Yoshiki Kashimori

An olfactory recognition model based on spatio-temporal encoding of odor quality in the olfactory bulb

Abstract. In order to study the problem how the olfactory neural system processes the odorant molecular information for constructing the olfactory image of each object, we present a dynamic model of the olfactory bulb constructed on the basis of well-established experimental and theoretical results. The information relevant to a single odor, i.e. its constituent odorant molecules and their mixing ratios, are encoded into a spatio-temporal pattern of neural activity in the olfactory bulb, where the activity pattern corresponds to a limit cycle attractor in the mitral cell network. The spatio-temporal pattern consists of a temporal sequence of spatial firing patterns: each constituent molecule is encoded into a single spatial pattern, and the order of magnitude of the mixing ratio is encoded into the temporal sequence. The formation of a limit cycle attractor under the application of a novel odor is carried out based on the intensity-to-time-delay encoding scheme. The dynamic state of the olfactory bulb, which has learned many odors, becomes a randomly itinerant state in which the current firing state of the bulb itinerates randomly among limit cycle attractors corresponding to the learned odors. The recognition of an odor is generated by the dynamic transition in the network from the randomly itinerant state to a limit cycle attractor state relevant to the odor, where the transition is induced by the short-term synaptic changes made according to the Hebbian rule under the application of the odor stimulus.

Role of itinerancy among attractors as dynamical map in distributed coding scheme

A basic frame, based on which certain features of sensory stimuli can be extracted systematically in a distributed coding scheme, has not been clarified yet. This paper proposes that the basic frame can be a dynamical map represented by itinerancy among attractors. Features of entities are encoded by attractors of neural networks. Relations between the features in each modality are mapped on dynamical links between the attractors in each network relevant to each modality. The itinerant states, in which the network dynamic state itinerates chaotically or cyclically among the attractors, can work as dynamical maps. The recognition of features is carried out by a phase transition from an itinerant state to a constituent attractor. The phase transition is induced by a short-term synaptic change based on the Hebbian rule under application of a relevant stimulation. A theta-like global oscillation is necessary for self-organized formation of the chaotically itinerant state.

Model of P- and T-electroreceptors of weakly electric fish.

To clarify the microscopic mechanisms by which P- and T-receptors encode amplitude modulation and zero crossing time of jamming signals, we present a model of P- and T-receptors based on their physiological and anatomical properties. The model consists of a receptor cell, supporting cells, and an afferent nerve fiber. The basal membrane of the receptor cell includes voltage-sensitive Ca2+ channels, Ca(2+)-activated K+ channels, and leak channels of Na+, K+, and Cl-. The driving force of potential change under stimulation is generated by the voltage-sensitive Ca2+ channels, and the suppressing force of the change is generated by Ca(2+)-activated K+ channels. It has been shown that in T-receptor cells the driving force is much stronger than the suppressing force, whereas in P-receptor cells the driving force is comparable with the suppressing force. The difference in various kinds of response properties between P- and T-receptors have been consistently explained based on the difference in the relative strengths of the driving and suppressing forces between P- and T-receptor cells. The response properties considered are encoding function, probability of firing of afferent nerve, pattern of damped oscillation, shape of tuning curves, values of the optimum frequency, and response latency.

A Hierarchical Dynamical Map as a Basic Frame for Cortical Mapping and Its Application to Priming

A hierarchical dynamical map is proposed as the basic framework for sensory cortical mapping. To show how the hierarchical dynamical map works in cognitive processes, we applied it to a typical cognitive task known as priming, in which cognitive performance is facilitated as a consequence of prior experience. Prior to the priming task, the network memorizes a sensory scene containing multiple objects presented simultaneously using a hierarchical dynamical map. Each object is composed of different sensory features. The hierarchical dynamical map presented here is formed by random itinerancy among limit-cycle attractors into which these objects are encoded. Each limit-cycle attractor contains multiple point attractors into which elemental features belonging to the same object are encoded. When a feature stimulus is presented as a priming cue, the network state is changed from the itinerant state to a limit-cycle attractor relevant to the priming cue. After a short priming period, the network state reverts to the itinerant state. Under application of the test cue, consisting of some feature belonging to the object relevant to the priming cue and fragments of features belonging to others, the network state is changed to a limit-cycle attractor and finally to a point attractor relevant to the target feature. This process is considered as the identification of the target. The model consistently reproduces various observed results for priming processes such as the difference in identification time between cross-modality and within-modality priming tasks, the effect of interval between priming cue and test cue on identification time, the effect of priming duration on the time, and the effect of repetition of the same priming task on neural activity.

Spatiotemporal burst coding for extracting features of spatiotemporally varying stimuli

Encoding features of spatiotemporally varying stimuli is quite important for understanding the neural mechanisms of various sensory coding. Temporal coding can encode features of time-varying stimulus, and population coding with temporal coding is adequate for encoding spatiotemporal correlation of stimulus features into spatiotemporal activity of neurons. However, little is known about how spatiotemporal features of stimulus are encoded by spatiotemporal property of neural activity. To address this issue, we propose here a population coding with burst spikes, called here spatiotemporal burst (STB) coding. In STB coding, the temporal variation of stimuli is encoded by the precise onset timing of burst spike, and the spatiotemporal correlation of stimuli is emphasized by one specific aspect of burst firing, or spike packet followed by silent interval. To show concretely the role of STB coding, we study the electrosensory system of a weakly electric fish. Weakly electric fish must perceive the information about an object nearby by analyzing spatiotemporal modulations of electric field around it. On the basis of well-characterized circuitry, we constructed a neural network model of the electrosensory system. Here we show that STB coding encodes well the information of object distance and size by extracting the spatiotemporal correlation of the distorted electric field. The burst activity of electrosensory neurons is also affected by feedback signals through synaptic plasticity. We show that the control of burst activity caused by the synaptic plasticity leads to extracting the stimulus features depending on the stimulus context. Our results suggest that sensory systems use burst spikes as a unit of sensory coding in order to extract spatiotemporal features of stimuli from spatially distributed stimuli.

