Functional near-infrared spectroscopy (fNIRS) monitors brain hemodynamics by sending NIR light into the head from light sources placed on the scalp and measuring the light that scatters back. While this approach is noninvasive, its accuracy can be affected by signal contamination from blood flow in the scalp and skull.
Using a special source-detector geometry, researchers at Tufts University measured how light travels through a layered-tissue model of the head. This approach allowed the researchers to isolate brain-specific signals without the need for large datasets for tomographic reconstructions.
Traditional fNIRS measurements often use a single distance between a light source and a detector. While easy to implement, this homogeneous setup is highly sensitive to blood flow changes in the scalp and skull. More advanced methods can separate surface and brain signals, but frequently require multiple light sources and detectors, dense sensor arrays, and heavy computation.
The newly developed approach uses a dual-slope source detector configuration with two light sources and two detectors placed at different distances on the scalp. This arrangement produces several measurements, each with a different balance of sensitivity to superficial and deep tissue.
The dual slope approach to measurement reduced the influence of signals coming from surface tissue, and enhanced sensitivity to signals coming from cerebral blood flow.
To translate the measurements into meaningful estimates of brain activity, the researchers developed tissue models for analyzing the data. Instead of creating a homogeneous (single layer) model, they developed a two-layer model with one layer of superficial tissue and another of brain tissue, and a three layer model with an additional middle layer to denote the cerebrospinal fluid that surrounds the brain.
The team used Monte Carlo simulations to track how light propagated through the layered structures. It generated dual-slope frequency-domain functional near-infrared spectroscopy (DS FD-NIRS) data from models with a range of tissue thicknesses and optical properties to reflect biological variability. The team characterized how each DS FD-NIRS data type responded to simulated functional activation in the deepest layer, reproducing the main qualitative features of in vivo functional data.
To validate the simulations, the researchers collected in vivo data from healthy volunteers with different extracerebral tissue thicknesses. During the experiment, participants viewed a visual stimulus designed to activate the occipital cortex. The researchers measured changes in light intensity and phase during the experiment using DS FD-NIRS. They also used ultrasound imaging to estimate each participant’s scalp and skull thickness, providing an independent measure of superficial anatomy.
The team compared the simulation results to data collected in vivo from the volunteers. Its goal was to identify a two- or three-layered medium that could reproduce, at least qualitatively, the behavior of DS FD-NIRS data collected in vivo and that could serve as a basis for a more accurate determination of cerebral hemodynamics, improving upon the oversimplistic homogeneous tissue model.
The results clearly favored the three layer model. Only when the model included a low scattering, low absorbing layer representing cerebrospinal fluid did the simulated data reproduce the main qualitative features of the human measurements.
The researchers found that cerebrospinal fluid changed the relative sensitivity of different fNIRS measurements to brain tissue, playing an outsized role in how light traveled through the head.
In the three layer model, the differences between subjects could be explained primarily by the variations in scalp and skull thickness that aligned with known anatomy. The two layer model could match the experimental data only by assuming large, unlikely differences in tissue scattering properties between individuals. This finding suggests that explicitly representing cerebrospinal fluid is important for the realistic modeling of light transport in the head.
The three layer model also enabled the researchers to estimate how much of the measured signal came from superficial tissue compared to brain tissue. When they analyzed the visual stimulation data, they found that the detected responses were dominated by cerebral changes, with minimal contribution from the scalp.
This result is consistent with prior work showing that visual tasks produce strong, localized brain responses with relatively small systemic effects at the surface. More importantly, it demonstrates that a modestly complex model can help distinguish these contributions using standard fNIRS measurements.
Although a two- or three-layer model is a simplification of the actual head anatomy, if the model can reproduce the main features of in vivo data, it can serve as a tool for taking robust, noninvasive measurements of cerebral hemodynamics. The study shows that moving from a homogeneous model to a three layer model could provide a significant improvement in measurement results without greatly increasing complexity.
The three-layer model reproduced in-vivo data collected with the two-source, two-detector configuration without the need for large sensor arrays or MRI scans. This could make the method useful for settings where portability and ease of use are required.
The work could enable more reliable, noninvasive brain monitoring in clinical environments and real-life settings. By clarifying how light interacts with layered head tissues, the study helps bring medical imaging closer to the goal of using fNIRS for accessible, accurate measurement of brain activity.
More Info: Visit: biophotonicsresearch.com Nominate Link: https://biophotonicsresearch.com/award-nomination/?ecategory=Awards&rcategory=Awardee Registration Link: https://biophotonicsresearch.com/award-registration/
No comments:
Post a Comment