Japanese journal of medical electronics and biological engineering
Online ISSN : 2185-5498
Print ISSN : 0021-3292
ISSN-L : 0021-3292
Volume 34, Issue 3
Displaying 1-9 of 9 articles from this issue
  • Shigeru MATSUSHIMA, Keiichi HASEGAWA, Yasutomi KINOSADA
    1996 Volume 34 Issue 3 Pages 209-214
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    The longitudinal (T1) relaxation mechanisms involving magnetization transfer (MT) were analytically studied using spoiled gradient recalled acquisition in steady state with MT pulse (MTSPGR) sequences on gelatin and MnCl2 phantoms. The (Ge-P) phantoms contained gelatin of various concentrations. The latter (Mn-P) phantoms had various MnCl2 concentrations in 20% gelatin gel. Magnetization transfer ratios (MTRs) were calculated with and without the MT pulse applied. The change of T1 value was analysed on the assumption that it was related to the MT. An apparent saturation transfer ratio (1/T1*) was separated into the conventional longitudinal ratio (1/T1) and the associated cross relaxation rate (1/TIS). The values of 1/T1 and 1/T1* decreased with in crasing gelatin and MnCl2 concentrations of the phantoms. The MTRs of the Ge-P and Mn-P showed different characteristics, which were presumably attributed to their dominant relaxation mechanisms. In the Ge-P, the relaxation mechanism was possibly dominanted by cross relaxation. On the other hand, in the Mn-P, the relaxation mechanism was possibly dominanted by apparent saturation transfer. Still more, it was accounted that the change of the MTR reflected the tissue malignancy. Results of this study were explicative to the MTR changes in clinical cases, and the analysis of T1 mechanism in this study should prove to be suggestive of a new approach to tissue characterization of human subjects.
    Download PDF (772K)
  • Hirokazu YOKOI, Hideki KUMA
    1996 Volume 34 Issue 3 Pages 215-223
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    In this paper, using the time-continuous model of short-term memory, we have studied the optimum-apportionment mode of the verbal presentation time of seven randomly-selected alphabet letters that were sequentially presented through an auditory display to four young males. This was done to determine the best mode for optimum retention in the short-term memory. An average optimum-apportionment mode was also obtained by averaging the individual scores of the optimum-apportionment mode. This mode has shown the same tendency as that in a visual display. That is, the verbal presentation time of the alphabet letters presented in the first part is long while the presentation time in the latter part is short. Individual differences can be seen in the latter part and can be separated into two groups. In the first group, the verbal presentation time of one alphabet letter is longer than that of the other letters. In the other group, the verbal presentation time of the letters decreases monotonically in proportion to the serial position. However, the minimum verbal presentation time is 100ms longer than that in the case of the visual display. For three of the subjects, the number of recalled verbal alphabet letters in the optimum-apportionment mode is larger than that in the case of the visual display.
    Download PDF (1352K)
  • Seigo KOREMATSU, Teruyuki OGAWA, Keisuke SATO, Kanoko SAWAGUCHI
    1996 Volume 34 Issue 3 Pages 224-229
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    It is well known that development of sleep state in neonates is closely related to maturation of central nervous system. In this study, we investigated the characteristic changes of blood pressure during sleep in a neonate by a new method. A series of 48 hour blood pressure records from a healthy preterm neonate (gestational age: 36 weeks) was stored in a computer system (A-50, Fujitsu). Subsequently autoregressive analysis was carried out on the blood pressure. The resultant characteristic values, such as power and bio-informing activity were compared between quiet sleep and active sleep. The result showed that the powers of both systolic and diastolic blood pressure increased in active sleep.
