It is believed that attributes of objects, such as colors or shapes, are expressed with time series of neuron spikes in the brain. We do not have, however, any answer to the following question “what are the neuron dynamics working at the processes and how is it able to express the processes?”. To solve this problem, we need to study the neural coding mechanism where necessary information is expressed by the spike sequences. In this research, we propose a neural coding method which uses the morphoelectorotonic transform theory to formalize the cell assembly. Then, a simulation model to solve the binding problem is constructed by using this coding model and its effectiveness is verified.
The typicality effect is one of important findings on concept hierarchies in the field of cognitive psychology. The followings are some of well-known phenomena: if asked to list up all the instances of a concept, we recall them in descending order of their typicality; if asked whether an instance belongs to a concept, we can judge faster as its typicality is higher. As the first step toward a neuronal basis of the typicality effect, this paper proposes a neural netwrok model which recalls a typical instance of a concept with a higher probability and also in a shorter period of recalling time than an atypical one.
A lot of researches try to realize a useful system in real world, which is mimicing from information processing system of living thing, especially animals. The features of information processing system of an animal are adaptability strong against disturbance, and autonomous emergency of an information processing rule, but there are many unknown portions about the mechanism yet. We reconstruct a dissociated culture system of neuronal network using the nerve cell taken out from the hippocampal region of a rat brain, and elucidate whether the information processing system can perform perception for the outer world and whether the origin of intelligence emerges in the system. For the first step, we applied feed-back stimulation into the living neuronal network. When the feed-back stimulation system detects synchronous action potentials at 2 recording sites simultaneously, it applies single stimulation. The feed-back stimulation evoked drastic change of spatiotemporal pattern of the activities of action potentials. The result suggested that the living neuronal network can perform modification of their own responding rules against outer world by interaction. Now we are developping a system for the living neuronal network interacting to outerworld mediated by miniature moving robot.
Assessment of hemispheric dominance for language is one of principal factors of pre-surgical planning for neurosurgery, researches on higher brain function, etc. Conventionally, hemispheric dominance for language has been assessed by using Wada test, functional magnetic resonance imaging and so on. However, these methods are invasive for subjects or require large-scale imaging modalities. This paper proposes a new method for assessing language dominance using near-infrared spectroscopy (NIRS), which is a non-invasive and small-scale modality. The proposed method introduces fuzzy degree which belongs to language area for each measuring point into NIRS data analysis. The fuzzy degree estimates which brain function area the measuring point is assigned. By using the estimated fuzzy degrees, we can calculate language laterality index using only the measuring points that arranged on the language area. The performance of the proposed method has been evaluated by applying it to fourteen healthy subjects with multiple trials.
To investigate how the human brain processes words and symbols with the same meanings, the following four words (Kanji: Chinese characters using in Japanese language) and four symbols (arrows) were utilized as visual stimuli: four words for upward, downward, leftward and rightward (??, ??, ?? and ??, respectively) and four arrow symbols for upward, downward, leftward and rightward (↑, ↓, ← and →, respectively). Electroencephalography (EEG) was performed while subjects were observing these visual stimuli. In each experiment, each visual stimulus was presented 30 times, for a total of 240 presentations of the eight visual stimuli. Subjects comprised four women aged from 20 to 21. EEGs were averaged for each stimulus type, and event-related potentials (ERPs) were determined. Tendencies in ERPs were compared, and marked changes in amplitude were seen near a latency of 420 ms for ?? (up) and ?? (down) and 500 ms for ?? (left) and ?? (right). Marked changes in amplitude were seen near a latency of 500 ms for all arrow symbols. When comparing ERPs between kanji characters and arrow symbols, differences in latency were noted, as were similarities in marked amplitude changes. When comparing ERPs between kanji characters and arrow symbols with opposing meanings, peak latencies for marked amplitude changes were predominantly similar, but polarities were opposite. Peak latency of ERPs was subjected to equivalent current dipole source localization (ECDL). ECD was estimated at a latency of around 110 ms in the MT field and then around 300 ms in the precentral gyrus. No marked differences in this tendency were noted among the 8 stimuli. After ECD was estimated in the precentral gyrus, with the kanji characters, ECD was estimated in the right middle temporal gyrus regardless of direction. ECD was then estimated in areas related to language, such as the Wernicke field in the left middle temporal gyrus, the left angular gyrus and the left lingual gyrus. ECD was later estimated in the left middle frontal gyrus, left inferior frontal gyrus and prefrontal area. ECD was estimated in the precentral gyrus just before the amplitude of ERPs changed markedly. With arrow symbols, ECD was estimated in the right middle temporal gyrus, then ECD was estimated in areas related to working memory for spatial perception, such as the right inferior or right middle frontal gyrus. Then, as with kanji characters, ECD was estimated in the prefrontal area and precentral gyrus. When comparing visual stimuli with opposite meanings, ECD was estimated in almost the same area at latencies where ERPs exhibited opposite peaks, but moments differed.
