Computational neuroscience: Memory-making is all about the connection

Nov. 8, 2012 — A model that shows how connections in the brain must change to form memories could help to develop artificial cognitive computers

Exactly how memories are stored and accessed in the brain is unclear. Neuroscientists, however, do know that a primitive structure buried in the center of the brain, called the hippocampus, is a pivotal region of memory formation. Here, changes in the strengths of connections between neurons, which are called synapses, are the basis for memory formation. Networks of neurons linking up in the hippocampus are likely to encode specific memories.

Since direct tests cannot be performed in the brain, experimental evidence for this process of memory formation is difficult to obtain but mathematical and computational models can provide insight. To this end, Eng Yeow Cheu and co-workers at the A*STAR Institute for Infocomm Research, Singapore, have developed a model that sheds light on the exact synaptic conditions required in memory formation.

Their work builds on a previously proposed model of auto-associative memory, a process whereby a memory is retrieved or completed after partial activation of its constituent neural network. The earlier model proposed that neural networks encoding short-term memories are activated at specific points during oscillations of brain activity. Changes in the strengths of synapses, and therefore the abilities of neurons in the network to activate each other, lead to an auto-associative long-term memory.

Cheu and his team then adapted a mathematical model that describes the activity of a single neuron to incorporate specific characteristics of cells in the hippocampus, including their inhibitory activity. This allowed them to model neural networks in the hippocampus that encode short-term memories. They showed that for successful formation of auto-associative memories, the strength of synapses needs to be within a certain range: if synapses become too strong, the associated neurons are activated at the wrong time and networks become muddled, destroying the memories. If they are not strong enough, however, activation of some neurons in the network is not enough to activate the rest, and memory retrieval fails.

As well as providing insight into how memories may be stored and retrieved in the brain, Cheu thinks this work also has practical applications. "This study has significant implications in the construction of artificial cognitive computers in the future," he says. "It helps with developing artificial cognitive memory, in which memory sequences can be retrieved by the presentation of a partial query." According to Cheu, one can compare it to a single image being used to retrieve a sequence of images from a video clip.

The A*STAR-affiliated researchers contributing to this research are from the Institute for Infocomm Research.

Share this story on Facebook, Twitter, and Google:

Other social bookmarking and sharing tools:

Story Source:

The above story is reprinted from materials provided by The Agency for Science, Technology and Research (A*STAR).

Note: Materials may be edited for content and length. For further information, please contact the source cited above.

Journal Reference:

Eng Yeow Cheu, Jiali Yu, Chin Hiong Tan, Huajin Tang. Synaptic conditions for auto-associative memory storage and pattern completion in Jensen et al.’s model of hippocampal area CA3. Journal of Computational Neuroscience, 2012; 33 (3): 435 DOI: 10.1007/s10827-012-0394-8

Note: If no author is given, the source is cited instead.

Disclaimer: This article is not intended to provide medical advice, diagnosis or treatment. Views expressed here do not necessarily reflect those of ScienceDaily or its staff.

View the original article here

Biologists create first predictive computational model of gene networks that control development of sea-urchin embryos

ScienceDaily (Aug. 29, 2012) — As an animal develops from an embryo, its cells take diverse paths, eventually forming different body parts -- muscles, bones, heart. In order for each cell to know what to do during development, it follows a genetic blueprint, which consists of complex webs of interacting genes called gene regulatory networks.

Biologists at the California Institute of Technology (Caltech) have spent the last decade or so detailing how these gene networks control development in sea-urchin embryos. Now, for the first time, they have built a computational model of one of these networks.

This model, the scientists say, does a remarkably good job of calculating what these networks do to control the fates of different cells in the early stages of sea-urchin development -- confirming that the interactions among a few dozen genes suffice to tell an embryo how to start the development of different body parts in their respective spatial locations. The model is also a powerful tool for understanding gene regulatory networks in a way not previously possible, allowing scientists to better study the genetic bases of both development and evolution.

"We have never had the opportunity to explore the significance of these networks before," says Eric Davidson, the Norman Chandler Professor of Cell Biology at Caltech. "The results are amazing to us."

The researchers described their computer model in a paper in the Proceedings of the National Academy of Sciences that appeared as an advance online publication on August 27.

The model encompasses the gene regulatory network that controls the first 30 hours of the development of endomesoderm cells, which eventually form the embryo's gut, skeleton, muscles, and immune system. This network -- so far the most extensively analyzed developmental gene regulatory network of any animal organism -- consists of about 50 regulatory genes that turn one another on and off.

To create the model, the researchers distilled everything they knew about the network into a series of logical statements that a computer could understand. "We translated all of our biological knowledge into very simple Boolean statements," explains Isabelle Peter, a senior research fellow and the first author of the paper. In other words, the researchers represented the network as a series of if-then statements that determine whether certain genes in different cells are on or off (i.e., if gene A is on, then genes B and C will turn off).

By computing the results of each sequence hour by hour, the model determines when and where in the embryo each gene is on and off. Comparing the computed results with experiments, the researchers found that the model reproduced the data almost exactly. "It works surprisingly well," Peter says.

Some details about the network may still be uncovered, the researchers say, but the fact that the model mirrors a real embryo so well shows that biologists have indeed identified almost all of the genes that are necessary to control these particular developmental processes. The model is accurate enough that the researchers can tweak specific parts -- for example, suppress a particular gene -- and get computed results that match those of previous experiments.

Allowing biologists to do these kinds of virtual experiments is precisely how computer models can be powerful tools, Peter says. Gene regulatory networks are so complex that it is almost impossible for a person to fully understand the role of each gene without the help of a computational model, which can reveal how the networks function in unprecedented detail.

Studying gene regulatory networks with models may also offer new insights into the evolutionary origins of species. By comparing the gene regulatory networks of different species, biologists can probe how they branched off from common ancestors at the genetic level.

So far, the researchers have only modeled one gene regulatory network, but their goal is to model the networks responsible for every part of a sea-urchin embryo, to build a model that covers not just the first 30 hours of a sea urchin's life but its entire embryonic development. Now that this modeling approach has been proven effective, Davidson says, creating a complete model is just a matter of time, effort, and resources.

The title of the PNAS paper is "Predictive computation of genomic logic processing functions in embryonic development." In addition to Peter and Davidson, the other author on the PNAS paper is Emmanuel Faure, a former Caltech postdoctoral scholar who is now at the École Polytechnique in France. This work was supported by the National Institute of Child Health and Human Development.

Share this story on Facebook, Twitter, and Google:

Other social bookmarking and sharing tools:

Story Source:

The above story is reprinted from materials provided by California Institute of Technology. The original article was written by Marcus Woo.

Note: Materials may be edited for content and length. For further information, please contact the source cited above.

Journal Reference:

I. S. Peter, E. Faure, E. H. Davidson. Predictive computation of genomic logic processing functions in embryonic development. Proceedings of the National Academy of Sciences, 2012; DOI: 10.1073/pnas.1207852109

Note: If no author is given, the source is cited instead.

Disclaimer: Views expressed in this article do not necessarily reflect those of ScienceDaily or its staff.

View the original article here