    Download PDF (669K)
  • Norihiko HAREYAMA, Mamiko FUJII, Katsuyuki SAKAMOTO, Hiroshi KANAI
    1996 Volume 34 Issue 3 Pages 230-237
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    It is very difficult to heat deep seated tumor surrounded by low conductivity tissues, such as bone and fat. We proposed a new RF magnetically induced method for this deep regional hyperthermia. To get high density of induced current in deep tissues, we arranged 4 coils around the sphere agar phantom. The frequency of exciting current is 13.56MHz. In this paper, we will show the calculation method of SAR distribution and temperature distribution in 3D inhomogeneous model. We also discuss the possibility of this heating method from both experiment and calculated results. In the phantom experiment we examined two types of the coils. The calculated results and experimental results show a good agreement, these therefore, the reliability of this calculation method was confirmed. The heating pattern of these coils were not very effective for deep regional hyperthermia. We also calculated temperature distribution in the brain model which was constructed from CT images. We assumed tumor in the brain with half of the blood flow compare to normal tissue. Deep seated tumor was heated by the effect of inhomogeneity and blood flow, except there were hot spots near the coil. In the case of shallow region heating, we can get effective localized heating pattern. From these results, we cannot surely say that this heating method is quite sufficient for deep regional hyperthermia. But since there is no other method to successfully heat the brain non-invasively, this kind of examination with other coil patterns must be continued.
    Download PDF (1211K)
  • Satoshi ITO, Osamu SATO, Yoshitsugu KAMIMURA, Yoshifumi YAMADA, Kunio ...
    1996 Volume 34 Issue 3 Pages 238-248
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    MRI fast-imaging technique such an echo-planar imaging (EPI) has begun to be employed in commercial MRI system and data acquisition time may be shorten to less than 100ms. With this imaging system, if MRI images were reconstructed in a moment, we would be allowed to get a real-time moving image. Reconstructing images in synchronization with data acquisition may be executed numerically by using Digital Signal Processing board or parallel combined CPU in the case of small data-size. However, these methods even take a few time to reconstruct images and it may be the limiting factor for the real-time imaging. Optical information processing has the potential to overcome the time limitation related to image reconstruction, because optical computation has the advantage of very fast parallel processing. Holography is the most popular technique for optically reconstructing images and has the potential of producing a three-dimensional image. The expression of NMR signal in Fresnel transform technique is similar equation to that of the Fresnel diffraction equation in light. Therefore, holographic reconstruction of NMR images is feasible, making a hologram from NMR spin-echo signal and reconstructing images using a coherent optical system. In this paper, we describe a new method of opto-electronic reconstruction of NMR images using a liquid crystal-spatial light modulator (LC-SLM) as a hologram display, addressing a hologram electrically and using a coherent optical system. Experimental results show considerably good images are obtained even from the commercially available LC-SLM. The results also indicate a real-time reconstruction of NMR images by combining the NMR fast-speed imaging technique, because the hologram pattern on the LC-SLM is refreshed at a video-rate that is fully catch up with the data-acquisition time of high-speed imaging.
    Download PDF (8818K)
  • Hiroyuki ITAGAKI, Yukari ONODERA, Etsuji YAMAMOTO, Hideaki KOIZUMI
    1996 Volume 34 Issue 3 Pages 249-254
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    We have been investigating functional MRI (fMRI) using Student t-test analysis for selecting activated regions. This analysis is sensitive to the signal to noise ratio (SNR) level, so that a low SNR reduce the activated size. In this report we describe new analysis, a combination of Student t-test analysis and signal change analysis. This combination of two analyses is less sensitive to the SNR level. Experiments of fMRI were performed using a 1.5-T prototype MRI system with local gradient coils. Healthy volunteers carried out a left-finger tapping task in these experiments. Clinical images were acquired using spin-echo sequence to identify the structure of the brains. Functional images were acquired using two sequences, echo-planar imaging (EPI) sequence and interleaved EPI (IEPI) sequence. Each imaging sequence had the following parameters: 1) EPI: TE=15ms, data acquisition time=75ms, spatial resolution=4×2mm and temporal resolution=2s; and 2) IEPI: TE=15ms, data acquisition time=75ms, spatial resolution=2×1mm, and temporal resolution=8s. The SNR of IEPI is about a quarter that of EPI. The activated regions were selected by using Student t-test analysis and the combination of two analyses respectively. The activated size was calculated based on the number of pixels in the activated region and the pixel size. As a result of Student t-test analysis the activated sizes of IEPI reduced to less than half those of EPI. But there were not a significant differences in the activated sizes estimated by the combination of two analyses between EPI and IEPI. We conclude that our proposed combination analysis method can estimate the size of the activated areas more precisely.
