Scalable mapping of myelin and neuron density in the human brain with micrometer resolution

Abstract

Optical coherence tomography (OCT) is an emerging 3D imaging technique that allows quantification of intrinsic optical properties such as scattering coefficient and back-scattering coefficient, and has proved useful in distinguishing delicate microstructures in the human brain. The origins of scattering in brain tissues are contributed by the myelin content, neuron size and density primarily; however, no quantitative relationships between them have been reported, which hampers the use of OCT in fundamental studies of architectonic areas in the human brain and the pathological evaluations of diseases. Here, we built a generalized linear model based on Mie scattering theory that quantitatively links tissue scattering to myelin content and neuron density in the human brain. We report a strong linear relationship between scattering coefficient and the myelin content that is retained across different regions of the brain. Neuronal cell body turns out to be a secondary contribution to the overall scattering. The optical property of OCT provides a label-free solution for quantifying volumetric myelin content and neuron cells in the human brain.

Figure 1

Histology and OCT optical property maps of the human somatosensory cortex. (a) Gallyas Silver stain shows contrast for myelin content. Red region: white matter. Green region: infragranular layers consist of layer IV, V and VI. Blue region: supragranular layers consist of layer I, II and III. Red arrow indicates a thin band of higher myelin content inside layer IV. (b) Optical density (OD) of Gallyas silver stain. The red arrow highlights the myelinated band inside the layer IV. (c,d) Nissl statin and COPA show contrast for cell bodies. The red region indicates the layer II and III with highest COPA value. The big black arrow highlights the high neuron density region. The small green arrows highlight the IV, V, VI layers within the infragranular layer with alternating contrasts. (e–g) Optical properties derived from the OCT images. (e) ${\mu }_s$ map. Small green arrows highlight the alternating contrasts in the infragranular layers similar to that in the COPA map, and the big red arrow indicates the myelinated band seen in Gallyas OD map. (f) ${\mu {}^{\prime }}_b$ map. Green arrow highlights the fibers with high intensity, possibly oriented within the imaging plane. (g) Ratio map of ${\mu {}^{\prime }}_b/{\mu }_s$ . Green arrow highlights the fibers with high intensity, similar to that in the ${\mu {}^{\prime }}_b$ map. The Blue arrow highlights the region with high signals in the supragranular layers. (h) OCT average intensity projection (AIP) image.

Figure 2

(a–e) Linear regression of ${\mu }_s$ and Gallyas OD for 5 different regions in human brain samples. Red dots: white matter data points. Blue dots: grey matter data points. The inset figure shows the Gallyas OD map of the corresponding sample. The Red circles in the inset figure represent the ROIs in the white matter. The blue, green and yellow circles represent the ROIs in different layers of the grey matter, for example, the green ROIs in (c) represent infragranular layers and the blue ROIs represent the supragranular layers. (a) Cerebellum, (b) hippocampus, (c) somatosensory cortex, (d) superior frontal cortex (SupFrontal), (e) middle temporal Brodmann area 21(BA21), (f) linear regression of all data points from 5 samples. The six panels have the same range on the X and Y axes for easier comparison.

Figure 3

Multivariate regression of ${\mu }_s$ vs Gallyas OD and COPA (black bars), compared against univariate regression where only Gallyas OD is considered (grey bars). (a) Gallyas OD slope $k_1$ of 5 brain regions resulted from univariate regression and multivariate regression. From left to right: cerebellum, hippocampus, somatosensory, SupFrontal and BA21. (b) Correlation coefficient with Gallyas OD in univariate and partial correlation coefficient (PCC) in multivariate regression. (c) R² of Pearson's correlation in univariate and multivariate regressions. (d) Normalized root mean square error (NRMSE) of univariate and multivariate regressions.

Figure 4

Evaluation of ${\mu }_s$ with COPA and remaining factors in multivariate regression. (a) COPA slope $k_2$ in the five brain regions. Regions with stars indicate significant $k_2$. (b) Partial correlation coefficient of COPA with respect to ${\mu }_s$. (c) Intercept b of the multivariate regression in the five brain regions.

Figure 5

(a) Average ${\mu }_s$, (b) average Gallyas OD in white matter and (c) average COPA in grey matter differentiating brain regions. The error bars represent the standard error from the ROIs.