Naoyuki Sato, Future University Hakodate

Top Page

Research

Publications

Links
-2008 2009-
Connectome-based neural network simulation on EEG traveling waves
This study aimed to evaluate the functional contributions of traveling waves using connectome?based network modeling. According to the simulation results, cortical traveling waves were thought to reflect the intrinsic frequency?dependent hierarchical sequencing of local regions, global traveling waves sequence the set of large?scale cortical networks, and local traveling waves sequence local regions within individual cortical networks.

Analysis of EEG coherence based on distributed semantic network
Our recent computational simulation using biologically plausible network demonstrated the coupling between subpopulation network structure and local field potential (LFP) coherence. This finding has importance in developing an analytical method for evaluating subpopulation network structure based on signals generated in larger spatial scales, such as in electroencephalogram (EEG) and electrocorticogram (ECoG). This method can be used for various types of networks associated with distributed representation, and it is compatible with recent machine learning techniques for visual and auditory signals, or natural languages. The present study aimed to develop a theoretical formulation of the method for estimating the subpopulation network, and to test it using experimental data.

EEG-fMRI measurement during natural reading of literature
Context-dependent memory encoding during natural reading of literature was evaluated using electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) based on a subsequent memory paradigm. Subsequent memory was quantified by semantic correlation between the read text and reports subsequently written by the participants, based on a recent natural language processing procedure. Our results demonstrated a positive correlation between subsequent memory and fixation-related EEG and a negative correlation with fMRI activity. Sentence-length and paragraph-length processes were associated with regions belonging to the salience network and the default mode network, respectively. This is the first demonstration that memory encoding during literature reading is associated with large-scale network deactivations, which might reflect the coordination of a range of voluntary processes during reading.

A study of memory encoding during reading of literature
Dynamics of hippocampal phase precession was applied to the experience of text reading that is characterized by a series of semantic events in a temporal context. Input to the model was given by word sequences determined by eye fxation data in a reading experiment. The phase precession was found to form asymmetric synaptic connections based on both of temporal order and hierarchical relationship (or temporally inclusion) in the word sequence and such asymmetry was found necessary for subsequent recall. These results suggested a general role of theta phase precession in the formation of episodic memory.

Computer simulation on the relationship between neural oscillation and firings
In order to evaluate local field potentials (LFP) coherence and neural firing activities in each area, a neural network of two cortical areas using leaky-integrate fire model was evaluated. Results showed that the spatial consistency of neural firing in the two areas was strongly correlated with LFP coherence. It is suggested that the spatial consistency of neural firing is essential for regulating long-range LFP synchronization, which would facilitate neuronal integration processes over multiple cortical areas.

A study of the brain dynamics during memory encoding by using a visual flicker
Visual flicker is known to induce electroencephalography (EEG) at a frequency of the flicker. The induced EEG on the scalp reflects neural synchronization of a number of brain regions, thus it can be expected that the visual flicker is available for a probe to synchronization network of multiple regions. In this study, a neural network model is proposed to analyze EEG synchronization in the cortico--hippocampal network in relationship to the flicker. According to results, it is predicted that the flicker phase--locking to the frontal EEG on the scalp can modulate EEG synchronization in the cortico--hippocampal network.

A theoretical study of mental image for view estimation at different viewpoint
Hippocampus damage is known to associate with disability of spatial recognition form a novel viewpoint. As an implementation of such hippocampal related function, two hypothesized mechanism was evaluated by using computer simulation. First mechanism is a scene memory with a scene expansion/reduction mechanism. Second mechanism is a scene memory with inverse-perspective mapping mechanism. Both mechanisms areexpected to be cooperated in the emergeence of novel spatial images at arbitral viewpoints.

A theory-experiment combined analysis to understand the episodic memory
Coding of visual experiences in the episodic memory is essential for understanding the episodic memory and for designing devices to support our memories. In this study, several coding were evaluated by those predictionabilities of human recalls of visual scene memory. In results, semantic coding was found to have better predictionability. The semantic coding is considered to be dominantly used to store the episodic memory based on visual experiences.