ModelDB: Learning spatial transformations through STDP (Davison, Frégnac 2006)

Learning spatial transformations through STDP (Davison, Frégnac 2006)

Accession: 64261

A common problem in tasks involving the integration of spatial information from multiple senses, or in sensorimotor coordination, is that different modalities represent space in different frames of reference. Coordinate transformations between different reference frames are therefore required. One way to achieve this relies on the encoding of spatial information using population codes. The set of network responses to stimuli in different locations (tuning curves) constitute a basis set of functions which can be combined linearly through weighted synaptic connections in order to approximate non-linear transformations of the input variables. The question then arises how the appropriate synaptic connectivity is obtained. This model shows that a network of spiking neurons can learn the coordinate transformation from one frame of reference to another, with connectivity that develops continuously in an unsupervised manner, based only on the correlations available in the environment, and with a biologically-realistic plasticity mechanism (spike timing-dependent plasticity).
Reference: Davison AP, Fregnac Y (2006) Learning Cross-Modal Spatial Transformations through Spike Timing-Dependent Plasticity J Neurosci 26:5604-5615 [PubMed]

Citations Citation Browser

Model Information (Click on a link to find other models with that property)

Model Type:	Network;
Brain Region(s)/Organism:	Generic;
Cell Type(s):
Channel(s):
Gap Junctions:
Receptor(s):	GabaA; AMPA;
Gene(s):
Transmitter(s):
Simulation Environment:	Neuron;
Model Concept(s):	Synaptic Plasticity; Long-term Synaptic Plasticity; Unsupervised Learning; STDP;
Implementer(s):	Davison, Andrew [Andrew.Davison at iaf.cnrs-gif.fr];

Search NeuronDB for information about: GabaA; AMPA;

Model files

Help downloading and running models


The bfstdp_demo.hoc (which runs on Auto-launch) runs the learning procedure for
10000 s (about 3 hours) and displays the weight matrix from Input to Output
layers together with the time-evolution of three of the weights.

The simulation takes 20 minutes or so, but you should be able to see the weight
patterns begin to form, as in Figure 2 of the J. Neurosci. paper, before this
time.

Data files are produced in a subdirectory called Data, and may be plotted using
Matlab, Gnuplot (use pm3d in map view), etc.  A sample graphing program,
bfstdp_many.gnu, is provided but must be modified by globally changing the date
and time from 20071217_1201 to the time recorded in the filenames in the Data
directory.  When you run the file in linux with the command
"gnuplot bfstdp_many.gnu" it will then produce a postscript file that looks similar to:

ModelDB Home SenseLab Home Help
Questions, comments, problems? Email the ModelDB Administrator
How to cite ModelDB
This site is Copyright 2012 Shepherd Lab, Yale University