Elucidating the brain’s neural network
29 May 2009 (Volume 4 Issue 5)
The brain comprises more than 100 billion neurons, which are mutually connected by means of dendritic projections and axons extending from their cell bodies to form an extremely complex network of neural circuits. If all dendritic projections and axons in the whole brain were joined together in a line, they would extend for 100,000 kilometers. Toshihiko Hosoya, Unit Leader of the Hosoya Research Unit at RIKEN’s Brain Science Institute, is working to clarify neuron connections and information processing in these neural circuits. This article describes his work on elucidating the huge complexity of the brain’s neural network (Fig. 1).
Figure 1: The six-layer structure of the cerebral cortex.
The cerebral cortex is about 2 mm thick, covering the surface of the cerebrum. It consists of six layers, the outermost being layer 1, and the innermost layer 6. The fifth layer contains output cells that send information to the outside of the cerebral cortex, and other cells. Hosoya is exploring important circuit structures by means of genetic engineering techniques.
“Now is the time for me to begin doing what I dreamt of when I was a university student,” says Hosoya joyfully.
When he was a student in the Department of Physics at the University of Tokyo’s School of Science, Hosoya had the vague thought, “Perhaps I should study the brain.” One day he attended a lecture that made a deep impression on him. “While presenting some relevant data, Yoshiki Hotta (now director of the Research Organization of Information and Systems) said, ‘In the Drosophila brain, morphologically similar neurons are arranged, but they exhibit different patterns of gene expression. There are many types of neurons characterized by unique features, which are arranged with some regularity.’ This shocked me, because I had been convinced that the brain’s neural network is basically too complex to study meaningfully. I decided to perform research to clarify this complex neural network and how it undertakes signal processing by identifying the wide variety of neurons with a gene engineering approach.”
Looking at the brain from the neural circuit of the retina
The human brain comprises more than 100 billion neurons. From each cell body there extend a number of dendritic projections like branches from a tree, and one long axon, connecting to other neurons to form a neural circuit. On receiving an external stimulus, each neuron produces an electric signal, which travels to the adjacent neuron through the axon, resulting in signal transduction. “The neural network of the brain is huge and complex. It is difficult to get a handle on the whole neural network as it is,” says Hosoya. The strategy he first adopted was to examine in detail a neural circuit that is small but undertakes something integral. By clarifying the mechanism of the neural circuit and elucidating how information is processed in it, he aims at establishing a common principle that will be applicable to larger neural circuits.
The retina is an example of a small neural circuit that undertakes some kind of integral information processing. Located in the back of the eyeball, the retina is part of the central nervous system. It receives light coming from the outside as optical information, and this is transferred as electrical signals to the brain through the optic nerve. “If an eyeball is taken out and gently patted, the retina detaches easily. The neural circuit of the retina can thus be extracted intact, without damaging the complete process from input to output,” says Hosoya. By placing the removed retina on an electrode and stimulating it with a projected image—a light–dark pattern or something similar—one can determine how the neurons respond to a particular input, and what kind of information is transferred to the brain (Fig. 2). Hosoya uses the salamander retina, because it has long been used as research material, and also because its neurons are large and easy to handle.
Figure 2: Diagram of retina experiment.
An image displayed on the screen is projected on the retina, and the firing of neurons is detected by the electrode and stored in a computer. P, photoreceptor cell; B, bipolar cell; G, ganglion cell; H, horizontal cell; A, amacrine cell. The photograph shows an image of a salamander retina placed on electrodes (black dots). The lines in the upper portion of the photograph are transparent signal wires connecting the electrodes to the computer.
Another advantage of using the retina is that images and bright–dark patterns for stimulating it can easily be created by computer. “However, slapdash attempts would not attain the goal. So I used hints from information processing theory in drawing up my experimental plans,” says Hosoya. “In information processing, there are many established theories to the effect that ‘by processing in this way, its efficiency can be increased, without affecting the quality of data.’ With the expectation that the same thing might apply to neural circuits, I formulated an experimental hypothesis and performed experiments to verify it.”
In recording an image, for example, inputting the pieces of information from individual dots one by one would result in a vast amount of data. In addition, adjacent dots have very similar levels of brightness, so the transmitted data carries mostly similar information. This is very wasteful. For this reason, only the differences in brightness between adjacent dots are recorded, to reduce the volume of data required and to make the information processing more efficient. “It has long been known that the retina also extracts differences in brightness between adjacent dots and transmits the information to the brain. This is an effective way to use a neural circuit. There should be other features of neural circuits shared by information processing theories.”
When he was working as a research fellow in Markus Meister’s laboratory at Harvard University, Hosoya showed that the retina changes its signal transduction rule according to the nature of the image received. In the case of vertical stripes, for example, similar levels of brightness occur in the vertical direction, and it is redundant to transmit this information as it is. Hence, it is more efficient to focus on brightness differences in the vertical direction. The reverse applies to horizontal stripes. In fact, the retina streamlines its process of signal transduction by changing the rules according to the nature of the image. “When the data on a video image is compressed, the information processing rule is changed according to the direction of movement of the object, from right to left, or from above to below. By doing so, the amount of data can be reduced without any deterioration in the quality of the image. I thought that the same thing is being done by the retina.”
