The morphology of dendrites shapes synaptic connectivity and therefore neuronal function. To better understand the principles governing the formation of dendritic arbors, we study two subclasses of sensory neurons in the embryo and larva of the fruit fly Drosophila. These neurons, known as class I and class IV neurons, have markedly differentfunctions and morphologies. Class I neurons, involved in proprioception, adopt a comb-like geometry that covers onlypart of each hemisegment in the embryo and then the larva. By contrast, class IV neurons, involved in nociception, deploy a dense mesh innervating all segments. In both cases, the dendritic tree is built from three elementarybehaviors: (i) side-branching, where new sprouts emerge from existing branches; (ii) tip dynamics, characterized byalternating phases of extension and retraction; and (iii) tip retraction upon encountering another branch of the sameneuron.
To identify the parameters that differ between the two neuronal classes and ultimately determine their respective forms, we combine high-resolution in vivo imaging with automated image-analysis tools to quantify branch-formationdynamics in wild-type and perturbed conditions. To predict the impact of these parameters on final neuronal morphology, we perform in silico simulations of growth for both neuron classes using computational models. We show that models based solely on independent, short-duration increments fail to reproduce the observedmorphologies. In contrast, introducing long-term correlations between increments, modulated by microtubules, proves essential to account for branch growth in both neuron classes. These results highlight the central role of microtubule regulation in shaping dendritic architecture and propose a quantitative framework linking cytoskeletaldynamics to neuronal form.
The thesis is organized around four themes:
- Development of image-processing tools: segmentation, branch-tracking, and neuronal-reconstructionalgorithms.
- Data-driven growth modeling: comparison of several branch-growth models calibrated on quantitativedynamical measurements.
- Cytoskeletal dynamics: experimental and quantitative analyses of microtubule- and actin-dependent processesthat govern short- and long-term branch dynamics.
- Theoretical analysis: a mathematical model describing the morphology of class I neurons as an emergentconsequence of self-repulsion.