Neuroanatomical underpinning of diffusion kurtosis measurements in the cerebral cortex of healthy macaque brains

Abstract

Purpose: To investigate the neuroanatomical underpinning of healthy macaque brain cortical microstructure measured by diffusion kurtosis imaging (DKI), which characterizes non-Gaussian water diffusion.

Methods: High-resolution DKI was acquired from 6 postmortem macaque brains. Neurofilament density (ND) was quantified based on structure tensor from neurofilament histological images of a different macaque brain sample. After alignment of DKI-derived mean kurtosis (MK) maps to the histological images, MK and histology-based ND were measured at corresponding regions of interests characterized by distinguished cortical MK values in the prefrontal/precentral-postcentral and temporal cortices. Pearson correlation was performed to test significant correlation between these cortical MK and ND measurements.

Results: Heterogeneity of cortical MK across different cortical regions was revealed, with significantly and consistently higher MK measurements in the prefrontal/precentral-postcentral cortex compared to those in the temporal cortex across all six scanned macaque brains. Corresponding higher ND measurements in the prefrontal/precentral-postcentral cortex than in the temporal cortex were also found. The heterogeneity of cortical MK is associated with heterogeneity of histology-based ND measurements, with significant correlation between cortical MK and corresponding ND measurements (P < .005).

Conclusion: These findings suggested that DKI-derived MK can potentially be an effective noninvasive biomarker quantifying underlying neuroanatomical complexity inside the cerebral cortical mantle for clinical and neuroscientific research.

Keywords: biomarker; cerebral cortex; diffusion MRI; diffusion kurtosis imaging; histology; microstructure.

Figure 1:

Flowchart of evaluating relationship between mean kurtosis (MK) and histology in the cerebral cortex. The flow chart consists of three technical components, estimation of MK maps in the left panel, calculation of neurofilament density (ND) in the right panel and systematic and quantitative comparison between MK and ND map. For kurtosis estimation in the left panel, the procedures included from top to bottom acquisition of diffusion weighted images (DWIs) with multiple b-value, linear registration of DWIs to the b0 image, constrained linear kurtosis fitting with registered DWI volumes to generate kurtosis maps including MK maps and upsampling of MK maps. For ND estimation in the right panel, the procedures included contrast inversion, calculation of anisotropy index (AI) for each pixel with structure tensor analysis, thresholding, blocking histological image at the similar resolution to upsampled diffusion MRI, and calculation of ND with the equation.

Figure 2:

High resolution diffusion tensor imaging (DTI) and diffusion kurtosis imaging (DKI) maps from a representative macaque brain (sample #2). First two rows show DTI parameter maps, including fractional anisotropy (FA), color-encoded, mean diffusivity (MD), axial diffusivity (AD) and radial diffusivity (RD) map, along with the averaged diffusion-weighted image (aDWI). Third row shows DKI parameter maps, including axial kurtosis (AK), radial kurtosis (RK), and mean kurtosis (MK) maps. MK map shown in the yellow box was used for comparison with histology.

Figure 3:

Coronal MK maps from anterior to posterior from a representative macaque brain (sample #1). Significant MK difference is clear between the temporal cortex (blue regions pointed by white arrows) and prefrontal/precentral-postcentral cortex (yellow-green regions pointed by white arrows). The color bar encodes MK values.

Figure 4:

A representative MK map in a corresponding coronal slice for each of all six postmortem macaque brains included in this study. Relatively higher MK at the prefrontal/precentral-postcentral cortex (indicated by red arrow) and relatively lower MK at the temporal cortex (indicated by yellow arrow) was reproducible across all 6 scanned macaque brains.

Figure 5:

Qualitative comparison between MK and ND map in four different coronal slices 1, 2, 3 and 4 from a representative macaque brain (sample #2). In each panel, top left, top right, bottom left and bottom right shows aDWI, corresponding histology image, MK map and ND map, respectively. aDWI is shown as an anatomical reference. High MK value in the prefrontal/precentral-postcentral cortical regions is consistent to high ND value at corresponding locations indicated by pink arrow, while low MK value in the temporal cortical regions is consistent to low ND value at corresponding locations indicated by yellow arrows. The color bar and grayscale bar encode MK and ND values, respectively.

Figure 6:

a. Reproducibly significantly higher (P<0.00005, Bonferroni corrected) cortical MK measurements at regions of interest (ROI) in the prefrontal/precentral-postcentral cortex than cortical MK measurements at ROI in the temporal cortex across all 6 macaque brain samples. MK measurements from ROI in the prefrontal/precentral-postcentral and temporal cortex were plotted for each sample as blue and orange boxplots, respectively. To match the thickness of MK coronal plane, ND measurements from the corresponding ROI on three histology slices were plotted alongside. Significant higher (P<0.00001) ND measurements in corresponding prefrontal/precentral-postcentral cortex than ND measurements in corresponding temporal cortex, shown on the right, were consistent with MK measurements. b. Cortical FA measurements at the same ROIs in the prefrontal/precentral-postcentral and temporal cortex across all 6 macaque brain samples showed no consistent differences between FA values in different cortical ROIs.

Figure 7:

Significant linear correlation between MK and ND measurements at ROIs of a representative macaque brain (sample #1). a. MK map of a representative coronal slice. b. Corresponding three neurofilament-stained histological slices matching thickness of the MK map. c-d. Enlarged red and green ROIs show denser staining in the red box than in the green box. e. Three ND maps calculated from the three histological slices in panel b. f. With MK and ND measurements at the 8 corresponding ROIs demonstrated in panels a (black circles), b (black circles) and e (white circles), statistically significant (P<0.005) linear correlation was found between MK and ND.