Purkinje cell three-dimensional reconstruction and visualization system
Neuroscientists study the relationship between neuronal dendritic morphology and function by searching for relationships between behavior and morphology, as well as between morphology and disease. Research in this area is important because in humans, Purkinje cells (neurons) are affected in a variety of diseases. Developing new processing techniques for two-photon excited fluorescence microscopy and continuing research on Purkinje Cells can aid in slowing down the progression of these diseases and conditions. A new system that reconstructs and visualizes 3D Purkinje cells from two-photon microscopy images will be presented. The main components of the system are nonlinear diffusion filtering for denoising of each two-photon microscopy slice, increasing the image resolution of each Purkinje cell slice using cubic spline interpolation, DCT-II based global image enhancement of each slice, using a measure of enhancement to aid in selecting the optimal enhancement parameters as well as local pixel based image enhancement of each slice using our previously presented algorithm. 3D reconstruction and visualization of the Purkinje cell is accomplished using ImageJ. Finally, a 2D binary skeletonization algorithm will be used to create a simplified topological representation of a single Purkinje cell. Computer simulations in Matlab will illustrate the elucidated improvements over the original Purkinje cell images and other traditional image enhancement techniques. Specifically, the presented system aided in bringing out hidden dendrite branches and allowed for a clearer picture of the entire Purkinje cell, including the ability to visualize dendritic spines on the Purkinje cell skeleton. The presented system will ultimately allow for quantification of the Purkinje cell's morphology, including length, volume, radius and other features of dendrite branches and dendritic spines in the future.