2. Regulation mechanism of neuronal network of hippocampus

(a) Inhibitory regulation of signal flow in the hippocampal circuit

The lamellar hypothesis in the hippocampus is based on physiological data showing that stimulation of the entorhinal cortex activates only a limited number of CA1 cells arranged in a direction along the alvear fibers, nearly transversely to the longitudinal axis of the hippocampus. Classical anatomical observations with Golgi staining of hippocampal neurons, the simple tri-synaptic circuit (DG-CA3-CA1), are consistent with the lamellar hypothesis. However, recent anatomical work has revealed much richer synaptic connections between hippocampal neuron subfields (DG, CA3, CA2, CA1) and wider distribution of axons along the longitudinal axis of the hippocampus than those expected in the simple tri-synaptic concept. The presence of functional lamellar organization, despite wide spread of fibers, indicates that inhibitory mechanisms must suppress propagation of excitation beyond the lamellae. Here we addressed a new concept for the inhibitory regulation of signal flow in the hippocampal circuit, gating of signal flow, which might be regulated the motivating state of memory and regulate memory acquisition into the hippocampus. In order to examine this hypothesis, slices were made from the middle part of the rat hippocampal formation along the bundle of alvear fibers, as described in Andersensﾕs work which propose the lamellar hypothesis. Then the field potentials were recorded by electrophysiological and optical-imaging techniques.

(b) Supuramammilary neucleus and Hippocampus.

It has been known that the hippocampal activity is strongly related to the learning. The theta oscillations are recorded at the hippocampus of an animal, when the animal explore novel environment. We focus on the supramammillary nucleus (SuM) that is one of the hypothalamic nuclei, because the SuM neurons send axons to neurons in the hippocampal formation directly. We hypothesize that the SuM is the source of the hippocampal theta oscillations and controls the hippocampal memory system. To confirm our hypothesis we are analyzing animal behavior and the hippocampal theta oscillation using immunohistochemical analysis together. In addition to the experimental analysis, we study a machine learning theory, which is used for the robot control, to construct a computational brain model that is able to explain the learning behavior from the viewpoint of the neuron.