Diese Anwendung erfordert Java-Skript.

Bitte aktivieren Sie Java-Script in den Browser-Einstellungen.
Brains for Brains Award
Bisio, Martha; Bourdoukan, Ralph; Huber-Brösamle, Andrea (2013)
clipboard
mla
clipboard		 Bisio M., et al. "Brains for Brains Award.", timms video, Universität Tübingen (2013): https://timms.uni-tuebingen.de:443/tp/UT_20130927_008_bestcon_0001. Accessed 23 Nov 2024.
apa
clipboard		 Bisio, M., Bourdoukan, R. & Huber-Brösamle, A. (2013). Brains for Brains Award. timms video: Universität Tübingen. Retrieved November 23, 2024 from the World Wide Web https://timms.uni-tuebingen.de:443/tp/UT_20130927_008_bestcon_0001
harvard
clipboard		 Bisio, M., Bourdoukan, R. and Huber-Brösamle, A. (2013). Brains for Brains Award [Online video]. 27 September. Available at: https://timms.uni-tuebingen.de:443/tp/UT_20130927_008_bestcon_0001 (Accessed: 23 November 2024).
file download bibtex   endnote
Information
title: Brains for Brains Award
alt. title: Bernstein Conference 2013: Award Session and Outlook
creators: Bisio, Martha (author), Bourdoukan, Ralph (author), Huber-Brösamle, Andrea (annotator)
subjects: Bernstein Conference, Computational Neuroscience, Award Session, Brains for Brains Award, Electrophysiological Activity, Confined Neuronal Populations, Development, Spiking Networks, Balance, Excitation, Inhibition, Representation, Martha Bisio, Ralph Bourdoukan
description: Bernstein Conference 2013, 24. bis 27. September 2013
abstract: Martha Bisio: The development of in vitro models of patterned neuronal networks is of significant interest in the neuroscientific community and requires the convergence of electrophysiological studies with micro/nano-fabrication of adapted devices. Neuronal assemblies coupled to Micro Electrode Arrays (MEAs) constitute a peculiar neurobiological model for investigating the strategies employed by the brain to represent and process information. Considering the multitude of connections arising in un-patterned neuronal cultures, the constraint imposed to neurite outgrowth along specific pathways ensures a considerable control over network complexity. Indeed, in vitro models should be intrinsically modular, in order to provide a 'simplified' but plausible representation of in vivo nervous systems. In order to facilitate the realization of reproducible patterned assemblies, a technique to induce self-organization of networks into two clusters connected by two macro-channels (50-100 µm width), on commercially available MEAs, has been developed. The spontaneous activity has been recorded during the development of patterned networks, from a few days up to 8 weeks. Unlike in vivo networks, in which multiple activation pathways impinge on any recorded region, these partially confined networks can be studied in a controlled environment. Moreover these networks exhibit synchronized events at later stages of the development with respect to homogeneous cultures. The obtained results constitute important evidence that engineered neuronal networks are a powerful platform to systematically approach questions related to the dynamics of neuronal assemblies. Moreover, this engineered system can be easily interfaced to artificial artifacts in order to investigate coding properties towards the final goal of integrating brains and machines. Ralph Bourdoukan: Cortical activity is typically irregular, asynchronous, and Poisson-like. This variability seems to be predominantly caused by a balance of excitatory and inhibitory neural input (Haider, B. et al., 2006). However, the learning mechanisms that develop this balance and the functional purpose of this balance are poorly understood. Here we show that a Hebbian plasticity rule drives a network of integrate-and-fire neurons into the balanced regime while simultaneously developing an optimal spike-based code. The remarkable coincidence of balance and optimality in our model occurs when synaptic plasticity is proportional to the product of the postsynaptic membrane potential and presynaptic firing rate. This plasticity rule acts to minimise the magnitude of neural membrane potential fluctuations. Balance develops because, without it, membrane potential fluctuations are too large. Meanwhile, an optimal representation develops because membrane potentials correspond to representation errors for a signal encoded by the network (Boerlin, M. and Deneve, S., 2011). This signal may be a sensory signal or the result of some network computation. It can be extracted from the network spike trains with a fixed linear decoder (a summation of postsynaptic potentials), with a precision on the order of 1/N, where N is the number of neurons. This is much more precise than a typical rate model. Our work suggests that several of the features measured in cortical networks, such as the high trial-to-trial variability, the balance between excitation and inhibition, and spike-time-dependent plasticity are all signatures of an optimal spike-based representation.
publisher: ZDV Universität Tübingen
contributors: Bernstein Center for Computational Neuroscience Tübingen (BCCN) (producer), Bethge, Matthias (organizer), Wichmann, Felix (organizer), Lam, Judith (organizer), Macke, Jakob (organizer)
creation date: 2013-09-27
dc type: image
localtype: video
identifier: UT_20130927_008_bestcon_0001
language: eng
rights: Url: https://timmsstatic.uni-tuebingen.de/jtimms/TimmsDisclaimer.html?638679510849996568