Like the brain, the study of the brain has an extensive hierarchical structure that incorporates the molecular biology of synapses near the base as well as theoretical modeling of cognition at higher levels. The interactions between all these layers are surprisingly rich. Working near the base, we want to understand the characteristics of relatively local neuronal circuits. Our goal is to define the structures of local circuits, their expression patterns, and the information they convey through those patterns.
As with brain science, local circuits have several layers of significance. First we need to understand how the functional mechanisms of the receptor fields of various neurons are formed. Accurately obtaining this information requires precise control of the input; however, whether one is working with humans or non-human animals, this control is difficult. Local circuits are more easily observed when complicated operational processes are involved. Nonetheless, understanding the intricate operations of local circuits may shed insight on conventional behavioral tests and thereby make interpreting those results more meaningful. Yet, despite the acknowledged importance of understanding local circuits, our understanding is far less developed than work in molecular biology at the level of a single neuron or synapse.
To expand the available knowledge of local circuits, we are taking two approaches. For one, we study inputs and outputs that can be easily observed to identify those properties common to all neuronal circuits. For the other, we develop new techniques to establish a foothold within those circuits to investigate more complex activities.
The former approach requires precise analysis of complicated input/output characteristics. For this, we used retina from vertebrae. Information processing in this retina is unidirectional, so mapping input/output responses reveals most functions. With this retinam, it is also possible to extract a fully functional circuit and apply input via computer. In this way, a search of functions, even complex ones, becomes possible. Clearly, the retina is an excellent model of local circuits that enables us investigate precise processes and retrieve information on expression patterns.
We know that precise temporal patterns of neuronal spikes and spike frequency contain information. In our studies, we found that the retina can mediate processing according to the image patterns it receives. Since information patterning and processing are properties for a wide range of neuronal circuits, we believe that they can serve as prototypes for the properties of neuronal circuits in general.
Work using the second approach has just started. While we are still evaluating various possibilities, we hope that early results will lead to a publication.
As new theories and technologies are developed, a diverse array of ideas about the functions of local circuits have been tested. An individual, or small group, with the ability to acquire this knowledge will contribute significantly to the advancement of brain science. Our work at RIKEN BSI, a unique place where theoretical and practical discussions with many different researchers is possible, may provide such a role. Anyone who shares our passion and commitment is invited to visit, the door is open...
