An expert knowledge-based method integrated with amendment function is presented in order to develop an effective automatic sleep stage determination technique for clinical practice. The main methodology is based on the expert knowledge of visual inspection by qualified clinician. A knowledge database is consisting of probability density functions of parameters for various sleep stages. The sleep stage is decided automatically based on conditional probability. An amendment process is developed to modify the decision on sleep stages. We applied the methodology to the sleep data of patients with sleep disorder under usual recording condition in hospital. The determination result was satisfactory. With the amendment function, the determination system showed better performance on stage change and continuity detection. The developed expert knowledge-based method can be an assistant tool for clinics.

View full abstract

In this paper, the displacement and the strain behaviors under a forward bending moment at a spine model that consists of four vertebras and three intervertebral discs were investigated. The spine model used for the measurement test is calf's lumbar vertebra. Intervertebral discs that compose a spine are main elements that enable free movement of a human body, and they also play roles of a buffer corresponding to the load applied to the spine. In consideration of such a fact, we tried to measure an overall deformation behavior of the spine and a detailed strain distribution of an intervertebral disc simultaneously, and the relation of the bending deformation of the spine and the strain behavior of the intervertebral disc was examined. The spine is considered as if a structure consists of solid vertebras linked by flexible intervertebral discs, the bending angle at the intervertebral disc was obtained from the overall deformation of a test spine that has been measured with laser displacement sensor at typical points. On the other hand, detailed displacements distribution on a surface of the intervertebral disc was measured by Electronic Speckle Pattern Interferometry (ESPI) method, and strain distribution were calculated based on the displacements distribution. The results showed that although changes of the strain in the vertebral body area were little, the strain concentration at the intervertebral disc became larger according to an increase of forward bending moment. Moreover, compressive strain on the side of the belly of the intervertebral disc was measured, and tensile strain on the backed side was observed. In addition, the relation between the bending angle at the intervertebral disc that dominated overall deformation of a spine and maximum strain value obtained at the surface of the intervertebral disc was shown.

View full abstract

In the field of rehabilitation or care, it is extremely important that medical staff clearly understand the appearance of daily activities of the subjects to make and improve rehabilitation programs or care plans. Information on daily activities of the subjects is usually obtained in each facility by questionnaires answered by subjects or their families, and through observation by medical staff. However these methods have some difficulties; questionnaires might include unreliable information; observations differ due to each observers subjectivity; observing is stressful for medical staff and observations invade subject's privacy. With these points in mind, several works on some sensors attached to the subjects and the sensing data analyzed in order to understand daily activities of the subjects have been reported. In this paper, a novel method is proposed that can judge daily activities, such as lying, sitting, standing, walking and wheelchair-driving, by using neural networks from the sensing data of the accelerometer attached to the subject. The network concerning posture of the subject when lying, sitting and standing and the network also concerning actions when walking and wheelchair-driving are constructed in a cascade. From the experimental results for 7 subjects in a facility, their matching ratios of evaluated the Level of Daily Activities between a use of video and the proposed method are 86.0% to 95.7%, which shows the effectiveness of the proposed method.

View full abstract

Holter electrocardiogram has spread with medical institutions widely to be able to detect transient arrhythmia and ischemia. It is necessary to be analyzed automatically, because of a large quantity of data that is recorded for hours. However, the accuracy of automated classification for QRS complexes is not sufficient. This paper presents the application of a hierarchical self -organizing map (SOM) to classify QRS complexes. Each beat is divided into some sections, and the characteristics are learned by the first SOM. The qualitative attributes are classified for characteristics of each section. The qualitative attributes are arranged in time series for each beat, and they are learned in the second SOM. As a result, each beat is classified by their characteristics. We evaluated this method using MIT-BIH Arrhythmia Database of 17 cases 33,362 beats and compared to a correlation coefficient method. The classification error rate was 0.41% and the number of the classifications was 28.7. Using this method the classification of QRS complexes is significantly improved.

View full abstract

Electrical impedance (or admittance) cardiography is a simple and still potential method for non-invasive and continuous monitoring of stroke volume and cardiac output (CO) by detecting the electrical impedance of a thorax which is roughly assumed to be a two-compartment coaxial cylindrical model composed of the aorta and its surrounding thoracic tissues. Although the band-electrode method proposed by Kubicek et al. has been widely used for the detection of the electrical impedance, long-term use of the band-electrodes is often uncomfortable and causes inflammation of the skin. Replacement of band-electrodes with spot-electrodes is therefore highly required for clinical practice. We have previously investigated an optimal spot-electrodes array from such a view point, and reported these results in this journal. However, the current distribution during cardiac blood ejection period showed little uniformity, and detailed investigation of this phenomenon was remained as a future problem. The present study concerns with this problem through the analysis of pulsatile components of the thoracic impedance along the medial line of the thorax by a finite element method and those by actual measurement. From these investigations, an optimal electrodes-location for voltage pick-up was to be determined as the medial portion at the level of the clavicle and the portion above the xiphisternum. Furthermore, we have investigated the accuracy of CO measurement using impedance cardiography by the new spot-electrode array to that using a pulse dye-densitometry method. From the CO comparison experiments, bias and precision calculated for the impedance cardiography versus a pulse dye-densitometry method averaged -0.43±1.42 L/min, the validity and reasonable accuracy of the new spot-electrodes array was clearly demonstrated.

