Inferring deep-brain activity from cortical activity using functional near-infrared spectroscopy.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Cognitive Science

Background:

• Functional near-infrared spectroscopy (fNIRS) offers non-invasive brain activity measurement with high temporal resolution.
• fNIRS is limited in detecting hemodynamic changes in deep-brain regions compared to fMRI.
• Inferring deep-brain activity from cortical signals could enhance fNIRS utility.

Purpose of the Study:

• To develop and validate a computational method for inferring deep-brain activity using fNIRS cortical data.
• To assess the feasibility of predicting deep-brain activity from surface fNIRS measurements.
• To overcome the spatial resolution limitations of fNIRS for deep-brain structures.

Main Methods:

• Simultaneous fNIRS and fMRI data acquisition from 17 participants during cognitive tasks.
• Application of a support vector regression (SVR) algorithm to predict deep-brain activity from fNIRS cortical signals.
• Comparison of predicted deep-brain activity with fMRI-measured activity using Pearson's correlation.

Main Results:

• The SVR model successfully predicted deep-brain activity using fNIRS cortical data.
• High prediction accuracy was achieved, with an average correlation coefficient of 0.67 for all deep-brain regions.
• Top predictions using fNIRS achieved an accuracy of 0.7, demonstrating the method's potential.

Conclusions:

• This study presents the first investigation into inferring deep-brain activity from cortical fNIRS signals.
• The developed computational method shows promise for extending fNIRS applications in cognitive and clinical neuroscience.
• This approach could significantly enhance the capabilities of fNIRS for brain research.