstrating the model's reliability.
• Analysis of finger-tapping tasks in healthy subjects revealed statistically significant estimations of activity strength parameters.

Conclusions:

• The proposed HR model enhances the precision of functional brain mapping with fNIRS by capturing individual response variations.
• This advancement offers improved cortical functional maps, crucial for understanding brain activity in diverse neurological contexts.
• The method provides a statistically robust approach for analyzing fNIRS data, paving the way for more refined neuroimaging studies.