Computed Tomography (CT), a new method of medical X-ray imaging has become the center of wide interest in clinical medicine and it is about to achieve general use in Japan. In this review, the technical background, the principle, present status and future prospect of this technique are described with some explanations of reconstruction algorithms.
The technical background of CT consists of X-ray transverse tomography, RI imaging devices with on-line data acquisition system and algorithms of 3-D reconstruction from projections. A CT-Scanner is a kind of apparatus which make the transverse tomograph, but is different from conventional ones in principle. The features of this apparatus are as follows. 1) It uses a narrow beam of X-ray. 2) The radiation detector is a scintillation counter. 3) It reconstructs distribution of absorption coefficients in the cross section from intensities of transmitted X-ray. The CT-Scanner can show variations of absorption coefficients in the soft tissues. EMI's brain-scanner is described as an example of the first generation scanners. New types of CT-Scanners, which are more rapid in scanning speeds and finer in resolutions, are also described. Among them, Dynamic Scanner produced by JEOL is promising because it scans X-ray not mechanically but electro-optically. In future, one may develop a dynamic CT-Scanner by the use of an electro-optical scan.
Algorithms of 3-D reconstruction from projections are explained in relation to CT. Iterative methods, Fourier transform methods and convolution methods are mentioned. For the parallel beam geometry, convolution methods are best of all algorithms. But, for the fan beam geometry, many methods are being studied in order to obtain the optimum one.

抄録全体を表示

A new type of computer method for classifying aperiodic EEG waveforms is presented. The EEG waveforms are trapped into a time-window with certain duration and threshold levels, and various activities are detected as to whether or not these activities coincide with the logic patterns stored in the memory of the computer.
Continuous EEG records are digitalized at 100 samples per second into each group with a duration of 10. 24 seconds successively, and the zero level is determined by subtracting the average value of the 1 024 sampled data. A time-window of 3 seconds width with threshold levels of ±50 μV and ± 30μV is passed along the time-axis of the EEG time series, and the waveforms arecoded into a series of 1, 0, -1 patterns comparing whether or not amplitudes exceed the threshold levels. To simplify this threshold pattern composed of 300 samples, codes that are consistently the same are condensed into a single code, and an adjunct-window pattern is derived. Zero-crossing patterns are also constructed. Paroxysmal activities are detected by the adjunct-window patterns and information on the previous window. Stationary activities are classified by the sum total of 1 and-1 in the adjunct patterns and the number of zero-crossings.
In order to verify the feasibility of the method, typical EEG patterns during sleep in normal adults are filed and classified into several categories, and experimental tests are performed by an off-line system.

抄録全体を表示

The implantable electrode assembly consisting of collagen and metalic electrodes has been developed to fill the need for the continuous recording of neural signals from the intact nerve and other bioelectrical signals (EKG and @EMG) in both anesthetized and awake animals. The electrical properties of collagen electrodes were compared with that of conventional metalic electrodes in physiological saline (0.9% NaCl) solution. When the electrodes were implanted on the nerves of cats, the collagen electrodes filled with extracellular fluid operated as a recording electrodes of neural signals. The mechanical noises in the collagen electrodes due to movement of the body or displacement of the electrodes were negligibly small as compared to those in the metalic electrodes. Using this implantable collagen electrode the neural signals were recorded in awake cats 0 to 8 days after imnlantation suroerv.

抄録全体を表示

Studies on the development of the artificial arm are gradually increasing recently, but hands closely similar to natural ones that move very skillfully have not been produced. Although artificial hands capable only of simple grip actions have been made, most of these customary artificial hands move by such external powers as electricity, gas or oil pressure. The finger-tip of these artificial limbs are hook-shaped, so they can only grasp objects, but not manipulate them skillfully.
This paper describes the development of the arm equipment that can bend the joints of each finger. The actuator of these finger-tips is made of the vinyl tube moved by gas pressure. The merits of this equipment are as follows :
1. The possibility of grasping object smoothly and naturally.,
2. The ability of changing the grasping power by the change of gas pressure.
3. It is easy to put on and well-fitting.
4. It is not heavy.
The weak points are as follows :
1. It is noticeable because of its shaped finger joints.
2. It is difficult to control accurately without development of a micro magnet valve

抄録全体を表示

The analysis of circulatory system using polynomials has two characteristic features; it can combine various parameters of the circulatory system in a global sense to form a mathematical description, w hich, in turn, suggests a clue to solving the regulatory question. Thus, a research was porformed to investigate the relationship among cardiac output (F), heart rate (H) and blood pressure (P), abbreviated as FHP relationship, using the polynomial models consisting of methematical and biological models. The former is general mathematical expression of FHP relation while the latter implies some biological mechanisms, i. e., in our study a tonus which, corresponding to the tonic activity of an assumed cardiac output center, are introduced here in the form T=aH+bP where a and b are constants. Then cardiac output is given by F=k₀+k₁T+k₂T² where k₀, k₁ and k₂ are constants, which can be determined by the least square method. The equation with highest correlation coefficient of those between cardiac outputs observed and calculated by the equations of various coefficients b (for a=1), was adopted actually as a biological model.
Ten dog experiments resulted in the coefficient of 0.977 on average for the methematical and 0.959 for the biological model. These high correlations and the small difference between them suggest that the relation, expressed by the biological equation, among the circulatory parameters can be well applied to the actual circulatory phenomenon in the aorta. This is veryfied by the figures representing the FHP relationship observed. Furthermore, the biological equation indicates that cardiac output increases with increasing heart rate while it decreases with increasing blood pressure in most cases under the conditions induced by infusion with noradrenaline, isoproterenol and acetylcholine.

抄録全体を表示

The design method of closed-circuit TV pupillometer is discussed, especially from two viewpoints, and a model is made.
From the viewpoint of the safety of infrared illumination against the eye, illumination power 1 mW/cm² was adopted, based on the fact of infrared exposed level of workers at glass or steel manufacturers and calculation of natural light exposed in our daily life.
From another viewpoint of sensitivity and dynamic performance, silicon-vidicon tube was adopted, together with good optical system. Its sensitivity is enough under the level of safe infrared illumination. By processing TV video signal electronically, the area, vertical diameter and horizontal diameter of pupil can be measured. As the after image of this tube is reduced so much, the dynamic performance of this pupillometer is greatly improved to have the band width of about 5 Hz.
A binocular electropupillometer was also made in order to measure both pupil diameter at a time. It will be possible to apply this binocular electropupillometer to clinical use.

抄録全体を表示

The opposition of the arterial vessel to the flow of cardiac output was simulated by the most minimized three component model consisting of two resistors and a capacitor. Thus, using the cardiac flow wave calculator, designed on the theoretical basis of the three component models, a trial was performed to calculate the flow wave of cardiac output from cardiac pulses such as blood pressure. The calculator has two adjusting mechanisms : time constant of a low pass filter and a potentiometer. The former is selected so that the wave matching the actual flow wave may be calculated, while the latter is adjusted so as to put the diastolic flow signal to zero level. Close agreement existed between cardiac flow waves observed and calculated by this calculator from blood pressure under the conditions induced by infusion with noradrenaline, isoproterenol and acetylcholine. Furthermore, each component of the three component models was computed from the calculated flow wave and the pressure wave under the same conditions described above.

抄録全体を表示