In order to simplify the procedure of its execution in our pendulum-test, an examiner had supported the lower-leg of the subject in a sitting situation manually until just before its start. The supporting force had however been apt to remain for a short time after its start. This study analyzed the influence of the residual supporting force on the lower-leg motion. It was assumed that the residual force decreased linearly with time. A simulator for the analysis was constructed by supplementing the function of the residual force to the existing one under such assumption. The influence on the lower-leg motion was investigated by comparing simulated waveforms with measured ones. Time when the force disappears was determined by means of reverse simulation of the measured waveform, so that both waveforms agreed as much as possible. As the results, the accuracy of the simulator increased dramatically.

At the beginning of the pendulum-test, some fluctuation is induced in the femur by its upward shifting with releasing of lower-leg. This study analyzes its influence upon lower-leg motion by simulations. The fluctuation is controlled by a viscoelastic element of the iliopsoas muscle including the sitting seat. Behavior of all over the leg is modeled as a double pendulum constructed based on the viscoelastic elements and an existing lower-leg sub-model. We analyzed the femoral motion and the lower-leg motion under the double pendulum model. It was seen from the results that a tiny and nonlinear fluctuation like that appeared in measured angular acceleration waveforms exists also in the simulated lower-leg ones, and that its amplitude increased with increases of the upward shifting or advance of spasticity. One of the urgent works is to investigate some way to remove the fluctuation from measured waveform.

It has been known from our experience that femur-tibia contact point shifting can occur inside the knee-joint in the pendulum-test by a tiny difference in lower-leg and return to the original position on the way of its execution. This study analyzes its influence on the lower-leg motions. Its shifting is expressed as an angle between the lower-leg and the straight line connecting the contact point with the knee-joint axis. The function of the return from the shifting is modeled by introducing imaginary viscoelastic elements, whose values are proportional to the magnitude of the angle. The lower-leg motion is simulated by the use of the existing pendulum-test model supplemented with the viscoelastic elements. It is shown from the results that, the larger the angle and/or the higher the spasticity of the subject is, the more the disturbance increases.

The extensive discomfort of heat and sweat donning plastic socket is the cause of discontinuing the limb prostheses use. This research's objective is to develop a system for body surface temperature rise control. Experiments were conducted to compare the thermal migration types' cooling effect with a cylindrical heat source model resembling the temperature range and transient properties of forearm body surface temperature. Peltier element and fan were installed for heat conduction and convection, and properties were measured. Temperature and energy flow were measured and evaluated of a system with vaporization heat of water's phase-transition applied by moistening the fabric installed on socket's interlayer and convection with fan. By comparing the experimental results to the non-countermeasure, vaporization heat with convection was most effective and the temperature rise was reduced significantly. By adjusting air flow, temperature rise of the socket's inner surface temperature was reduced to less than 2 deg C.

Patients with spinal cord injury have deteriorated thermoregulatory function because of the autonomic nervous system is damaged. Thus, body temperature control in summer for wheelchair athletes is an important issue from the viewpoints of improving training efficiency and preventing of accidents caused by heat stroke.To solve these problems, we have developed the thermoregulatory system that can be used by patients with spinal cord injured using a wheelchair. The developed system consisted of cooling jacket with cold water tubing. , Exercise loading experiment was performed to wear and not wear a cooling jacket by two healthy subjects. To measure the cooling effect of the system, the deep body temperature (DBT) as an index of body temperature was measured every 10 seconds and . DBT with the cooling jacket was declined after the exercise, while DBT without the cooling jacket was increased.

To support health care and rehabilitation in patient with spinal cord injury for a long time, this study is concerned with the initial attempt to evaluate daily cardiopulmonary function using a non-intrusive monitoring system. The capacitance-sensitive pressure sensor was located under the mattress and thus pulse and respiratory variability could be detected through an analysis of spectral ranges. These data could be led to a cloud server and browsed using an application software. The wearable sensor units were also attached to the subject's head, trunk, waist (near sacral spine), thigh, and calf, respectively, to evaluate the posture changes during rehabilitation. As the medical assessment, the pulse variability, the central apnea, and the effectiveness of phlegm expulsion care were successfully evaluated in patients (55 and 74 yrs.) in the bed. In another patient (21 yrs.), the motion characteristics of sitting, transfer, lying and push-up position could be also evaluated in rehabilitation.

