Sponsors of the 2 nd Annual EPFL Life

Number Abstract Title Authors Authors 2nd Annual EPFL Life Science Symposium Abstract Number Abstract TitleNumber Abstract Title 02 Dynamic I-V curves are reliable predictors of naturalistic pyramidal-cell voltage traces Badel, Laurent 1 | Lefort, Sandrine 2 | Petersen, Carl C.H. 2 | Gerstner, Wulfram 1 Richardson, Magnus J.E. 1-3. 1 Laboratoty of Computational Neuroscience, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland | 2 Laboratoty of Sensory Processing, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland | 3 Warwick Systems Biology Centre, University of Warwick, Coventry, UK. Neuronal response properties are typically probed by time-intensive intracellular measurements of current-voltage (I-V) relationships by step-change stimuli, in either current-clamp or voltage-clamp mode. Here we demonstrate a novel I-V curve which is measured while the neuronal voltage is fluctuating and producing spikes under more naturalistic stimulation protocols. This dynamic I-V curve requires only a few 10s of seconds of experimental time and so lends itself readily to the rapid classification of cell type and to the generation of reduced analytical models. We apply this technique to layer-5, cortical pyramidal cells and show that their dynamic I-V curve is invariant across all cells measured. The shape of the curve is very well fitted by a linear component, characterizing the passive properties of the cell, and an exponential component characterizing the onset of the sodium spike-generating current providing experimental evidence for a recently proposed theoretical model [1]. An important and novel property of the dynamic I-V curve is that it can be used to dissect the changing neuronal response function after a spike. This property is used to quantify the post-spike refractory properties of pyramidal cells which are characterized by transient changes in the parameters of the I-V curve, particularly in the level of the spike threshold. These observations are cast in a simple mathematical model which is used to predict the response of pyramidal cells to novel stimuli. The predictions of the resulting model are in excellent agreement with experiment and close to the intrinsic neuronal reproducibility to repeated stimuli. Task-modulated interactions in primary visual cortex induced by perceptual learning Schäfer, Roland 1 | Vasilaki, Eleni 2 | Senn, Walter 1. 1 University of Bern, Switzerland | 2 Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland The neuronal processing in primary visual cortex (V1) strongly depends on the perceptual task subjects perform based on a visual stimulus. Depending on the task, the same visual stimulus can evoke entirely different interactions among V1 neurons. For instance, the mutual suppression of edge-detecting V1 neurons in a side-by-side configuration can turn into strong facilitation when switching from a fixation task to a bisection task (Crist et al., 2001). Yet, bisection training does not change the classical receptive field properties of V1 neurons, even when these are determined while the task is performed. These seemingly contradictory observations place strong constraints on the architecture and processing in the primary visual cortex. We suggest a model of perceptual learning which naturally explains the selective modulation of V1 activities. The core idea is that perceptual learning develops a task-specific top-down modulation of intrinsic V1 interactions. Although these top-down signals vary with the task, they operate through common mechanisms on V1: gain modulation of pyramidal neurons and modulation of recurrent inhibition. During the fixation task top-down input is weak, recurrent inhibition dominates, and lateral suppression emerges. During the bisection task top-down signals strengthen the gain of excitatory neurons and lateral suppression is turned into excitation. Nevertheless, receptive field properties of the involved V1 neurons change only weakly. The proposed model is further compatible with recent psychophysical results on bisection learning with different bisection widths (Otto et al., 2006). 01 Affiliations Affiliations