Abstract

The structure of neuronal dendrites and dendritic spines has been realized to reveal many important aspects of cognitive functions. We present an MDL (Minimum Description Length) principle based algorithm to identify the morphologic structure of dendritic branching and analyze spine density and distribution. An algorithm utilizing the gradient vector field to locate the skeletons of the tubular objects is applied after an anisotropic diffusion process. In order to generate a graph structure from the 3D skeleton, a minimum spanning tree algorithm based on density weighted edges (DWMST) is employed. MDL models are created and optimized under consideration of the tree branches in dendritic structure. We have the prior probabilities of the dendritic spines, which should be involved in the descriptive language. In the MDL strategy, the spine features are investigated and considered as the observed data for any given models. We choose the models that minimizes the number of bits required to describe the features given the model plus the number of bits required to describe the model. Many experimental results show the efficiency of our algorithms on 3D fluorescence images.

Notes

Poster presented at the 2007 Thrust R2B Localized Probing and Mosaicing Methods Conference

Keywords

MDL Principle, Extraction, spines, DWMST

Subject Categories

Dendrites--Analysis

Disciplines

Biomedical Engineering and Bioengineering

Publisher

Bernard M. Gordon Center for Subsurface Sensing and Imaging Systems (Gordon-CenSSIS)

Publication Date

2007

Rights Holder

Bernard M. Gordon Center for Subsurface Sensing and Imaging Systems (Gordon-CenSSIS)


View Full Text

Click button above to open, or right-click to save.

Share

COinS