In recent years, because of the increase of elder population and the advancement of the woman in society, countermeasures are becoming essential for burdensome works. The tendency is similar in the medical and welfare fields, and there is a high interest in the development of support tools. Work that is physically burdensome has work that keeps the same posture as work using power. For the former, power assist equipments and the like are being developed. But for the latter the development of the corresponding equipment is late compared with the former. Therefore, we decided to develop an EAM brake which the braking force can be controlled by the applied voltage, and to develop an upper limbs support orthosis using it.

Blood sampling is one of the most difficult skills for beginners because it requires hand dexterity. The problem is that many students cannot acquire the skill sufficiently before graduating nursing course. In our previous study, we developed a blood sampling technique measuring system with high spatial resolution. Then, we developed a system that allows students to practice by imitating the movement of teacher's syringe that was reproduced by CG from the measured nursing teacher's motion data and superimposed on the real space via the optical see-through HMD. We believe that by adding a quantitative evaluation function of blood sampling technique to this system, the learning efficiency of blood sampling can be improved. As a basic study for this purpose, we propose an index needed for the development of the aforementioned quantitative evaluation function.

In this study, we develop an automatic evaluation system of motor function of fingers for a home rehabilitation device for fingers. A patient developing hemiparesis cannot perform dexterous finger movements by paralysis at a hand. Therefore, hand recovery is an important target for rehabilitation. However, self-rehabilitation of paralysis by a patient is extremely difficult especially for fingers. From this, a device that automates finger rehabilitation is required. To this, it is necessary to show the change of motor function for finger movement by using the device. In our previous study, we proposed a quantification by a simple integration of finger pressure time series comparison. In this paper, we describe a better integration method which is relevant in terms of physiological validity.

In rehabilitation of hemiplegic patients, it is important to get an ability of loading weight on the paretic lower limb. Generally, physical therapists evaluate the ability by using weight scales in the standing position or observation of movements during walking. Therefore, a method of objective and quantitative evaluation of the limb-loading ability in a series of movements including both standing and walking has not been realized. The purpose of this study was to develop an estimation method of center of mass (CoM) position using inertial sensors. It is expected that estimated relationship between CoM position and foot position makes possible to evaluate movement of the weight. In this report, at first, the attachment positions of inertial sensors were examined. Then, movements were measured during moving the weight in a standing position and during gait. From the results, problems to be solved for improving the accuracy for practical application were discussed.

To find out more effective parameter of electrical stimulation, we examined the effect of an asymmetric pulse in retinal prostheses electrophysiologically. An anodic-first biphasic current pulse, with both symmetric and asymmetric pulses, was applied to the rat's eye using a stimulating electrode. The symmetric pulse had a duration of 0.5 ms and a stimulation current amplitude of 0.1 mA. The pulse duration of the anodal phase was modified to create an asymmetric pulse, and the charge balance was adjusted by changing the amplitude of the electrical current. Evoked potentials were recorded from the superior colliculus. All rats, the pulse duration of the anodal phase was shorter, and as the current amplitude increased, the amplitude of evoked potential increased significantly. The results suggest that the asymmetric pulse with a shorter pulse duration of the anodal phase and a higher current amplitude induces a greater response than the symmetric pulse.

