| Dendritic morphology contributes greatly to synaptic integration and nervous system connectivity. This dissertation describes a series of computational models in which measurements taken from digitally reconstructed real neurons are stochastically resampled to create virtual dendritic trees. The resulting virtual structures can be statistically compared to the original data, while the parameter and algorithm choices allow the testing of different hypotheses of development. Morphological modeling is an iterative process in which even unsuccessful attempts can usefully point out gaps in our understanding. We found that an existing model in which branching probability was based on local branch diameter produces explosive growth when applied to CA1 apical trees, even though it had previously been shown to work for several other neuronal tree types. Based on these findings, we designed a new model in which all parameters, and not just branching probability, depend on local diameter. This choice limits explosive growth, but still produces CA1 apical trees that are too varied in size. Finally, in order to address whether these results are specific to CA1 apical trees or to the dependence on local diameter, we introduce a comparative approach in which a wide variety of dendritic tree types are modeled with a suite of closely related models. In these models, the sampling of parameters controlling branching, diameter changes, and elongation are dependent on diameter, path distance from the soma, branch order, or a combination of the three. By applying such model variations to a wide variety of tree morphologies, results can be directly compared, allowing one to uncover general as well as model and morphology specific trends. We find that apical and basal trees generally respond to the models in opposite ways, with the response patterns of non-pyramidal cells typically falling somewhere in between. This intriguing behavior suggests that pyramidal cells develop differently than other cell types, such as reacting differently in different cell layers, or competing for intracellular chemicals. |
