We assess how well the recently created MRI reconstruction technique, Multi-Resolution Convolutional Neural Network (MRCNN), performs in the core medical vision field (classification). The primary goal of MRCNN is to identify the best k-space undersampling patterns to accelerate the MRI. In this study, we use the Figshare brain tumor dataset for MRI classification with 3064 T1-weighted contrast-enhanced MRI (CE-MRI) over three categories: meningioma, glioma, and pituitary tumors. We apply MRCNN to the dataset, which is a method to reconstruct high-quality images from under-sampled k-space signals. Next, we employ the pre-trained VGG16 model, which is a Deep Neural Network (DNN) based image classifier to the MRCNN restored MRIs to classify the brain tumors. Our experiments showed that in the case of MRCNN restored data, the proposed brain tumor classifier achieved 92.79% classification accuracy for a 10% sampling rate, which is slightly higher than that of SRCNN, MoDL, and Zero-filling methods have 91.89%, 91.89%, and 90.98% respectively. Note that our classifier was trained using the dataset consisting of the images with full sampling and their labels, which can be regarded as a model of the usual human diagnostician. Hence our results would suggest MRCNN is useful for human diagnosis. In conclusion, MRCNN significantly enhances the accuracy of the brain tumor classification system based on the tumor location using under-sampled k-space signals.

We characterized prion disease by comparing brain functional connectivity network (BFCN), which were constructed by 16-ch scalp-recorded electroencephalograms (EEGs). The connectivity between each pair of nodes (electrodes) were computed by synchronization likelihood (SL). The BFCN was applied to graph theory to discriminate prion disease patients from healthy elderlies and dementia groups.

We propose an idea generation support system known as the “GUNGEN-Heartbeat” that uses heartbeat variations for creating high quality ideas during brainstorming. This system shows “An indication of a check list” or “An indication to promote deep breathing” at time beyond a value with variance of heart rates. We also carried out comparison experiments to evaluate the usefulness of the system.

Machine learning algorithms are becoming more and more popular in current era. Data preprocessing especially feature selection is helpful for improving the performance of those algorithms. A new powerful feature selection algorithm is proposed. It combines the advantages of ant colony optimization and brain storm optimization which simulates the behavior of human beings. Six classical datasets and five state-of-art algorithms are used to make a comparison with our algorithm on binary classification problems. The results on accuracy, percent rate, recall rate, and F1 measures show that the developed algorithm is more excellent. Besides, it is no more complex than the compared approaches.

Brain-computer interfaces (BCIs) have been used by users to convey their intentions directly with brain signals. For example, a spelling system that uses EEGs allows letters on a display to be selected. In comparison, previous studies have investigated decoding speech information such as syllables, words from single-trial brain signals during speech comprehension, or articulatory imagination. Such decoding realizes speech recognition with a relatively short time-lag and without relying on a display. Previous magnetoencephalogram (MEG) research showed that a template matching method could be used to classify three English sentences by using phase patterns in theta oscillations. This method is based on the synchronization between speech rhythms and neural oscillations during speech processing, that is, theta oscillations synchronized with syllabic rhythms and low-gamma oscillations with phonemic rhythms. The present study aimed to approximate this classification method to a BCI application. To this end, (1) we investigated the performance of the EEG-based classification of three Japanese sentences and (2) evaluated the generalizability of our models to other different users. For the purpose of improving accuracy, (3) we investigated the performances of four classifiers: template matching (baseline), logistic regression, support vector machine, and random forest. In addition, (4) we propose using novel features including phase patterns in a higher frequency range. Our proposed features were constructed in order to capture synchronization in a low-gamma band, that is, (i) phases in EEG oscillations in the range of 2-50 Hz from all electrodes used for measuring EEG data (all) and (ii) phases selected on the basis of feature importance (selected). The classification results showed that, except for random forest, most classifiers perform similarly. Our proposed features improved the classification accuracy with statistical significance compared with a baseline feature, which is a phase pattern in neural oscillations in the range of 4-8 Hz from the right hemisphere. The best mean accuracy across folds was 55.9% using template matching trained by all features. We concluded that the use of phase information in a higher frequency band improves the performance of EEG-based sentence classification and that this model is applicable to other different users.

Traffic engineering refers to techniques to accommodate traffic efficiently by dynamically configuring traffic routes so as to adjust to changes in traffic. If traffic changes frequently and drastically, the interval of route reconfiguration should be short. However, with shorter intervals, obtaining traffic information is problematic. To calculate a suitable route, accurate traffic information of the whole network must be gathered. This is difficult in short intervals, owing to the overhead incurred to monitor and collect traffic information. In this paper, we propose a framework for traffic engineering in cases where only partial traffic information can be obtained in each time slot. The proposed framework is inspired by the human brain, and uses conditional probability to make decisions. In this framework, a controller is deployed to (1) obtain a limited amount of traffic information, (2) estimate and predict the probability distribution of the traffic, (3) configure routes considering the probability distribution of future predicted traffic, and (4) select traffic that should be monitored during the next period considering the system performance yielded by route reconfiguration. We evaluate our framework with a simulation. The results demonstrate that our framework improves the efficiency of traffic accommodation even when only partial traffic information is monitored during each time slot.

