Dissection of bidirectional synaptic plasticity into saturable unidirectional processes

Daniel H. O'Connor, Gayle M. Wittenberg, Samuel S.H. Wang

Research output: Contribution to journalArticlepeer-review

50 Scopus citations

Abstract

In populations of synapses, overall synaptic strength can undergo either a net strengthening (long-term potentiation) or weakening (long-term depression). These phenomena have distinct induction pathways, but the functional outcome is usually measured as a single lumped quantity. In hippocampal CA3-CA1 synapses, we took two approaches to study the activity dependence of each phenomenon in isolation. First, we selectively blocked one process by applying kinase or phosphatase inhibitors known, respectively, to block potentiation or depression. Second, we saturated depression or potentiation and examined the activity dependence of the converse process. The resulting unidirectional learning rules could be recombined to give a well-known bidirectional frequency-dependent learning rule under the assumption that when both pathways are activated kinases dominate, resulting in potentiation. Saturation experiments revealed an additional process in which potentiated synapses can be locked at high strength. Saturability of the components of plasticity implies that the amount of plasticity contributed by each pathway depends on the initial level of strength of the synapses. Variation in the distribution of initial synaptic strengths predicts a form of metaplasticity and can account for differences in learning rules observed under several physiological and genetic manipulations.

Original languageEnglish (US)
Pages (from-to)1565-1573
Number of pages9
JournalJournal of neurophysiology
Volume94
Issue number2
DOIs
StatePublished - Aug 2005
Externally publishedYes

ASJC Scopus subject areas

  • General Neuroscience
  • Physiology

Fingerprint

Dive into the research topics of 'Dissection of bidirectional synaptic plasticity into saturable unidirectional processes'. Together they form a unique fingerprint.

Cite this