Abstract: The TAM (Topographic Attentive Mapping) Network based on a biologically-motivated neural network is an especially effective model. When the network makes an incorrect output prediction, the attentional feedback circuit modulates the learning rates, and adds a node to the category layer in order to improve the network's prediction accuracy. In this paper, Gabor type receptive field is formulated in the input layer of TAM network. By using Gabor filtering, we can detect orientation selectivity of visual images, and simulate the mechanism of early vision of brain as a computational model.
The adaptive consensus formation models are useful for obtaining harmonious consensus of a group when all members of the group pursue their individual preference. In the process of consensus formation, the agents access to macro information and adapts their preference to the overall group preference. In this paper, we propose and analyze some quantitative measures of the difference between the individual preferences and the agreed preference. Moreover, we introduce a model of the agents with the different speeds of adaption and analyze its effects to consensus formation by means of simulation. We also incorporate a kind of time constraint into our model and see its effects to consensus formation.
Condition search is mainly used by the data retrieval technique. Since it is difficult to judge which is adjusted when there are two or more search conditions, we have to repeat adjustment re-search repeatedly. Furthermore, depending on a database, even a condition setup may become difficult if the searcher has not familiar with the database. A real estate housing database is considered as an example. The conditions of a housing of wanting to live have a house rent is cheap, a building is new,. a housing space is large, close to a landmark, etc, and the housing which fills these all is the most desirable. However, the concrete value of an ambiguous expression, such as “cheap”, “new”, “large”, “close”, changes greatly with areas which the housing locates. The searcher has to learn these relations through searching. The same problem exists in many fields including job information search and used car search. In order to solve the above-mentioned problem, new method was proposed that, display feature of information in 2 dimensional map using Self-Organizing Map technique, clustering data on the map by index data, and adjusts conditions visually on the map. As a result of experiments using actual real estate housing data, it was shown that this method is especially effective to look for information on the ambiguous conditions.
In this paper, we propose a method for evaluating and/or determining the blurred quantity and range of the degraded image by the texture feature quantities of the coefficients in wavelet transform which decomposes into the multiple hierarchical resolution level. The evaluation process of the proposal method is composed of two steps. The first stage is decomposition into multiple hierarchical components by multi-resolution analysis based on wavelet transform. The next stage is calculation of the texture feature quantities from co-occurrence matrix of the decomposed components. Then the blurred quantity and range are evaluated and/or determined by the texture feature quantities. When there is the standard image of known blurred quantity, the blurred level and/or range of the image are evaluated from the comparison with the standard image. On the other, when there is not the standard image, the level and/or range of the degraded image are estimated by the ratio of the change for the addition of the locale degradation. We confirmed that the texture feature quantities in the wavelet coefficient of degraded image was effective to evaluate the image degradation level. And then the contrast quantity within texture feature quantities was most suitable for the blurred quantity evaluation from the experiments which applied this proposal method in 4 kinds of images. We also confirmed that the proposal method was effective in evaluating the degraded image under the noise existence.
From the viewpoint of safety and noise reduction, the study on anti-lock braking systems of railways is very important. But, the skid of wheels of railways is strongly uncertain phenomenon so that it is too hard to control it and there are no general guides to construct anti-lock braking systems. By the way, it is well-known that fuzzy reasoning is useful to controlled systems with strong uncertainty. Thus, we can consider that fuzzy reasoning is also available to braking systems of railways. In this paper, we propose a new system with fuzzy reasoning, which is the braking system for train sets into which ABS with fuzzy reasoning is integrated. Moreover, through numerical simulations and practical tests, we show that the proposed system can shorten braking distance and reduce wheel damages.
How to proceed the higher education for fostering researchers is ought to be one of the whole point s which would inevitably affects and determines the future of a knowledge based country. In graduate school in universities, specialization knowledges are given to the students. After graduation, many of them will spend the research life as job whereby they will face not only research as the main job but also chore which may indirectly support research activities. Thus, so to speak harmonious fosterage of human resources mentioned here deals with harmonious education of specialization knowledge and support knowledge, whose vision is the target of the consciousness survey. This paper discusses how valid university teachers and university students imagine the vision will be, and the statistical analysis of the data. It concludes, that students are more interested in ‘distant educational method’ than teachers which is one factor axis, that students have more ‘interest in the vision of harmonious fosterage of human resources’ than that teachers have which is another factor axis, and so forth.