Neuroscience 2005 Abstract
| Presentation Number: | 160.1 |
|---|---|
| Abstract Title: | Limits on the improvement of memory storage due to multiple levels of synaptic efficacy. |
| Authors: |
Fusi, S.*1
; Abbott, L. F.1
1Center for Neurobiology and Behavior, Columbia Univ. Col. of Physician and Surgeons, New York, NY |
| Primary Theme and Topics |
Neural Excitability, Synapses, and Glia: Cellular Mechanisms - Synaptic Plasticity -- LTP: Physiology and behavior |
| Session: |
160. LTP Physiology and Behavior: Modulation In Vivo Poster |
| Presentation Time: | Sunday, November 13, 2005 8:00 AM-9:00 AM |
| Location: | Washington Convention Center - Hall A-C, Board # L7 |
| Keywords: | synaptic plasticity, LTP, LTD, learning |
Old memories, maintained in patterns of synaptic connectivity, can be destroyed by overwriting when plasticity mechanisms modify synaptic efficacies during the storage of new memories. Such overwriting leads to forgetting, with memory lifetimes that only grow proportional to the logarithm of the number of synapses being used to store the memories. Previous, it has been suggested that the factor multiplying this logarithm could be increased dramatically, and thus memory performance improved, if synapses have many different levels of strength between which long-term potentiation and depression (LTP and LTD) can induce transitions (Amit & Fusi, Neural Computation, 1994 and Fusi, Biol. Cybernetics, 2002).
In agreement with previous results, we find that memory lifetimes grow proportional to the square of the number of levels of synaptic efficacy, but only if there is a near-perfect balance between the effects of LTP and LTD. If this balance is not precisely maintained, the improvement decreases significantly. In the models we consider, synaptic plasticity cannot increase or decrease synaptic efficacy beyond specific bounds. If the change in synaptic efficacy induced by plasticity is smoothly reduced as the bound is approached (soft bounds) and LTP and LTD are not perfectly balanced, the memory lifetime depends only linearly, not quadratically, on the number of synaptic levels. If synaptic plasticity only stops changing synaptic efficacy when the boundary is reached (hard bounds), the effect is even more drastic and the memory lifetime becomes insensitive to the number of synaptic levels. These results are robust to potentiation/depression imbalance and are shared by a large class of models. In the case where synaptic efficacy can vary continuously, the ratio of the total range of synaptic efficacy to the magnitude of the typical change induced by LTP or LTD plays an analogous role to the number of synaptic levels and the same results apply.
Research supported by National Institutes of Health Grant MH-58754 and an NIH Director s Pioneer Award.
In agreement with previous results, we find that memory lifetimes grow proportional to the square of the number of levels of synaptic efficacy, but only if there is a near-perfect balance between the effects of LTP and LTD. If this balance is not precisely maintained, the improvement decreases significantly. In the models we consider, synaptic plasticity cannot increase or decrease synaptic efficacy beyond specific bounds. If the change in synaptic efficacy induced by plasticity is smoothly reduced as the bound is approached (soft bounds) and LTP and LTD are not perfectly balanced, the memory lifetime depends only linearly, not quadratically, on the number of synaptic levels. If synaptic plasticity only stops changing synaptic efficacy when the boundary is reached (hard bounds), the effect is even more drastic and the memory lifetime becomes insensitive to the number of synaptic levels. These results are robust to potentiation/depression imbalance and are shared by a large class of models. In the case where synaptic efficacy can vary continuously, the ratio of the total range of synaptic efficacy to the magnitude of the typical change induced by LTP or LTD plays an analogous role to the number of synaptic levels and the same results apply.
Research supported by National Institutes of Health Grant MH-58754 and an NIH Director s Pioneer Award.
Sample Citation:
[Authors]. [Abstract Title]. Program No. XXX.XX. 2005 Neuroscience Meeting Planner. Washington, DC: Society for Neuroscience, 2005. Online.
Copyright © 2005-2025 Society for Neuroscience; all rights reserved. Permission to republish any abstract or part of any abstract in any form must be obtained in writing by SfN office prior to publication.
