Geometrical Modeling and Visualization of Pre- and Post-Synaptic Structures in Double-Labeled Confocal Images

3-D volume samples of selected dendritic segments are scanned by a confocal laser-scanning microscope to investigate the impact of alterations in the shape of dendritic spines for their synaptic functioning. Double-labeling techniques using green fluorescent protein-labeled neurons and synaptophysin-labeled presynaptic terminals allow us to capture images of dendrites and spines in one image channel and the presynaptic boutons in a second image channel. To detect spines and to reconstruct spine shape a parametric model is fitted in the noisy confocal image. Starting with the geometrical model of the spine head, voxels in the second image channel are labeled by a distance map to detect possible associated presynaptic terminals. This method enables a flexible visualization of geometrical models of spines and dendrites together with selected presynapses and an adjustable, distance-dependent quantification of boutons. These functional features are combined with morphological features from the geometrical model for every spine and establish a basis for advanced statistical analysis of the underlying learning processes in brain. The cellular events of changes in spine morphology can be studied in relation to functional properties.