Effect of Syncytium Structure of Receptor Systems on Stochastic Resonance Induced by Chaotic Potential Fluctuation

To study a role of syncytium structure of sensory receptor systems in the detection of weak signals through stochastic resonance, we present a model of a receptor system with syncytium structure in which receptor cells are interconnected by gap junctions. The apical membrane of each cell includes two kinds of ion channels whose gating processes are described by the deterministic model. The membrane potential of each cell fluctuates chaotically or periodically, depending on the dynamical state of collective channel gating. The chaotic fluctuation of membrane potential acts as internal noise for the stochastic resonance. The detection ability of the system increases as the electric conductance between adjacent cells generated by the gap junction increases. This effect of gap junctions arises mainly from the fact that the synchronization of chaotic fluctuation of membrane potential between the receptor cells is strengthened as the density of gap junctions is increased.

A neural mechanism of hyperaccurate detection of phase advance and delay in the jamming avoidance response of weakly electric fish.

Abstract. The weakly electric fish Eigenmannia can detect the phase difference between a jamming signal and its own signal down to 1 s. To clarify the neuronal mechanism of this hyperaccurate detection of phase difference, we present a neural network model of the torus of the midbrain which plays an essential role in the detection of phase advances and delays. The small-cell model functions as a coincidence detector and can discriminate a time difference of more than 100 s. The torus model consists of laminae 6 and 8. The model of lamina 6 is made with multiple encoding units, each of which consists of a single linear array of small cells and a single giant cell. The encoding unit encodes the phase difference into its spatio-temporal firing pattern. The spatially random distribution of small cells in each encoding unit improves the encoding ability of phase modulation. The neurons in lamina 8 can discriminate the phase advance and delay of jamming electric organ discharges (EODs) compared with the phase of the fishs own EOD by integrating simultaneously the outputs from multiple encoding units in lamina 6. The discrimination accuracy of the feature-detection neurons is of the order of 1 s. The neuronal mechanism generating this hyperacuity arises from the spatial feature of the system that the innervation sites of small cells in different encoding units are distributed randomly and differently on the dendrites of single feature-detection neurons. The mechanism is similar to that of noise-enhanced information transmission.

A model describing collective behaviors of pedestrians with various personalities in danger situations

We present a model describing motions of individual pedestrians to study collective behaviors of pedestrians in various situations such as normal or emergency situations. Each individual pedestrian has ability to decide own desired action depending on circumstances and own personality. In the present work the personality of pedestrian is expressed as patient and impatient character. The main frame of the present model is constructed based on the social force model, and the individual ability to decide own desired action is represented by the counterpropagation neural network. The combination of neural network and dynamic model makes motions of pedestrians more natural and reliable in various situations. It has been shown from the computer simulation of passing of many pedestrians through a road that there exists an optimal ratio of patient persons number to impatient persons number for the passage time and amenity of pedestrians.

Neural mechanism of dynamic responses of neurons in inferior temporal cortex in face perception

Understanding the neural mechanisms of object and face recognition is one of the fundamental challenges of visual neuroscience. The neurons in inferior temporal (IT) cortex have been reported to exhibit dynamic responses to face stimuli. However, little is known about how the dynamic properties of IT neurons emerge in the face information processing. To address this issue, we made a model of IT cortex, which performs face perception via an interaction between different IT networks. The model was based on the face information processed by three resolution maps in early visual areas. The network model of IT cortex consists of four kinds of networks, in which the information about a whole face is combined with the information about its face parts and their arrangements. We show here that the learning of face stimuli makes the functional connections between these IT networks, causing a high spike correlation of IT neuron pairs. A dynamic property of subthreshold membrane potential of IT neuron, produced by Hodgkin–Huxley model, enables the coordination of temporal information without changing the firing rate, providing the basis of the mechanism underlying face perception. We show also that the hierarchical processing of face information allows IT cortex to perform a “coarse-to-fine” processing of face information. The results presented here seem to be compatible with experimental data about dynamic properties of IT neurons.

A neural model of electrosensory system making electrolocation of weakly electric fish

Abstract We address the question of how the fish recognizes the location of objects at a given position, independently of the electric properties of objects, although the distortion pattern of electric field is varied depending on the electric property. To know how the electric field generated by fishs electric organ discharge (EOD) is spatially and temporally distorted by resistive and capacitive objects, we calculated the electric field using a reasonable model of fish body. P and T electroreceptors distributed on the fish body surface respond to the amplitude and phase modulations of EOD, respectively, and propagate their output signals to electrosensory lateral-line lobe (ELL) and torus semicircularis (TS), respectively. As the first step to clarify the neural mechanism of electrolocation, we made a neural network model of electrosensory system consisting of electroreceptors, afferent nerves, ELL, and TS, and investigated how the amplitude and phase modulations are encoded in ELL and TS, respectively.