Dynamics in networks of spiking neurons
Networks of spiking neurons have several interesting dynamic properties,
for example very rapid and characteristic transients,
synchronous firing and asynchronous states.
A better understanding of typical
phenomena has important implications
for problems associated with neuronal coding (spikes or rates).
For example, the population activity is a rate-type quantity, but does not
need temporal averaging - which suggests
fast rate coding as a potential strategy.
The idea of the workshop is to start from
mathematical models of network dynamics,
see what is known in terms of results,
and then try to find out what the implications
for 'coding' in the most general sense could be.
A few points that could be raised in the workshop:
Potential audience: Researchers interested in
questions of coding and dynamics.
The populatation activity can respond
rapidly to changes of the stimulus
(rapid switching, e.g. Treves, 1992; Tsodyks and Sejnowski, 1995;
Van Vreeswijk and Somolinsky, 1996).
This suggests the capability of fast
signal transmission, consistent with
reaction time experiments (Thorpe et al. 1996).
What are the implications for 'rapid computation'?
Networks of spiking neurons
with balanced excitation/inhibition are noisy
(e.g., Van Vreeswijk and Sompolinsky, 1996) and so are cortical neurons
(e.g., Softky and Koch, 1993).
Is it `useful' for the net to have a certain noise level?
What is the signal-to-noise ratio?
How do correlations in the firing times
of different neurons affect the signal transmission
capacity of the population (Spiridon et al., 1998; Panzeri et al., 1998)?
In simulations simplified models of population activity
must be used, e.g., Population Density Methods or
What is a good description of the dynamics of population activity?
What are the differences in the dynamics between homogeneous and
heterogeneous networks? Which coding strategies work
better in homogeneous, which better in inhomogeneous networks?
What would be desirable structures/connectivity patterns
to exploit the dynamical properties of spiking neurons?
What is the use of spontaneous acitivity in networks
of spiking neurons?
In view of all the above, what are coding strategies
in large networks of spiking neurons?
The workshop is planned for 2 sessions on Friday, Dec. 4.
Each session will be started off
by a couple of talks with enough time for lots of discussions.
(4 x 25 min talks + 40 minutes discussion in the morning,
3x 25 minutes talks + 90 minutes discussion in the evening).
- 7.30-7.50 Introduction (W.Gerstner, Lausanne)
- 7.50-8.20 C. van Vreeswijk (London/Jerusalem),
Asynchronous chaos-like firing in
- 8.20-8.50 D. Hansel (Paris),
Robust synchrony in large heterogeneous neuronal networks:
the role of inhibition
- 8.50-9.15 Discussion
- 9.15-9.45 D. Nykamp (NYU),
A Population Density Method that Facilitates
Large-Scale Modeling of Neural Networks
- 9.45-10.15 J. Eggert (TU Munich),
Dynamics of assemblies of spiking neurons:
Integral and differential equation approaches
- 10.15-10.30 Discussion
- ----- BREAK ------
- 4:00-4:30 S. Panzeri (Oxford),
The role of correlations in fast information transmission
- 4:30-5:00 W. Maass (Graz),
Neural Systems as non-linear Filters
- 5:00-5:30 3 short contributions from additional participants
- 5:30-7:00 PANEL DISCUSSION
including B.J. Richmond (NIH) and the speakers:
What do we learn from all of this for coding?
- role of noise
- sources of noise
- role of heterogeneity
- fast information processing schemes
will be held on December 4 and 5
in Breckenridghe (Colorado). The workshops follow
the main conference
Neural Information Processing Systems, NIPS'98
which will be held in Denver on December 1-3.