Uni-Logo
You are here: Home BCCN Research A3: Functional plasticity in recurrent neuronal networks
Document Actions

A3: Functional plasticity in recurrent neuronal networks

A3: Functional plasticity in recurrent neuronal networks

Stefan RotterB, Marc-Oliver GewaltigD and Markus DiesmannQ


B = Bernstein Center for Computational Neuroscience;
D = Honda Research Institute Europe GmbH
Q = Riken Brain Science Institute


Scientific background

Learning in mammals depends on system-level processing involving both cortical and sub-cortical areas, presumably at least partially implemented by local mechanisms of synaptic plasticity. A number of biologically inspired algorithms have been developed which enable a machine to learn complex tasks. However, it remains largely unclear how these algorithms could be implemented by the brain, and how they are constrained by the dynamics of neuronal activity and its underlying network architecture. In our current research, we investigate spike-timing-dependent plasticity in recurrent networks that exhibit realistic dynamics and find that these two concepts are not compatible in a trivial way, most likely due to correlations. A two year development program has ensured that our simulation software, in contrast to other neural simulation tools, can cope with the computational demands of this project.


Objectives

In this project we want to bridge the gap between abstract functional models of learning and neurophysiological mechanisms believed to underlie it. Based on numerical simulations of a model system that learns and recalls temporal activity patterns (e.g. for the control of movement), we will study how new system-level functionality is acquired, how adaptation to a changing context occurs, and how compositionality of sub-functions is realized. This also yields specific predictions for physiological recordings in learning animals. We will formalize the results of our numerical studies and explain them by creating a biological theory of learning systems, linking plasticity of neuronal interactions with adaptivity of function and behavior. A better understanding of the constraints imposed by structure and dynamics of the biological substrate (e.g. the trade-off between network stability and flexibility) will result from this. New strategies to control and utilize plastic networks emerging from this work will be applied in other projects (C2-C5).


List of project-related publications

  1. Morrison A, Mehring C, Geisel T, Aertsen A, Diesmann M (2005). Advancing the boundaries of high connectivity network simulation with distributed computing. Neural Computation 17(8): 1776–1801.
  2. Morrison A, Aertsen A, Diesmann M (2007) Spike-timing dependent plasticity in balanced random networks. Neural Computation 19: 1437–1467.
  3. Morrison A, Diesmann M, Gerstner W (2008) Phenomenological Models of Synaptic Plasticity based on Spike-Timing. Biological Cybernetics 98: 459–478.
  4. Potjans W, Morrison A, Diesmann M (2009) A spiking neural network model of an actor-critic learning agent. Neural Computation 21: 301–339.

 

Abstracts:

  1. Morrison A, Aertsen A., Diesmann M (2006). Spike-timing dependent plasticity in balanced random networks. T79. In Computational Neuroscience Meeting, Edinburgh, UK.
  2. Morrison A, Aertsen A, Diesmann M (2007). Spike-timing dependent plasticity in balanced random networks. In: 10th Tamagawa-Riken Dynamic Brain Forum, Hakuba, JP.
  3. Potjans W, Morrison A, Diesmann M (2007) Reinforcement learning in an actor-critic spiking network model. In: 10th Tamagawa-Riken Dynamic Brain Forum, Hakuba, JP.
  4. Potjans W, Morrison, A, Diesmann, M (2007) Reinforcement learning in an actor-critic spiking network model. T28-9C. In: Neuroforum, Feb. vol. XIII, supplement. 7th Göttingen Meeting of the German Neuroscience Society.
  5. Schrader S, Morrison A, Diesmann M (2007) A composition machine for complex movements. TS18-1C. In: Neuroforum, Feb. vol. XIII, supplement. 7th Göttingen Meeting of the German Neuroscience Society.
  6. Potjans W, Morrison, A, Diesmann M (2007) A spiking neural network model for the actor-critic temporal-difference learning algorithm. Proceedings of the Society for Neuroscience 37th annual meeting.
  7. Potjans W, Morrison A, Diesmann M (2008) Synaptic plasticity rules for actor-critic temporal-difference learning, P2-r06, Neuroscience 2008 - The 31st Annual Meeting of the Japan Neuroscience Society, Tokyo, Japan.
  8. Potjans W, Morrison A, Diesmann M (2008) Linking synaptic plasticity to system-level learning in the framework of temporal-difference learning. First Status Seminar of the Helmholtz Alliance on Systems Biology. June 22nd - 24th, Potsdam, Germany.
  9. Potjans W, Morrison A, Diesmann M (2008) Properties of synaptic plasticity rules implementing actor-critic temporal-difference learning, #72, CNS*2008, July 19th - 24th, Portland, USA.
  10. Hanuschkin A, Kunkel S, Helias M, Morrison A, Diesmann M (2008) Time-Driven Simulation with Fully Asynchronous Pulse Coupling, Network Synchronization: from dynamical systems to neuroscience, (Lorentz Center), Leiden, The Netherlands.
  11. Hanuschkin A, Kunkel S, Helias M, Morrison A, Diesmann M (2008) Comparison of methods to calculate exact spike times in integrate-and-fire neurons with exponential currents, 17th Annual Computational Neuroscience Meeting 2008 (CNS2008), (OCNS), Portland, USA.
  12. Diesmann M, Hanuschkin A, Kunkel S, Helias M, Morrison A (2008) The performance of solvers for integrate-and-fire models with exact spike timing, 1st INCF Congress of Neuroinformatics: Databasing and Modeling the Brain (Neuroinformatics 2008), (INCF), Stockholm, Sweden.
  13. Hanuschkin A, Kunkel S, Helias M, Morrison A, Diesmann M (2008) Time-driven simulation with fully asynchronous pulse coupling, 4th Bernstein Symposium for Computational Neuroscience, (Bernstein Network), Munich, Germany.
  14. Ponulak F., Rotter S. (2008) Biologically inspired spiking neural model for motor control and motor learning. Proc. of “Perceptual learning, Motor learning and Automaticity”, Amsterdam, pp. 56.
  15. Ponulak F., Belter D., Rotter S. (2008) Adaptive Movement Control with Spiking Neural Networks, Part I: Feedforward Control. Proc. of “Recent Advances in Neuro-Robotics”, Symposium on Sensorimotor Control, Freiburg, pp. 47.
  16. Ponulak F., Belter D., Rotter S. (2008) Adaptive Movement Control with Spiking Neural Networks, Part II: Composite Control. Proc. of “Recent Advances in Neuro-Robotics”, Symposium on Sensorimotor Control, Freiburg, pp. 14.
  17. Ponulak F. (2008) Neural Mechanisms Improving Spike Timing Reliability under Noisy and Unreliable Conditions. Frontiers in Neuroscience, Conference Abstract: Bernstein Symposium 2008. doi: 10.3389/conf.neuro.10.2008.01.105.
  18. Hanuschkin, A., Herrmann J.M., Morrison A., Diesmann M. (2009). A model of free monkey scribbling based on the propagation of cell assembly activity. BMC Neuroscience 2009, 10(Suppl 1):P300 (13 July 2009).
  19. Kunkel S, Hanuschkin A, Helias M, Morrison A, Diesmann M (2009) Time-driven simulation as an efficient approach to detecting threshold crossings in precisely spiking neuronal network models, T26-7B, Meeting of the German Neuroscience Society, March 25-29 2009, Göttingen
  20. Potjans W., Morrison A. & Diesmann M. (2009) Linking synaptic to system-level learning through neuromodulated plasticity. German Symposium on Systems Biology, May 12th - 15th, Heidelberg, Germany.
  21. Potjans W., Morrison A. & Diesmann M. (2009) A spiking temporal-difference learning model based on dopamine-modulated plasticity. Eighteenth Annual Computational Neuroscience Meeting CNS*2009, July 18th - 23rd, Berlin, Germany (BMC Neuroscience 2009, 10(Suppl 1):P140
  22. Potjans W., Morrison A. & Diesmann M. (2009) Implementing neuromodulated plasticity in distributed simulations. 2nd INCF Congress of Neuroinformatics, September 6th - 8th, Pilsen, Czech Republic.
  23. Potjans W., Morrison A. & Diesmann M. (2009) Implementing temporal-difference learning through dopamine-modulated plasticity. Neuroscience 2009 - The 32nd Annual Meeting of the Japan Neuroscience Society, September 16th - 18th, Nagoya, Japan
  24. Potjans W., Morrison A. & Diesmann M. (2009) Distributed simulation of a spiking temporal-difference learning model based on dopamine modulated plasticity. Next Generation Supercomputing Symposium 2009, October 7th - 8th 2009, Tokyo, Japan
     
Personal tools