It is well known that microsaccades (MSs) reflect cognitive function such as concentration of attention. To detect MSs, a threshold processing to the velocity component of fixation eye movements is generally used. The detection precision of MSs, however, is influenced by the setting of the threshold value. Recent studies have shown that reccurrent neural networks (RNNs) with deep learnig are effective for pattern recognition in time series data. In this study, we applied an RNN to detect MSs. We used a gated reccurrent unit (GRU) with bidirectional RNN structure to avoid long term dependency problem. After identifying MSs from the differentiation signals of the eye movement data and labeling them, we trained the RNN with these time series data. As a result, time periods which include a MS was able to be extracted, and the MSs were classified with an accuracy of 99.34% for 97920 points eye movement data.

Later

A cochlea is an organ that transduces mechanical energy of sound to electrical signal. Acoustic stimuli vibrate a cochlear sensory epithelium by 0.3-10 nm in amplitude. The vibration is more effectively enhanced as sound intensity is smaller. This called "nonlinear amplification" characterizes hearing. How the motion in the sensory epithelium is controlled in vivo remains uncertain. Here we describe an advanced laser interferometry that monitors both the amplitude and baseline shift of the vibration in real time. With this method, the nonlinear response of the amplitude was observed in the epithelia of live guinea pigs. Moreover, a continuous lift of the baseline by several nanometers was measured only with stimulus sounds at > 70 dB. This motion was absent in postmortem animals. The baseline shift may be involved in modulation of the amplitude of the epithelial vibration at loud acoustic inputs and contribute to nonlinear response essential for audition.

Human audition can distinguish frequencies that are only 0.2% apart as well as perceive trillionfold differences in sound pressure level. These properties stem from sound-induced nanoscale vibrations in sensory epithelium of the inner ear. The epithelium is composed of three layers. Although spatially different distribution of the vibrations in each layer seems to be involved in the extraordinary hearing properties, this profile has not yet been measured by conventional technologies. Therefore, we have developed an imaging system that can 3-dimensionally detect the object's vibrations. The underlying technique is based on the optical coherence tomography. First, a supercontinuum broadband light source is used to achieve high imaging performance. Second, a vibrometry technique and a high-speed camera are incorporated into the system. Through an equipped microscope we scanned the vibrations within the sensory epithelium in a live guinea pig.

To maintain good interpersonal relationships, cooperativeness to think about others from the viewpoint of counterparties is important. In this research, brain functions were measured and analyzed by hyperscanning using fNIRS for multiple people during cooperative work. The performed task is synchronous tapping task. The correlation coefficient of cerebral blood flow change between the pairs was used as an index to evaluate temporal synchronization of brain activity. As a result, the correlation during the synchronization task of the region containing dlPFC and mPFC was higher than at rest. These areas are reported to be related to the assumption and understanding of the state of the partner's mind. Thus, these results suggest that the brain activities of these areas are synchronized during cooperation.

Functional near-infrared spectroscopy (fNIRS) is a non-invasive optical imaging technique for assessing functional activity of the cerebral cortex by measuring hemodynamic responses. fNIRS has been used for the assessment of the cerebral activities of patients with stroke and mental diseases because it allows measuring the hemodynamics in safety, quiet, and unconstrained environments. On the other hand, however, it has been reported that the hemodynamic responses to neuronal activities are contaminated by physiological and instrument noise which is non-negligible. In the present study, causes of the noise in the rehabilitation and those removal methods will be discussed.

Purpose: Direct mail (DM) is one means of informing potential clients of products or services, and to build a client base. We compared differences in the readability of printed material(PM) versus a slate display(SD), and compared with/without address or graphic lay-out differences.Subjects: 32 adult volunteers were employed (males: 14, females: 18).Methods: Subjects were recorded using an eye-tracker and NIRS to evaluate the following conditions:1) PM VS SD, 2) Readability of text, effect of illustration layout, impact of ones own address. Results 1) Paper media induced more activity in PFC regardless of image data or written content.Results 2) Printed material with the subjects own names printed triggered more attention. Readability was varied by Image or character layout.Conclusions: Using NIRS made it possible to perform objective evaluation without relying on experience or intuition.

