Modeling Information Transfer in the Brain

The group of Prof. Louis Tao from the Center for Quantitative Biology (CQB) and the School of Life Sciences at Peking University recently published a paper titled “Graded, Dynamically Routable Information Processing with Synfire-Gated Synfire Chains” in PLOS Computational Biology. Based on their previous work of pulse-gated synfire chains, they came up with a model network called synfire-gated synfire chains (SGSCs). SGSCs can rapidly cascade graded information through a neural circuit with multiple layers and are capable of performing complex computations. It provides a possible mechanism for information transfer in the brain, andthe construction of SGSCs also provides a practical pathway for designing computational neural circuits.

Coherent neural activities are believed to be important in many cognitivefunctions, such as memory, attention, binding and information transfer. Synfire chains are one of the main theoretical construct that have been applied to describe these phenomena. However, synchronous activity in feedforward networks of synfire chain will either approaches to an attractor or decay to zero. This limits the synfire chain’s ability to explain graded neuronal response and perform complex computations.

In their previous work, they have shown that pulse-gated synfire chains are capable of propagating graded information coded in mean population firing rates. In particular, they showed that it is possible to use one synfire chain to provide gating pulses and a second, pulse-gated synfire chain to propagate graded information. These circuits are called synfire-gated synfire chains (SGSCs). SGSCs can rapidly cascade graded information through a neural circuit, andthey can be well explained by a mean-field model in which gating pulses overlap in time. The graded transfer in SGSCs is robust in the presence of variability in population size, pulse timing and synaptic strength.

Fig1：Graded information transfer in synfire-gated synfire chains.

SGSC-based information coding can perform complex computations. In this work, they implemented a self-contained, spike-based, modular neural circuit that was triggered by streaming input, processed the input, and then made a decision based on the processed information and shut itself down. The construction of SGSCs also provides a practical pathway for designing computational neural circuits, either for the purpose of forming hypotheses about circuits in the brain, or for implementing algorithms on neuromorphic chips.

Fig 2: An autonomous decision making circuit based on SGSC.

The first author (Zhuo Wang) is a graduate student in the School of Life Sciences at Peking University. Andrew T. Sornborger (Department of Mathematics, University of California, Davis) and Louis Tao are the co-corresponding authors. This work was supported by the Ministry of Science and Technology of China through the Basic Research Program (973), by the Natural Science Foundation of China, by the Open Research Fund of the State Key Laboratory of Cognitive Neuroscience and Learning, by the Beijing Municipal Science and Technology Commission, and by the US National Institutes of Health CRCNS program.

Link to the paper:

http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1004979