A new hybrid brain-computer interface (BCI), which is based on sequential controls by eye tracking and steady-state visual evoked potentials (SSVEPs), has been proposed for high-speed spelling in virtual reality (VR) with a 40-target virtual keyboard. During target selection, gaze point was first detected by an eye-tracking accessory. A 4-target block was then selected for further target selection by a 4-class SSVEP BCI. The system can type at a speed of 1.25 character/sec in a cue-guided target selection task. Online experiments on three subjects achieved an averaged information transfer rate (ITR) of 360.7 bits/min.

To enable the vision of precision medicine, evidence-based medicine is the key element. Understanding the natural history of complex diseases like brain aneurysm and particularly investigating the evidences of its rupture risk factors relies on the existence of semantic-enabled data preparation technology to conduct clinical trials, survival analysis and outcome prediction. For personalized medicine in the field of neurological diseases, it is very important that multiple health organizations coordinate and cooperate to conduct evidence based observational studies. Without the means of automating the process of privacy and semantic-enabled data preparation to conduct observational studies at intra-organizational level would require months to manually prepare the data. Therefore, this paper proposes a semantic and privacy enabled, multi-party data preparation architecture and a four-tiered semantic similarity algorithm. Evaluation shows that proposed algorithm achieves a precision of 79%, high recall at 83% and F-measure of 81%.

Functional near-infrared spectroscopy (fNIRS) is a noninvasive neuroimaging technique, suitable for measurement during motor learning. However, effects of contamination by systemic artifacts derived from the scalp layer on learning-related fNIRS signals remain unclear. Here we used fNIRS to measure activity of sensorimotor regions while participants performed a visuomotor task. The comparison of results using a general linear model with and without systemic artifact removal shows that systemic artifact removal can improve detection of learning-related activity in sensorimotor regions, suggesting the importance of removal of systemic artifacts on learning-related cerebral activity.

To help elderly and physically disabled people to become self-reliant in daily life such as at home or a health clinic, we have developed a network-type brain machine interface (BMI) system called “network BMI” to control real-world actuators like wheelchairs based on human intention measured by a portable brain measurement system. In this paper, we introduce the technologies for achieving the network BMI system to support activities of daily living.

A brain-computer interface (BCI) translates the brain activity into commands to control external devices. P300 speller based character recognition is an important kind of application system in BCI. In this paper, we propose a framework to integrate channel correlation analysis into P300 detection. This work is distinguished by two key contributions. First, a coefficient matrix is introduced and constructed for multiple channels with the elements indicating channel correlations. Agglomerative clustering is applied to group correlated channels. Second, the statistics of central tendency are used to fuse the information of correlated channels and generate virtual channels. The generated virtual channels can extend the EEG signals and lift up the signal-to-noise ratio. The correlated features from virtual channels are combined with original signals for classification and the outputs of discriminative classifier are used to determine the characters for spelling. Experimental results prove the effectiveness and efficiency of the channel correlation analysis based framework. Compared with the state-of-the-art, the recognition rate was increased by both 6% with 5 and 10 epochs by the proposed framework.

Electroencephalography (EEG) and magnetoencephalography (MEG) measure the brain signal from spatially-distributed electrodes. In order to detect event-related synchronization and desynchronization (ERS/ERD), which are utilized for brain-computer/machine interfaces (BCI/BMI), spatial filtering techniques are often used. Common spatial potential (CSP) filtering and its extensions which are the spatial filtering methods have been widely used for BCIs. CSP transforms brain signals that have a spatial and temporal index into vectors via a covariance representation. However, the variance-covariance structure is essentially different from the vector space, and not all the information can be transformed into an element of the vector structure. Grassmannian embedding methods, therefore, have been proposed to utilize the variance-covariance structure of variational patterns. In this paper, we propose a metric learning method to classify the brain signal utilizing the covariance structure. We embed the brain signal in the extended Grassmann manifold, and classify it on the manifold using the proposed metric. Due to this embedding, the pattern structure is fully utilized for the classification. We conducted an experiment using an open benchmark dataset and found that the proposed method exhibited a better performance than CSP and its extensions.