Purpose: Call center operators struggle everyday to obtain good dialogue with customers. Corporate brands are cultivated through the repeated provision of reproducible good dialogue. We attempted to evaluate operator condition and call quality based on biomarkers.Subjects: 5 operators were employed (male: 1, female: 4).Methods: The performance of subjects were evaluated under 3 conditions; standby, during the call and after the end. These were estimated from the values for PFC brain blood flow, stress value of pNN50 and voice rhythm analysis. Results: There were significant difference among a self-reported condition and PFC activity, pNN50 value and voice-fluctuation. Audio-visual stimulus was able to change bad condition and the condition of low stress and well voice-fluctuation could be made by themself who were well-experienced operators.Conclusions: Objective quantitative evaluation using NIRS etc. appears to be effective for improving call center work quality.

Physical activities such as sports and exercises have been proved to have positive effects on our cognitive functions.In this paper, we investigated how it affects the brain activity to investigate the possibility of driving the remote control (RC) car as cognitive training.The subjects were asked to drive the RC car in two race tracks: oval and 8-shaped tracks, and their brain activity were measured by functional near infrared spectroscopy (fNIRS).Fractional amplitude of low frequency fluctuation (fALFF) was calculated for the fNIRS data of each race track condition and baseline condition.As a result, the difference in the fALFF metric between 8-shaped track and baseline conditions was found in several regions in frontal brain regions.

The aim of this study is to evaluate impression of the robot by brain activity measurement using near infrared spectroscopy (NIRS). Uncanny Valley Hypothesis (UVH) is well known in robotics. The UVH describes the change in a person's affect from affinity to aversion that is evoked by robots that border a human-like appearance. The portion of the human-likeness spectrum in which such aversion is posited to occur is referred to as the uncanny valley.In this study, based on the prefrontal cortex(PFC) activity for 10 participants, the influence of UVH on the brain activity was measured. The result is confirmed that the oxy-Hb level increased, the central portion and right side of PFC, when averse robots were presented. It suggests the possibility of evaluating impression of robots from measurement of brain activity using NIRS.

This study examines effects of full-automatic driving system on drivers brain function. Drivers brain activity is measured using wearable near-infrared spectroscopy (NIRS). Whether the reduction of driving workload by automatic driving system can be evaluated from brain activity was discussed through experiments using a real autonomous car and a driving simulator. Experimental results suggest that the reduction of brain activity can be observed in autonomous car driving.

Toilets are used every day in human life, sanitation management in developing countries, and economic and comfortable in developed countries, and various improvements have been made so far. Recently, attention has been paid to monitor the state of human excretion at the medical site due to miniaturization and high performance of sensors. This is an important issue in terms of health management.Due to sophistication of recent sensing technology and information processing technology, research on state excretion monitoring at the toilet site is being promoted by various groups not only in university research institutes, but also in the real industry, some of which are commercialized It has been marketed in.In this report, we will introduce the present state of the latest excretion measurement technology in these various research and corporate groups, and consider current condition and future prospect of toilet medical engineering.

The authors are developing a home excretion monitoring system - Kawaya Diary - that monitors excretion activity of the elderly by the user's visual inspection of the Bristol scale. This system consists of user terminal at patient home, management terminal at medical institution and server system that runs on the cloud. Important feature of user side terminal is that the user need not set up the environment at the installation site so that even users who do not have knowledge about IT equipment can do installation by themselves. This service handles personal data such as health information of the user and patients medical record stored in electronic carte system. Therefore, we need to clarify the information security standards. In this report, we propose security standards and propose data protection method in Kawaya Diary service.

It is considered to observe fecal characteristics to grasp the state of intestinal environment. Various indicators for evaluating fecal characteristics have been proposed, but it is difficult to evaluate fecal characteristics appropriately due to the influence of the evaluator's subjectivity. Therefore, we propose a method to classify into 6 types objectively using features extracted from line sensor signals while feceses are falling. The captured line sensor signal are combined in the vertical direction in chronological order to generate a spatiotemporal image. From the spatiotemporal image, a total of three features representing the fecal size and ruggedness are extracted to use for the machine learning. As an experimental result for pseudo feceses, the classification accuracy had an average F value = 0.95. As future works, it is necessary to perform evaluation in a real toilet environment and classification for actual feces.