    Download PDF (2583K)
  • Shigeto NISHIDA, Masatoshi NAKAMURA, Shugo SUWAZONO, Manabu HONDA, Hir ...
    1996 Volume 34 Issue 3 Pages 255-263
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    In order to extract the waveform characteristics from the single sweep evoked potentials (EPs), a morphological filter with functional structuring element was designed. The properties of the basic operations for the morphological filter; erosion, dilation, opening and closing were first clarified. Based on these properties, the morphological filter for the single sweep records was designed by taking into account the method employed for the visual inspection. In order to accurately extract the single sweep waveform, a method for determining the structuring element, which appropriately represents the features of the EP waveform, was presented. The designed morphological filter was evaluated using the simulation data which appropriately represent the ERP (event-related potential) and the EEG (electroencephalogram). Furthermore, the designed morphological filter was applied to the ERP data obtained from 16 normal adults, and variability of the peak latency and amplitude of P300 in single sweep ERP were demonstrated.
    Download PDF (1280K)
  • Yuko TANAKA, Tomohiko KIHARA, Yoshitsugu KAMIMURA, Yoshihumi YAMADA
    1996 Volume 34 Issue 3 Pages 264-273
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    We developed a new registration method based on surface matching of multi-modality medical images. This correlation procedure does not need any artificial markers but uses partial body surface information. This allows our registration method to be applied to patients whose body structure has been changed partially (e. g. by surgery or other medical treatment) in the interval between the acquisition of the two data to be correlated. This method consists of three steps. First, body surfaces of the two data are segmented by thresholding and represented in the form of z-buffers. Next, the operator selects ROIs for the matching on the surface of one of the data. Finally, the registration matrix relating the two data is calculated using an algorithm based on the downhill simplex optimization. The cost function of the optimization procedure is the root mean square of the distance the ROIs and the surface of the second data. We present a validation of the accuracy and reliability of this registration method by using X-ray CT and MR phantom data. In the application of surgical planning and diagnosis of cerebral and craniofacial surgery, a registration accuracy better than 1mm is requested. By using more than three ROIs (ROI diameter>40mm), this condition was achieved with a reliability better than 90% in the phantom study. The distortion of the MR images was estimated during the validation. To guarantee our condition of accuracy, the correction parameters of the distortion has to be taken into account during the registration procedure. The accuracy and the effectiveness of the method were confirmed visually on X-ray CT and MR images of the head.
    Download PDF (4468K)
  • Yasuhito YAMAMOTO, Takami YAMAGUCHI
    1996 Volume 34 Issue 3 Pages 274-278
    Published: 1996
    Released on J-STAGE: October 14, 2011
    JOURNAL FREE ACCESS
    The spontaneous alignment of arterial endothelial cells under non-uniform and unsteady flow conditions was simulated using a three-dimensional computational fluid mechanical model of cultured endothelial cells. Endothelial cells were simulated using a 2D Gaussian distribution function, and placed on a flat surface with randomly assigned rotational movement. The Navier-Stokes equations of a Newtonian fluid were solved using a finite volume method, and the absolute wall shear stress (WSS) in the case of non-uniform flow and the root mean square of WSS in the case of unsteady flow at the summit of cells was calculated. Only the movement of the cells which reduces the WSS was saved. The cell model eventually showed an alignment after long time simulation in the unsteady flow case. In the nonuniform flow case, cells were aligned along the local flow.
    Download PDF (821K)
feedback
Top