Computation in the brain is realized in complicated, heterogeneous, and extremely large-scale network of neurons. About a hundred billion neurons communicate with each other by action potentials called “spike firings” that are delivered to thousands of other neurons from each. Repeated integration and networking of these spike trains in the network finally form the substance of our cognition, perception, planning, and motor control. Beyond conventional views of neural network mechanisms, recent rapid advances in both experimental and theoretical neuroscience unveil that the brain is a dynamical system to actively treat environmental information rather passively process it. The brain utilizes internal dynamics to realize our resilient and efficient perception and behavior. In this paper, by considering similarities and differences of the brain and information networks, we discuss a possibility of information networks with brain-like continuing internal dynamics. We expect that the proposed networks efficiently realize context-dependent in-network processing. By introducing recent findings of neuroscience about dynamics of the brain, we argue validity and clues for implementation of the proposal.

Classification tasks in computer vision and brain-computer interface research have presented several applications such as biometrics and cognitive training. However, like in any other discipline, determining suitable representation of data has been challenging, and recent approaches have deviated from the familiar form of one vector for each data sample. This paper considers a kernel between vector sets, the mean polynomial kernel, motivated by recent studies where data are approximated by linear subspaces, in particular, methods that were formulated on Grassmann manifolds. This kernel takes a more general approach given that it can also support input data that can be modeled as a vector sequence, and not necessarily requiring it to be a linear subspace. We discuss how the kernel can be associated with the Projection kernel, a Grassmann kernel. Experimental results using face image sequences and physiological signal data show that the mean polynomial kernel surpasses existing subspace-based methods on Grassmann manifolds in terms of predictive performance and efficiency.

The relationship between different brain areas is characterized by functional networks through correlations of time series obtained from neuroimaging experiments. Due to its high spatial resolution, functional MRI data is commonly used for generating functional networks of the entire brain. These networks are comprised of the measurement points/channels as nodes and links are established if there is a correlation in the measured time series of these nodes. However, since the evaluation of correlation becomes more accurate with the length of the underlying time series, we construct in this paper functional networks from MEG data, which has a much higher time resolution than fMRI. We study in particular how the network topologies change in an experiment on ambiguity of words, where the subject first receives a priming word before being presented with an ambiguous or unambiguous target word.

Network structures can be found in almost any kind of natural or artificial systems as transport medium for communication between the respective nodes. In this paper we study certain key topological features of brain functional networks obtained from functional magnetic resonance imaging (fMRI) measurements. We compare complex network measures of the extracted topologies with those from Internet service providers (ISPs). Our goal is to identify important features which will be helpful in designing more robust and adaptive future information network architectures.

We have developed a portable NIRS-based optical BCI system that features a non-invasive, facile probe attachment and does not require muscle movement to control the target devices. The system consists of a 2-channel probe, a signal-processing unit, and an infrared-emission device, which measures the blood volume change in the participant's prefrontal cortex in a real time. We use the threshold logic as a switching technology, which transmits a control signal to a target device when the electrical waveforms exceed the pre-defined threshold. Eight healthy volunteers participated in the experiments and they could change the television channel or control the movement of a toy robot with average switching times of 11.5 ± 5.3 s and the hit rate was 83.3%. These trials suggest that this system provides a novel communication aid for people with motor disabilities.

We developed a novel movement-imagery-based brain-computer interface (BCI) for untrained subjects without employing machine learning techniques. The development of BCI consisted of several steps. First, spline Laplacian analysis was performed. Next, time-frequency analysis was applied to determine the optimal frequency range and latencies of the electroencephalograms (EEGs). Finally, trials were classified as right or left based on β-band event-related synchronization using the cumulative distribution function of pretrigger EEG noise. To test the performance of the BCI, EEGs during the execution and imagination of right/left wrist-bending movements were measured from 63 locations over the entire scalp using eight healthy subjects. The highest classification accuracies were 84.4% and 77.8% for real movements and their imageries, respectively. The accuracy is significantly higher than that of previously reported machine-learning-based BCIs in the movement imagery task (paired t-test, p < 0.05). It has also been demonstrated that the highest accuracy was achieved even though subjects had never participated in movement imageries.

We review recently obtained results for CMOS (Complementary Metal Oxide Semiconductor) imaging devices used in biomedical applications. The topics include dish type image sensors, deep-brain implantation devices for small animals, and retinal prosthesis devices. Fundamental device structures and their characteristics are described, and the results of in vivo experiments are presented.

The brain-machine interface (BMI) is a new method for man-machine interface, which enables us to control machines and to communicate with others, without input devices but directly using brain signals. Previously, we successfully developed a real time control system for operating a robot arm using brain-machine interfaces based on the brain surface electrodes, with the purpose of restoring motor and communication functions in severely disabled people such as amyotrophic lateral sclerosis patients. A fully-implantable wireless system is indispensable for the clinical application of invasive BMI in order to reduce the risk of infection. This system includes many new technologies such as two 64-channel integrated analog amplifier chips, a Bluetooth wireless data transfer circuit, a wirelessly rechargeable battery, 3 dimensional tissue-fitting high density electrodes, a titanium head casing, and a fluorine polymer body casing. This paper describes key features of the first prototype of the BMI system for clinical application.