Contribution to understanding the mechanisms underlying memory formation and coding in the brain via the use of computational models | News

All Science News


Contribution to understanding the mechanisms underlying memory formation and coding in the brain via the use of computational models

Work in the Computational Biology Laboratory of IMBB-FORTH, which is headed by Dr. Panayiota Poirazi, has used computational models to show that a single neuronal cell in the brain is capable of detecting spatiotemporal differences of incoming signals and encoding them into its response pattern.

This work, which is published today in PLoS Computational Biology, is particularly important as it predicts that a single neuron has the capacity to detect the spatial and temporal characteristics of memory-related signals, encode these characteristics into its response using a temporal code and transmit this information to higher brain regions. So far, such properties have been attributed only to large neuronal populations. The new work reveals for the first time the computational processing capability of individual cells in the brain, the extent of which had not been previously appreciated.

The hippocampus is a brain region that has been extensively studied because of its role in learning and memory. The Cornus Ammonis 1 (CA1) region is the output of the hippocampal formation and pyramidal neurons in this region are the elementary units responsible for the processing and transfer of information to the cortex. Dr. Poirazi, working with Dr Eleftheria Pissadaki (a Ph.D. student at the time, now a visiting researcher at the Oxford University) developed a detailed computational model of a CA1 pyramidal neuron and used it to investigate how the location of incoming signals and their temporal characteristics are recorded by the responding neurons. More specifically, Drs Poirazi and Pissadaki showed that the temporal separation between two incoming signals serves as a switch between regular firing and bursting: inputs that arrive with a large temporal offset cause attenuation of the neuronal response, while temporally associated or synchronous signals facilitate neuronal output leading to bursting.

It is possible that this temporal switch mechanism serves as a familiarity detector: known environmental stimuli, such as previously seen objects, are quickly recognized by the neuron leading to an enhanced response, thus indicating the recognition of a previously established memory. On the contrary, novel environmental stimuli are associated with long delays (which may result from the search process), leading to a suppression of the neuronal output, thus allowing some other region to process and store the new memory. This model prediction is in agreement with the existing literature and awaits experimental verification.

An important finding of this work concerns the way such information is coded in the brain. According to the computational model, individual CA1 pyramidal neurons utilize a temporal code comprising of the time-intervals prior to the onset of the response and between spikes within bursts to transmit information about the temporal association (extent of familiarity) as well as the location of incoming signals within their dendritic tree.

Overall, this work contributes significantly to our understanding of the mechanisms underlying memory formation in the hippocampus. In addition, this helps us to understand how memory deficits appear in normal aging and in neurodegenerative conditions, such as Alzhemeir’s and Parkinson’s disease. Furthermore, the findings of this work may be applied in neuromorphic engineering and neuroprosthetics for the development of intelligent artificial systems that mimic brain microanatomy.

For more information contact:
Dr. Panayiota Poirazi
Computational Biology Laboratory
Institute of MolecularBiology and Biotechnology, FORTH
Τηλ.: 2810 391139, e-mail:

Relevant Publication:
Pissadaki E.K., Sidiropoulou K., Reczko M., and Poirazi P. Encoding of Spatio-temporal Input Characteristics by a CA1 Pyramidal Neuron Model, PLoS Computational Biology, December, 2010 (