Modelling Memory Consolidation with STDP
-
1
Humboldt-University Berlin and BCCN Berlin, Institute for Theoretical Biology, Germany
Since the famous case of patient H.M. it is known that the acquisition of episodic and semantic memory relies on the hippocampus. After months, or even years, memories become hippocampus-independent and are stored in the neocortex. This process of information transfer is called memory consolidation. However, the detailed biological mechanisms underlying consolidation remain unknown.
As a basis for a mechanistic model of consolidation, we propose a canonical consolidation circuit. In this circuit, the inputs to a network are transformed to outputs via two possible pathways: an indirect pathway, that maps via an intermediate stage of processing, and a direct pathway, that maps directly to the output. In a hippocampal setting, the input stage could correspond to entorhinal cortex, the output to CA1, and the intermediate stage to the dentate gyrus together with CA3. CA3 is generally believed to be the place where new episodic memories are encoded. Hence, during consolidation, information stored in CA3 should, in a first step, get consolidated to the next level, i.e. the direct pathway. We call the circuit canonical because the direct-indirect pathway pattern seems to be ubiquitous in the brain.
As a mechanism for the transfer of information from the indirect to the direct pathway, we propose spike-timing-dependent plasticity (STDP). STDP is known to adapt synaptic strenghts to the earliest input spikes that correlate with a neuron's output. As the indirect pathway has an additional stage of processing, signals take longer to arrive at the output neurons as compared with the direct pathway. The indirect pathway shares its inputs with the direct pathway, and therefore spikes caused by the indirect pathway necessarily have correlations with input along the direct pathway. These correlations drive STDP to transfer information from the indirect pathway to the direct pathway.
We analyzed the information transfer in a linear rate-based model and confirmed the results by using a numerical implementation of the model. Because information transfer is bidirectional, memory transfer may oscillate with time. To increase the performance in our model, this oscillation can be prevented by decreasing the learning rates on the indirect pathway during consolidation, which is a testable prediction of the model. Finally, we implemented the model with integrate-and-fire neurons and confirmed the results of the linear model prediction.
Keywords:
computational neuroscience,
Learning and plasticity,
memory consolidation,
Spike-timing-dependent plasticity
Conference:
BC11 : Computational Neuroscience & Neurotechnology Bernstein Conference & Neurex Annual Meeting 2011, Freiburg, Germany, 4 Oct - 6 Oct, 2011.
Presentation Type:
Poster
Topic:
learning and plasticity (please use "learning and plasticity" as keyword)
Citation:
Bergmann
UM,
Sprekeler
H and
Kempter
R
(2011). Modelling Memory Consolidation with STDP.
Front. Comput. Neurosci.
Conference Abstract:
BC11 : Computational Neuroscience & Neurotechnology Bernstein Conference & Neurex Annual Meeting 2011.
doi: 10.3389/conf.fncom.2011.53.00193
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
19 Aug 2011;
Published Online:
04 Oct 2011.
*
Correspondence:
Dr. Urs M Bergmann, Humboldt-University Berlin and BCCN Berlin, Institute for Theoretical Biology, Berlin, Germany, urs.bergmann@biologie.hu-berlin.de