He learnt from Meister how to get hints from information processing theories. “He has discovered one new idea after another about information processing in neural circuits by using the retina. He always says, ‘Do something like giving a bang on the head.’ He suggests doing something that has not been thought of by anyone else. I learned many things from him.”
Signal transduction with an accurate firing pattern
In 2003, after returning to Japan, Hosoya organized his research unit in the RIKEN Brain Science Institute. One of its research aims is to continue to elucidate the important mechanisms common to neural circuits using the retina.
“It has been believed that a single neuron usually transmits a single piece of information. However, doesn’t that seem wasteful?” says Hosoya. “Information will be transmitted more efficiently if multiple pieces of information are carried in one signal. This also forms the fundamental basis of information processing theory.”
Therefore Hosoya decided to investigate extensively what information is transmitted by the retinal neurons. He used the neuron that responds to OFF, which ‘fires’ (generates an electrical signal) when it becomes dark. First, he repeatedly stimulated the retina using an image with changing dark areas, and examined the timing of the firing of the neuron that responds to OFF. It used to be thought that the responses of the retinal neurons involved a wide temporal fluctuation, with the timing of the firing differing between events, even in response to the same stimulation. In recent years, however, it has become evident that firing occurs with high reproducibility in terms of time in response to the same stimulation. Hosoya showed experimentally that the firing events in response to the same stimulation occurred with only a small fluctuation of several thousandths of a second. “This is quite an intriguing feature,” says Hosoya. “Because the retinal neurons fire with high reproducibility in response to the same stimulation, it has become easy to analyze what kind of information is conveyed when the neuron fires.”
The retina neurons that respond to OFF fire more frequently as the image input becomes darker and darker. “It has long been known that the firing frequency changes depending on ‘how dark it has become,’ or on the amplitude of the intensity change. Extensive examination shows, however, that both the firing frequency and the time interval are quite accurate (Fig. 3). This seems to carry some information.”
Figure 3: Responses of retinal neurons.
Firing (green points) of OFF-responding neurons in the salamander retina on stimulation by the brightness change shown by the gray line. This represents data for a single neuron receiving the same brightness change for 21 iterations. Firings caused by each iteration are shown in a row; cases of three firings are colored. The intervals between the first and second firings and between the second and third firings are found to be highly reproducible between all iterations.
Hosoya theorizes as follows. “The neurons may transmit information on ‘how it has become dark’ as well as on ‘how dark it has become’.” It has also been demonstrated that accurate firing patterns are transmitted from the retina to the brain through the optic nerve. Upsetting conventional common sense, it may be shown that a single neuron transmits multiple pieces of information by using an accurate firing pattern.
Exploring the circuit structure of the cerebral cortex
Another pillar of the research activities at Hosoya’s laboratory is the use of genetic engineering to understand a neural circuit that has been difficult to handle because of its complexity. “We are targeting the mammalian cerebral cortex, which is the newest in the context of evolution, and which controls the highest functions. This is quite an exciting challenge.”
Figure 4: Photomicrograph of part of the mouse cerebral cortex.
An example analysis of the arrangement of neurons. The nuclei of all cells in the visual field are shown in green. Purple represents mRNAs expressed in a particular type of neuron.
A basic problem arises, however, in choosing the approach to handling the neural circuits of the cerebral cortex, which has the greatest complexity. The cerebral cortex is a sheet of cells covering the surface of the cerebrum. Its functions vary between its different areas: the visual, auditory, motor, and association areas. The last of these undertakes higher levels of information processing by integrating the signals from the other areas. All these areas are about 2 mm thick, and in them a wide variety of neurons are basically stacked in six layers (Fig. 1). “The neurons in the cerebral cortex are highly diverse, involving a large number of types, with complex binding modes, and the details of the structure of their circuits are for the most part a riddle. However, it has been reported that cells with similar functions often arrange themselves to form a large number of thin columns. This suggests the presence of a repeat structure in the cerebral cortex network,” says Hosoya. “If so, it may be possible to understand the cerebral cortex as a whole by investigating in great detail one unit that repeats and by understanding its functions, rather than by handling the massive neural network as a whole.”
Figure 5: Patch clamp apparatus.
A hole is made in a neuron by pressing a glass electrode against the cell, and through it electric signals from the cell are measured. This apparatus is used to clarify cerebral cortex neural circuits and analyze their functions. “Glass electrodes have long been used as a tool to determine the activity of neurons,” says Hosoya. “I would like to develop a new method in cooperation with one of the engineering research laboratories at RIKEN.”
Hosoya showed a photograph (Fig. 4), saying, “We are working on how the various types of neurons are arranged in the mouse cerebral cortex. We detect the messenger RNAs of genes expressed in a particular type of cell, and analyze their spatial distributions in detail using a computer. Although neurons have been thought to have a basically random distribution, our analyses have indicated a variety of regularities that have so far not been known. We are examining these regularities in the expectation of finding repeat structures. We will investigate the basic unit comprising this repeat structure, if it exists, in great detail to deepen our understanding of the cerebral cortex as a whole.”
Hosoya wants to link his work on the brain to understanding the mind. “I want to discover what could be termed ‘a unit of mind’ in the same way that the cell is the unit of an organism. I am now conducting research to this end.”
As Hosoya proceeds with his research, he feels that the brain’s neural network, which as a university student he had thought impossible to handle, is being clarified little by little.