Research: Computational Modeling

Figure 1: In black, the firing rate of a model neuron from Area II of the goldfish hindbrain; the inset shows its morphology. A 15% reduction of dendritic diameters shifts the action potential to the left, increasing firing rate (Diam., green).

Figure 2: We predict compensatory perturbations of two active parameters (g-NaP, maximal conductance of persistent sodium channels, or RCa, the rate of buffering intracellular calcium) from the sensitivity of firing rate to these parameters (inset). The predicted perturbation of either parameter compensates almost exactly for the effect of the Diam. reduction in Figure 1.

Dendritic morphology, together with the active ion channels that drive membrane excitability, both contribute significantly to neuronal firing dynamics. However, the relative importance and interactions between these features remain poorly understood.

Our computational studies explore how morphology influences function in neurons subserving working memory in brain regions including the cortex and the brainstem. Constrained by the high resolution data collected within CNIC, our models provide insight into neuronal function at both the single cell and network levels.

We are developing computational techniques to characterize how neuronal output is influenced by interactions among the many parameters that define a model. We use these techniques to understand how realistic morphology participates in functional homeostasis, whereby a system maintains a certain level of output over time despite variability in its inputs.

In collaboration with colleagues at NYU Medical Center, one such project explores how morphology might contribute to the generation and maintenance of persistent neural activity. With colleagues within CNIC and at Boston University, we are also investigating how changes in active channels properties might compensate for morphologic changes observed in aging and neurodegenerative disorders.