View full abstract

To solve endoscopic surgical problems such as lack of depth information and hand-eye coordination, this paper proposes an intuitive human-machine interface for 3-D positioning of surgical instruments that uses the monitor as an input device of the desired point. This interface must achieve 3-D positioning from only 2-D image information. To cope with these problems, we propose a new 3-D positioning method using a laser pointer. This method achieves 3-D positioning of a surgical instrument by determining and controlling the direction of insertion of the instrument based on visual servoing, followed by estimating the distance between the instrument and organs using projective invariants. To evaluate validity of this interface, we conducted an experiment in which the interface system makes trajectory tracking with keeping the instrument-organ distance constant. As a result, it was confirmed that a laser beam was irradiated along the desired trajectory in few gaps while keeping the distance less variant.

View full abstract

This paper describes the development of a novel laparoscope manipulator using the medical hydraulic linear actuators. The manipulator is composed of a Stewart-Gough platform with six degrees of freedom and six hydraulic linear actuators, and can hold a general laparoscope. In the current prototype, the position of a laparoscope can be controlled through a joystick interface. To evaluate the performance of the manipulators, three experiments were conducted in which each operating surgeon used the manipulator to perform an in-vivo laparoscopic cholecystectomy on a pig,respectively. Also, in the experiments, we measured the time to set the manipulator, the time to switch to normal operation from robotic one, and the time to clean lens of endoscope. As a result, the manipulator could replace a human assistant during the operation.

View full abstract

Environmental noise is an inevitable obstacle for any speech-signal proccesing-based systems. Although bone-conducted microphone, which is robust for external noise, seems “a panacea for this syndrome”, quality of bone-conducted speech is deteriorated due to suppression of its high frequency component. Therefore, to comfort bone-conducted speech communication, signal proccesing to ease the deteration is indispensable. This paper proposed a method to improve quality of bone-conducted speech by restoring its high frequency component. The authors compared spectrum envelop of bone-conducted speech with that of air-conducted speech with cepstrum analysis, and revealed that replacement of spectrum envelop of bone-conducted speech with that of air-conducted speech can compensate the high frequency component of bone-conducted speech. Consequently, to apply spectrum envelop replacement for each phonemes appropriately, the authors designed a system using codebook-mapping which corresponds to given bone-conducted speech by each phonemes. Prototype of proposed speech enhancer was evaluated with cepstrum distance and hearing test. The evaluation ensured the prototype is enable to enhance bone-conducted speech especially to voiced speech. The proposed method may contributes to realize applications using speech signal under various environments.

View full abstract

In percutaneous puncture simulation, the finite element method is usually used to compute the needle bend and tissue deformation. This method, however, tends to be not suitable for real-time applications such as training systems due to its high computational cost. In this paper, we propose a novel deformable model with a low computational cost. The proposed model reproduces needle bend and tissue deformation in the direction perpendicular to the long axis of the needle. The proposed model consists of a soft tissue deformation model based on the Long Element Method (LEM) and a needle bending model which is designed to be consistent with the tissue model. The parameters for both models are put in the same simultaneous equation to be solved for the simulation. We verified that the proposed model shows a low computational cost by the computer simulation. The accuracy of the model was also evaluated by the experiment using needles and tissue phantoms. The experimental result using only needles showed that a bend could be estimated by the pre-optimized parameters. The optimized parameters were available even if different displacement was given to the manipulated point of the needle.