We developed both a non-contact urination measuring system as a toilet sheet shape and a cloud network system to store and to utilize the urination data for diagnosis or nursing care. The measuring system consists of a non-contact thermal sensor, personal identification sensors, and wireless communication. Urination is measured by the non-contact thermal sensor. Personal identification is achieved by walking vibration and ECG. These data are transmitted via Wi-Fi and a mobile virtual network operator, and are stored in one database. The urination data is displayed on a tablet PC. This study was partly supported by SCOPE (162305003), Minister for Internal Affairs and Communications.

In 1998, we have developed a self-management system using video-phone for children with type I diabetes. With video phones installed at home and schools, such an environment was created that made them possible to communicate with a support team. The attempt has shown its supportive effectiveness making to provide for children. For the next ten years, we introduced e-SMBG Blood Glucose Data Transfer System to mobile phones and supported children's self-management. These systems were effective in particular for supporting present condition of growing children who required self-management education at the time of early onset, lived at remote places and were in their adolescence. Currently, it is used as a blood glucose management system using i-Pad, in outpatients and diabetes summer camps.

I hope that our experience will be useful for future improvement and development of remote nursing care.

Introduce the efforts(medical support for the island, regional medical care, system to support home living) of telemedicine in Nagasaki prefecture.

Tokyo University of Information Sciences has been working on building and demonstrating the basic model of the telenursing care system that supports regional inclusive care. In this symposium, I will report about the operation of research center which we have been doing in about one year since the establishment in April 2017. Details will be reported on the day of the symposium.

It seems that we have heard the word Deep Learning (DL) for the past three or four years in Japan. In recent years, it has developed rapidly, and has achieved outstanding results in automatic driving, face recognition, natural language processing field and so on. It is noteworthy that artificial intelligence using DL began to demonstrate the same level as human beings or performance beyond humans in various fields including the above. The application of DL to medical data is also being studied, and how it is applied becomes important. It is expected that results beyond the conventional method can be obtained by applying DL on a large amount of data, such as MRI and CT images, pathological images, DNA sequences, and patient's medical records. In this presentation, I introduce from research which my research group has done recently such as medical/bio-image analysis and classification of DNA using DL.

Deep learning is playing more and more important roles in various fields of medical research. The first part of the talk aims to present the foundations of deep learning and to review the recent progress of deep learning. In the second part of the talk, the applications of deep learning to medical research will be described. The focus will be on the applications of deep learning to medical image analysis and computer aided diagnosis tasks.

This presentation introduces a data-driven strategy to mathematically extract dominant features of surgical procedures and synthesize patient-specific surgical plans from past planning data. The overall framework is developed based on the sparse coding theory. The preliminary experiments demonstrate the efficacy of sparse coding from virtual planning database and show the findings on data-driven formulation of mandibular reconstruction procedures.

Research on Computer-aided diagnosis of medical images using machine learning, especially deep learning, has been actively conducted. However, many machine learning techniques are based on supervised learning; thus, they need a large number of training data with correct annotations. Especially, deep learning sometimes requires tens of thousands of annotated data, and it is quite tough work for radiologists to give annotations to the images. This research aims to classify opacities of diffuse lung diseases in lung CT images, and we introduce an unsupervised classification algorithm without using annotated data, and semi-supervised classification algorithm using only a small number of annotated data. The unsupervised algorithm combines feature extraction using deep autoencoder and Bag-of-features and K-means clustering. Semi-supervised algorithm uses the same feature extraction as the above, but self-training and active learning are applied to Support Vector Machine. In the experiments, six kinds of opacities are classified and the results are analyzed.

