tmgsyn.mod is an implementation of a model of short-term synaptic plasticity using NEURON's model description language NMODL. The model has been described by Tsodyks and collaborators in publications that include Tsodyks, M., Pawelzik, K., Markram, H. Neural networks with dynamic synapses. Neural Computation 10:821-835, 1998 and Tsodyks, M., Uziel, A., Markram, H. Synchrony generation in recurrent networks with frequency-dependent synapses. J. Neurosci. 2000 RC50. This demonstration is patterned after Fig. 1A-C in the Neural Computation paper, except that the synapse is implemented with a conductance change, whereas Tsodyks et al. assumed that the synapse acted as a current source. NOTES: 1. The important part of this package of files is the synaptic mechanism, not the "cells." The synaptic mechanism is reusable in models of biophysical neurons. The "cells" are merely single-compartment passive models whose sole purpose is to serve as leaky integrators. 2. The presynaptic stimulus and its connection to the synaptic mechanism were implemented with NEURON's Artificial Cell and Network Builder tools. For tutorials about how to use these tools, and information about NEURON's model description language NMODL, see the Documentation page at NEURON's WWW site ( http://www.neuron.yale.edu/ ) especially the FAQ link on that page. 3. If you are interested in other models of short term synaptic plasticity, see ModelDB for an implementation of the model described by Varela, J.A., Sen, K., Gibson, J., Fost, J., Abbott, L.R., and Nelson, S.B.. A quantitative description of short-term plasticity at excitatory synapses in layer 2/3 of rat primary visual cortex. Journal of Neuroscience 17:7926-7940, 1997. (search under author's name, or under topics such as "synaptic plasticity"). Contact ted.carnevale@yale.edu if you have questions about this implementation of the model described by Tsodyks et al..