View full abstract

Transcatheter arterial embolization (TAE) is a way of occluding blood vessels such as arteriovenous fistula that are physiologically unnecessary. Although this technique becomes more popular in clinical practice for its minimal invasiveness, it is sometimes fraught with difficulties especially when a catheter cannot be placed directly in a target vessel. The risks associated with TAE may be reduced by better understanding of flowing behaviors of embolic agents in blood vessels. In the present study, we established a numerical model to simulate TAE and investigated influences of injection positions and intervals of spherical embolic agents (SEAs) on their flowing behaviors. Flows of blood and SEAs were modeled by the equation of continuity along with the Navier-Stokes equation, re-formulated based on the moving-particle semi-implicit method where continuum is represented by collective behaviors of particles. An SEA was modeled as an aggregate of particles. Particles consisting of the membrane of SEA were linked by springs which resist to stretching and bending. Based on the virtual work principle, a force acting on the SEA membrane was calculated. Using this numerical model we simulated injection of SEAs into a T-junction blood vessel where a target vessel to be occluded with SEAs is connected perpendicularly to a main vessel. The simulation showed that a distribution ratio of SEAs between the main and target vessels did not always coincide with a distribution ratio of blood flow evaluated in the absence of SEAs. Although the proximal variations of the injection positions were less influential, the radial position of injection significantly affected the flowing behavior of SEAs. Frequent injection of SEAs induced jam of SEAs in the target vessel, causing undesired flowing of SEAs into the main vessel. These results indicated the significant influence of injection position and interval on flowing behavior of SEAs.

View full abstract

In recent years, tissue engineering has been progressing, and myocardial sheets that consist of cultured myocytes are designed to regenerative medicine. Here we propose myocardial cell net, which is expected to be more flexible and richer in blood flow than dense sheet. To make myocardial cell net, we cultured neonatal rat cardiac myocytes on micro-lithographically fabricated dishes. Myocardial cells were obtained from neonatal rat hearts and cultured on the dish which have micro-lithographed squared pattern. Cell growth was observed with a time laps system. Some cardiomyocytes migrated between squares and made cell-to-cell connection, others stayed in squares, divided in two and gradurally made connection at the corners. Finally cardiomyocytes formed cell net. Myocardial cells were regularly directed and synchronous contraction was observed. Rhodamine phalloidin staining revealed that actin fibers were oriented. We succeeded in production of myocardial cell net using micropattern-attached culture dish. It may be a prospective candidate for application in regenerative medicine.

View full abstract

Ventricular fibrillation (VF) is the major cause of sudden cardiac death by lethal arrhythmia. Regional abnormalities such as myocardial infarction are regarded as the cause of VF; however, VF is clinically caused without the organic heart disease. Then we hypothesized that a human heart equips a protective mechanism against VF and thus the breakdown of the protective mechanism induces VF. We considered that the essence of the mechanism is the ventricular transmural gradient (electrophysiological heterogeneity) . To confirm this hypothesis, we have developed a 3-D ventricular wall model and analyzed the dynamics of spiral wave reentry and filament (reentrant organizing center) . It was found that the ventricular gradient forces an intramural filament to drift out a boundary and reduces the sustainment of VF. On the other hand, the pharmacological modifications of ventricular gradient to increase transmural dispersion of repolarization (TDR) break the protective mechanism and sustain VF.

View full abstract

In the brain, cooperative activity of many neurons is believed to play essential role in neuronal information processing. To evaluate correlation of neuronal activity, cross-correlation histogram (CCH) and cross-correlation coefficient (CCC) of spike timing of neuron have been widely used. The CCC is available to judge whether or not the two neurons have a synaptic connection. However, CCC cannot be used to evaluate the strength of the connection because the CCC is dependent on the firing rate of postsynaptic neurons ; even though the synaptic efficacy was invariant, increase in the firing rate of the postsynaptic neuron due to the increase in background inputs would result in decrease in CCC. In this paper, we propose a method to evaluate the strength of the neuronal connectivity (synaptic weight) and the excitability of postsynaptic neuron (threshold) . We assume a simple neuron model and stochastic property of background inputs. Both the weight and threshold are calculated based on the improved CCH method but the weight is unaffected by the firing rate of postsynaptic neuron. We applied this method to data recorded from hippocampal CA3 neurons in vitro to estimate the distribution of synaptic weight of CA3 neurons. It was found that the histogram of synaptic weight was well fitted by the unimodal gamma distribution, whereas the CCC was similar to the exponential distribution. According to the intracellular recording data, the histogram of excitatory postsynaptic potential amplitudes, corresponding to the synaptic weights, is a unimodal, gamma-like distribution. Therefore, it is suggested that our method provides a better estimation than the conventional method. Finally we investigated activity-dependent change in synaptic weight. It was found that the histogram unchanged after an application of conditioning stimulation facilitating synaptic modification, even thought individual synaptic weights changed. These results suggest a mechanism for regulating synaptic distribution in the neuron.