Heart Rate Variability has various indices and many of them correlate with each other, suggesting the presence of common underlying factors determining them. We therefore extracted common factors among indices using big data and analyzed the characteristics of each factor. We performed factor analysis on mean NN, SDNN, rMSSD, VLF, LF, HF, LF / HF, scaling exponent alpha 1, DC, the amplitude of cyclic variation of heart rate, and non-Gaussianity index of lambda 25s obtained from 254,357 subjects with sinus rhythm in ALLSTAR 24-h electrocardiogram big data. Two factors that reflect cardiac vagal function and complexity/irregularity of fluctuation, respectively, were extracted. They showed different relationships from each other with aging and daily physical activities.

There is a criticism against big data analyses that they would seldom provide new findings that are not recognized by statistical analysis of small samples because statistical analysis has been developed to estimate population features from a small number of samples. We examined this by analyzing randomly sampled 100 to 100,000 data from 24-h ECG big data. The results indicated that statistical significance provides an indicator of the representability of the samples for population features and that once statistical significance is reached, further increases in sample size do not result in substantial differences in outcome.

In order to obtain basic information on the weekly fluctuation of Japanese physical activity(PA) for gender and age, 24-hour activities of 33,840 males and 42,108 females recorded by Holter ECG recorders with built-in 3-dimensional acceleration sensor were statistically determined by sex, age, and day of the week in Allostatic State Mapping by Ambulatory ECG Repository (ALLSTAR).A significant weekly fluctuation was observed in both gender, the average of PA in 24 hours was the highest on Monday and then Thursday (men 99.0% ,females 99.4% for Monday 100%) was high. Tuesday(97.8%,98.9%), Wednesday(95.7%,98.2%) and Friday(96.1%,97.7%)were intermediate, Sunday (93.5%, 97%) was low, Saturday(90.4%, 91.2%) was the lowest .It is suggested that there is a weekly fluctuation of about 10% in Japanese PA.In addition, we will also consider weekly fluctuation of PA for regional differences in healthy life expectancy.

The recent development and prevalence of intelligent wearable devices enable us access large-scale databases of human biological signals measured in daily life. The utility of such big-data is thought to have a great potential for the early risk detection or diagnosis of diseases. Therefore, the establishment of a new framework to utilize those data, especially a method to extract useful information associated with diseases, has been much focused. In this presentation, we will report our findings on the analysis of the large-scale dataset (about 80,000 people) consisting of 24 h physical acceleration signals measured in daily life; the evaluation of sleep-wake statistics and classification of daily living behaviors based on machine learning approaches. We further discuss their relations with subjects' demographics and regionality, and biometeorological effects.

In order to prevent mental illness in working people, it is important to continuously monitor and evaluate various correlates such as psychophysiological states or habitual behaviors in their daily lives. Critical changes in mental conditions leading to the development of diseases, such as psychiatric disorders, are thought to be signaled by alterations in psychophysiological states or habitual behaviors. In this study, we used smartphone-based ecological momentary assessment (EMA) to record subjective symptoms in approximately 300 working people for one week. EMA involves repeated sampling of subjects' current experiences in real time, offers ways to reduce the effect of recall bias and affords a better picture of subjects' moment-to-moment emotional and behavioral experiences in their natural environments. In addition, locomotor activity and sleep were simultaneously measured. We show the utility of multilevel modeling approaches using specific examples analyzed using the large-scale hierarchical data set collected in this study.

[Objective] Fabry disease is a genetic disorder causing systemic accumulation of globotriaosylceramide (Gb3). In this study, we measured molecular vibrational spectra of Gb3. Vibrational spectra obtained using Raman spectroscopy provides molecular vibrational information without any labeling procedures. Measuring vibrational spectra enables us to visualize Gb3 within tissues in a molecular specific manner.

[Methods] Vibrational spectra of peripheral nerve sections from the patients were obtained using spontaneous Raman microspectroscopy. The data were linearly decomposed into 16 independent bases using sparse coding. The most highly correlated basis with Gb3 spectra in vitro were selected and spatially mapped.

