Direct Assessment of Synaptic Modification Rules
Abstract
We have spent considerable time reprogramming our computer programs for data acquisition and evaluation. This is an important effort since we were previously unable to study simultaneously the synaptic response and the cell discharge. This improvement has become particularly critical since the ongoing evaluation of data gathered last year shows strong support for the existence of two distinct adaptive processes. One process modifies the synaptic response, and another adaptive process modifies the conversion of synaptic current into cell firing. Note that this improvement in data gathering is an on-going task. We are continuing our study (i.e., data gathering and evaluation) of the quantitative manner in which asymptotic changes are induced independently at neighboring synapses. This study has at least two important implications. First, it corroborates our earlier claims that we are studying a process of individual synaptic modification. Second, it helps to establish the experimental conditions which allow us to distinguish a variety of adaptive modification processes. Finally, we have continued our theoretical work which considers various interpretations of the adaptive processes we have experimentally observed. The context of this interpretation now centers on optimally performing, adaptive pattern recognition systems. We are encouraged by the performance shown by multiplicative, recursive neural networks.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1984
- Accession Number
- ADA217198
Entities
People
- William B. Levy
Organizations
- University of Virginia