Inhibition in the neural network models of the 1920s and 1930s

Vincent Wan
Committee on the Conceptual Foundations of Science, University of Chicago. USA

Models of neural circuitry came into use at the close of the 19th century (e.g., Cajal, Exner). By the 1920s and 1930s, diagrams of small groups of neurons were being used to bridge the gap between the properties of nerves and the central nervous system's behavior. In a 1917 posthumous work, Keith Lucas suggested axons be used as a model for the behavior of central neurons. This idea fueled much of the work on axons during first decades of the 20th century. In 1922, Alexander Forbes popularised the use of circuit diagrams to show how neurons, behaving like axons, could account for central phenomena (e.g., spinal reflexes). Before the 1950s there were two major classes of models of inhibition. (1) models proposing axons with direct inhibitory chemical or electrical affects, and (2) models proposing that excitatory axons have inhibitory effects by increasing the relative refractory periods of the postsynaptic cells (Wedensky inhibition). In the work reported here, I trace the development of the second class of neural network models of inhibition from Forbes to the 1939 American Physiological Society symposium on the synapse. In the 1930s several basic arrangement of neurons were recognised as functional units (e.g., delay paths, reverberent circuits, and conditional pathways). I have used computer simulations to examine the behavior of these elements and their use in the production of inhibition. Recognising these basic elements helps us see the connections and contrasts between the circuits proposed by different investigators.


Poster Session II
Friday, 20 June 1997, 16.10 - 16.40

Second Annual Meeting of the International Society for the History of the Neurosciences (ISHN) and 6th Meeting of the European Club on the History of Neurology (ECHN)

Leiden, The Netherlands