View full abstract

In multiunit recording with multisite neural electrodes in the brain, an adjustment of electrode parameters such as the spatial arrangement of the recording sites to the target brain region is essential to obtain good signal-to-noise ratio data. However, in the most case, the parameters have been decided according to the experience of experimenter due to the cost of animal experiments. In this study, we propose a framework to optimize the electrode parameters by the virtual experiment of multiunit recording and data analysis including spike detection and sorting of the simulated multiunit data. For the purpose, we construct the 3-D models of neural tissue and multisite electrodes. The neural tissue model is composed of sphere-shaped model neurons which are randomly arranged to avoid overlap in a cubic region. Two types of model neurons (bursting and non-bursting) are included. Each model neuron independently and randomly emits action potentials which are recorded by the model electrode with the amplitude as a function of distance between the neuron and recording site. Model parameters are determined based on the anatomical and physiological data of rat hippocampus. Five models of multisite electrodes are constructed with identical shape but different spatial arrangements of recording sites to investigate how the arrangement of recording sites, one of the electrode parameter, affects the performance of multiunit recording and spike sorting error. Virtual experiments and the performance assessments of spike sorting are conducted. It will be shown that in the case of silicon tetrode with diamond-shaped arrangement of recording sites, the number of simultaneously recorded neurons slightly increases with increasing the interval of recording sites (IRS) from 15 to 40 µm. The frequency of spike sorting error is found to be minimal for the electrode with IRS=25-35 µm. In conclusion, the proposed framework will be useful for the optimization of parameters of multisite neural electrodes.

View full abstract

Using high-density electroencephalography (EEG) , the brain regions involved in tapping of the index fingers were investigated. Subjects were requested to perform voluntary alternate tapping movements with both index fingers as fast as possible. During the task, the tapping mode in which both index fingers moved simultaneously was interlaced. The alternate tapping (A-mode) and simultaneous tapping (S-mode) groups were extracted using a histogram of the inter-tapping intervals. EEG coherence was used to evaluate functional connectivity between cortical regions. Compared with the S-mode, connectivity increased significantly in the A-mode in the mesial-central circuitry including the supplementary motor area (SMA) and the primary motor area (M1) , in the fronto-centroparietal circuitry including the primary somatosensory area (S1) and the premotor area (PM) , and in the fronto-central circuitry including the PM and M1 in the dominant hemisphere. In addition, interhemispheric connectivity increased in the PMs of both hemispheres. These findings indicate that the connectivity involved in the internal onset of the movement (SMA-M1) , in the sensori-motor integration (S1-PM, PM-M1) and in the phase planning (both PMs) is associated with the involuntary switching between A-mode and S-mode.

View full abstract

A tissue-engineered muscle has a potential to realize a flexible actuator with high efficiency as a bioactuator, whereas mechanical actuators currently used are inflexible and low efficiency. In this study, acellular tissue and collagen gel were used as the scaffold for myoblasts to regenerate the tissue-engineered skeletal muscle and its isometric contractile force was measured. Furthermore, a solid object formed by the micro stereo-lithography system was driven by the tissue-engineered skeletal muscle. A cold suspension of C2C12 cells embedded in Type-I collagen gel was added on and between two collagenous vascular grafts placed 13 mm apart and cultured for 2 days in high-glucose Dulbecco's modified Eagle's medium (HG-DMEM) supplemented with 10% fetal bovine serum and 1% antibiotics. The culture medium was shifted to HG-DMEM supplemented with 7% horse serum and 1% antibiotic to enhance differentiation of the cells to the myotubes. The cells in the collagen gel differentiated to multinuclear myotubes at 12 days after differentiation induction. From the result of force measurement, the isometric contractile force was increased with increase of amplitude, duration, and frequency of the electrical pulse stimulation to the tissue engineered skeletal muscle. Unfused tetanus was generated at 20 Hz when the amplitude and duration were 50 V and 2 ms, respectively. The value of isometric contractile force of twich and unfused tetanus were 145 μN and 180 μN, respectively. The unfused tetanus was not observed at amplitude of 10 V and duration of 2 ms. The solid object having the tissue-engineered skeletal muscle could be driven by electrical pulse and the motion was visible without microscope. These results suggested that it could be possible to realize the development of the bio-actuator with tissue-engineered skeletal muscle.

View full abstract

The shape of focal spot is affected by the distribution of polarization and phase on the cross-section of the incident beam. We demonstrated the spatial resolution enhancement effect using the difference of the focal spots between a linearly polarized beam and an azimuthally polarized beam. The azimuthally polarized beam has the vortex of polarization on the cross-section of the beam. Since the vortex of the azimuthally polarized beam forms a doughnut shaped focal spot, the spatial resolution is expected to increase with the difference image between linearly polarized beam excitation/detection and azimuthally polarized beam excitation/detection. We applied an eight segmented polarization mode converter, which was developed by ourselves and could switch those polarization modes of the excitation beam electrically, to a commercial confocal microscope. We derived the suitable weight of subtraction theoretically, and demonstrated the spatial resolution enhancement effect by observing the stained cell cytoskeleton.

View full abstract