Towards a representative reference for MRI-based human axon radius assessment using light microscopy

Non-invasive assessment of axon radii via MRI bears great potential for clinical and neuroscience research as it is a main determinant of the neuronal conduction velocity. However, there is a lack of representative histological reference data at the scale of the cross-section of MRI voxels for validating the MRI-visible, effective radius (r(eff)). Because the current gold standard stems from neuroanatomical studies designed to estimate the bulk-determined arithmetic mean radius (r(arith)) on small ensembles of axons, it is unsuited to estimate the tail-weighted reff. We propose CNN-based segmentation on high-resolution, large-scale light microscopy (lsLM) data to generate a representative reference for r(eff). In a human corpus callosum, we assessed estimation accuracy and bias of r(arith) and r(eff). Furthermore, we investigated whether mapping anatomy-related variation of r(arith) and r(eff) is confounded by low-frequency variation of the image intensity, e.g., due to staining heterogeneity. Finally, we analyzed the error due to outstandingly large axons in r(eff). Compared to rarith, r(eff) was estimated with higher accuracy (maximum normalized-root-mean-square-error of r(eff): 8.5 %; r(arith): 19.5 %) and lower bias (maximum absolute normalized-mean-bias-error of r(eff): 4.8 %; r(arith): 13.4 %). While r(arith) was confounded by variation of the image intensity, variation of r(eff) seemed anatomy-related. The largest axons contributed between 0.8 % and 2.9 % to r(eff). In conclusion, the proposed method is a step towards representatively estimating r(eff) at MRI voxel resolution. Further investigations are required to assess generalization to other brains and brain areas with different axon radii distributions.

Automated summary

Non-invasive assessment of axon radii via MRI is important for clinical and neuroscience research, but lacks representative histological reference data. This study proposes a method using CNN-based segmentation to generate a representative reference for the effective radius (r(eff)) of axons. The results show that r(eff) estimation is more accurate and less biased compared to the traditional method of estimating the arithmetic mean radius (r(arith)).