Evaluating and improving the quality of time-dependent, diffuse reflectance spectroscopic signals measured from in vivo brain during craniotomy

Distinguishing motion artifacts during optical fiber-based in-vivo hemodynamics recordings from brain regions of freely moving rodents

Significance

Motion artifacts in the signals recorded during optical fiber-based measurements can lead to misinterpretation of data. In this work, we address this problem during in-vivo rodent experiments and develop a motion artifacts correction (MAC) algorithm for single-fiber system (SFS) hemodynamics measurements from the brains of rodents.

Aim

(i) To distinguish the effect of motion artifacts in the SFS signals. (ii) Develop a MAC algorithm by combining information from the experiments and simulations and validate it.

Approach

Monte-Carlo (MC) simulations were performed across 450 to 790 nm to identify wavelengths where the reflectance is least sensitive to blood absorption-based changes. This wavelength region is then used to develop a quantitative metric to measure motion artifacts, termed the dissimilarity metric (DM). We used MC simulations to mimic artifacts seen during experiments. Further, we developed a mathematical model describing light intensity at various optical interfaces. Finally, an MAC algorithm was formulated and validated using simulation and experimental data.

Results

We found that the 670 to 680 nm wavelength region is relatively less sensitive to blood absorption. The standard deviation of DM (σ D M ) can measure the relative magnitude of motion artifacts in the SFS signals. The artifacts cause rapid shifts in the reflectance data that can be modeled as transmission changes in the optical lightpath. The changes observed during the experiment were found to be in agreement to those obtained from MC simulations. The mathematical model developed to model transmission changes to represent motion artifacts was extended to an MAC algorithm. The MAC algorithm was validated using simulations and experimental data.

Conclusions

We distinguished motion artifacts from SFS signals during in vivo hemodynamic monitoring experiments. From simulation and experimental data, we showed that motion artifacts can be modeled as transmission changes. The developed MAC algorithm was shown to minimize artifactual variations in both simulation and experimental data.

Quantitative assessment of hemodynamic and structural characteristics of in vivo brain tissue using total diffuse reflectance spectrum measured in a non-contact fashion

Here we present a new methodology that investigates the intrinsic structural and hemodynamic characteristics of in vivo brain tissue, in a non-contact fashion, and can be easily incorporated in an intra-operative environment. Within this methodology, relative total diffuse reflectance spectra (R_TD(λ)) were acquired from targets using a hybrid spectroscopy imaging system. A spectral interpretation algorithm was subsequently applied to R_TD(λ) to retrieve optical properties related to the compositional and structural characteristics of each target. Estimation errors of the proposed methodology were computationally evaluated using a Monte Carlo simulation model for photon migration under various conditions. It was discovered that this new methodology could handle moderate noise and achieve very high accuracy, but only if the refractive index of the target is known. The accuracy of the technique was also validated using a series of tissue phantom studies, and consistent and accurate estimates of μ_s'(λ)/μ_a(λ) were obtained from all the phantoms tested. Finally, a small-scale animal study was conducted to demonstrate the clinical utility of the reported method, wherein a forepaw stimulation model was utilized to induce transient hemodynamic responses in somatosensory cortices. With this approach, significant stimulation-related changes (p < 0.001) in cortical hemodynamic and structural characteristics were successfully measured.