3. A Bayesian approach for fitting integrate-and-fire model parameters to electrophysiological data obtained from adult-born neurons

TUOMO MÄKI-MARTTUNEN; ANTONIA MARÍN-BURGIN; LUCAS A. MONGIAT; EMILIO KROPFF; M BELEN PARDI; MARJA-LEENA LINNE; ALEJANDRO F. SCHINDER

Huerta Grande, Cordoba

Congreso; Reunión Anual Sociedad Argentina de Investigación en Neurociencia; 2012

Sociedad Argentina de Investigación en Neurociencia

We have recently shown that immature granule cells (GCs) of the adult dentate gyrus display a low threshold for activation rendered by their high excitation/inhibition balance [1,2].  To better understand the underlying intrinsic and synaptic mechanisms we now present and apply a data integrative method for fitting integrate-and-fire model parameters. We show how prior data on neuronal properties, such as membrane resistance and threshold potential for spiking [1], can be combined with data obtained from electrophysiological recordings in acute hippocampal slices to obtain a single-neuron model capable of predicting the output to a given input. The data consists of 1) spiking frequency measured against varying injected current and 2) spike trains of GCs observed as a response for entorhinal inputs [2]. For the latter, post-synaptic excitatory and inhibitory currents are collected and used for the parameter fitting and prediction. Our results show that the presented Bayesian method succeeds well in the integrative fitting of the data and that qualitatively correct predictions can be obtained using the model, despite the simplicity of integrate-and-fire dynamics.