Abstracts from Rambam Research Day, December 29, 2011.

Abstract

Introduction:At the cornerstone of neuroscience lies the attempt to correlate behavior and thought to brain states and activity patterns and transitions between such states. The underlying assumption to the studies we will present is that the most relevant level of description allowing such correlations is that of neuronal populations and interactions between such populations. The function of the nervous system, at the population or neuronal network level, can be studied in terms of three axes: Representation, Development, and Learning.Materials and Methods:The experimental system consists of large, random, cortical networks developing ex vivo coupled to multi-electrode-arrays. The networks are relatively free of predefined constraints and intervening variables, yet the electrophysiological, biochemical, and pharmacological properties of their neurons are mostly identical to neurons in vivo. The ex vivo developing model system enables extensive, multi-site sampling and manipulating of the relevant variables: electrical activity and the chemical milieu. The results presented here were collected by long-term recordings from these networks both at the level of the single, synaptically isolated neuron and at the level of the intact network under various environmental conditions and stimulation environments.Results:Basic biophysical aspects of response dynamics and their long-term fluctuations will be presented, at two different levels of organization: the single, isolated neuron and the intact assembly. We show that individual neurons and networks display a very complex, history dependent response patterns that pose constraints on possible representation schemes. Moreover, we will show the feasibility of such representation schemes and implications of their usage. Representation schemes at the network level can be viewed as different “codes,” with a gross division between rate and time-based codes. Finally, we will show some preliminary results regarding such fluctuations in the response dynamics at the level of intact organisms.Conclusions:The neuronal system, at any given level of organization is not only very “noisy” in its response dynamics but moreover—it is inherently non-stationary. This non-stationarity reflects internal processes, some of which are activity dependent. Our results describe the dynamics of the fluctuations in the responses to stimulation; at the network level, we show how these dynamics affect possible representation schemes with interplay between rate-based and time-based schemes. We show the advantages to either scheme depending on the stimulus properties.