[Results] The spatial distribution of the selected basis generated an image with Gb3 streak in the perineurium, which is consistent with conventional lipid staining using toluidine blue.

[Conclusions] The Raman shifts worked as a molecular specific marker of Gb3 within the tissues from the patients. Sparse coding enabled us to extract quantitative localization of Gb3.

In this study, periodic noble metal nanostructure-based plasmonic optical device was fabricated for surface enhanced Raman scattering (SERS) application. Noble metal nanostructure can exhibit the specific optical characteristics based on localized surface plasmon resonance (LSPR). In addition, using LSPR, SERS can be accomplished high sensitively.Until now, we have been succeeded to develop the novel optical sensor for monitoring of biofunctions using LSPR. Hence, in this study, LSPR-based optical device was applied for SERS.

Brain is an organ that is the most vulnerable to a lack of oxygen in the body. In response to hypoxia, brain microvessels enlarge a diameter of capillaries in vicinity of neurons and partially develops new connections. Previous studies focused on plastic changes in brain microvasculature with cellular interactions of glia and vascular endothelium during each process of cerebral angiogenesis, such as sprout, enlargement, and connection to pre-existing capillaries. It was found that astrocytic processes cover newly-formed vessels with a preserved integrity of blood-brain barrier. Other types of the glial cells play a role in connecting a newly-formed vessel to pre-existing capillaries. However, it remains unclear whether glial cells are functionally involved in a process of capillary sprouts under hypoxia. Future works need to clarify a sensing mechanism of oxygen environment with cellular interactions of glia and microvessels, and impact of microvascular restructuring on neural functions and energy metabolism.

Near-infrared spectroscopy (NIRS) is one of the methods that can measure exercising muscle oxygenation in humans. This review deals with a recent progress in monitoring skeletal muscle oxygenation in response to hypoxic conditions during strength exercise. The presentation comprises the application of muscle NIRS to exercise physiology with special reference to the relationship between intramuscular hypoxic conditions and muscle hypertrophy during strength training. In a recent study, we found during a certain strength-training program, a positive correlation was detected between the increase in muscle cross-sectional area and the intramuscular hypoxic indicator before and after 8-week strength exercise training. NIRS provides not only the information on changes in muscle metabolism during aerobic exercise, but also instructions to training regimens in strength-trained athletes.

Hypoxia potentially impairs cognitive function, while acute aerobic exercise improves cognitive function. It is unclear how cognitive function is altered when acute exercise is combined with exposure to hypoxia. Until now, findings concerning the combined effects of hypoxia and acute exercise on cognitive function are inconsistent, probably due to differences in experimental conditions. The detrimental effects of hypoxia on cognitive function are mainly dependent on the severity and duration of exposure to hypoxia. Some cognitive domains (e.g., memory function) are more vulnerable to hypoxia than others (e.g., executive function). The beneficial effects of exercise are dependent on exercise intensity and/or duration. The interaction between the detrimental effects of hypoxia and the beneficial effects of exercise determines the effects of exercise under hypoxia on cognitive function. We propose that alterations in neurotransmitter function, cerebral blood flow, and possibly cerebral metabolism are the primary candidates that determine the physiological interaction.

Present brain functional localization was done by electrical cortical stimulation (ECS). Although ECS is simple and easy to use clinical routine, we never know the functional dynamics, network and the gradual modulation by training or rehabilitation. Recent functional MRI provide sequential changes of brain functions and indicate multiple-foci related to higher brain functions. Furthermore, electrocorticogram (ECoG) with implanted subdural grids provides us high oscillation signals with good time-resolution directly from the brain surface. ECoG has great potentials to identify the eloquent cortices in real-time and decode brain signals related to modulated functions by any stimuli by combining support vector machine, sparse logistic regression and others. ECoG it the best tool to reveal brain functional dynamics and connectivity, Decoding brain signals would contribute to developing brain computer interface for controlling robots and communication devices. It is indispensable to make tight and continuous collaboration with researchers with different backgrounds.