New techniques for imaging, digitization and analysis of three-dimensional neural morphology on multiple scales

Abstract

Cognitive impairment in normal aging and neurodegenerative diseases is accompanied by altered morphologies on multiple scales. Understanding of the role of these structural changes in producing functional deficits in brain aging and neuropsychiatric disorders requires accurate three-dimensional representations of neuronal morphology, and realistic biophysical modeling that can directly relate structural changes to altered neuronal firing patterns. To date however, tools capable of resolving, digitizing and analyzing neuronal morphology on both local and global scales, and with sufficient throughput and automation, have been lacking. The precision of existing image analysis-based morphometric tools is restricted at the finest scales, where resolution of fine dendritic features and spine geometry is limited by the skeletonization methods used, and by quantization errors arising from insufficient imaging resolution. We are developing techniques for imaging, reconstruction and analysis of neuronal morphology that capture both local and global structural variation. To minimize quantization error and evaluate more precisely the fine geometry of dendrites and spines, we introduce a new shape analysis technique, the Rayburst sampling algorithm that uses the original grayscale data rather than the segmented images for precise, continuous radius estimation, and multidirectional radius sampling to represent non-circular branch cross-sections and anisotropic structures such as dendritic spine heads, with greater accuracy. We apply the Rayburst technique to 3D neuronal shape analysis at different scales. We reconstruct and digitize entire neurons from stacks of laser-scanning microscopy images, as well as globally complex structures such as multineuron networks and microvascular networks. We also introduce imaging techniques necessary to recover detailed information on three-dimensional mass distribution and surface roughness of amyloid beta plaques from human Alzheimer’s disease patients and from the Tg2576 mouse that expresses the “Swedish” mutation of the amyloid precursor protein. By providing true three-dimensional morphometry of complex histologic structures on multiple scales, the tools described in this report will enable multiscale biophysical modeling studies capable of testing potential mechanisms by which altered dendritic structure, spine geometry and network branching patterns that occur in normal aging and in many brain disorders, determine deficits of functions such as working memory and cognition.