A voice-controlled computer system for multichannel functional electrical stimulation (FES) of the paralyzed hand has been developed. The computer system is composed of a personal computer NEC PC-8801 mk II with a voice recognition board NEC PC-8012-05, and a digitizer Logitec K-510.
Stimulation patterns for restoring three kinds of prehension, such as cylindrical grasp, key grip and parallel extension grip, are prepared with the digitizer and stored as three data files in a floppy disc system. Speech inputs from a head-set microphone are processed by the voice recognition system and are used as commands for selection of a prehension pattern and execution of the hand movement. Angles of the neck flexion and extension detected by an angular sensor utilizing mercury are used for proportional control of the hand movement.
As voice commands for selection of a prehension pattern, ‹CUP›, ‹KEY› and ‹CARDS›, which correspond to cylindrical grasp, key grip and parallel extension grip, respectively, are used. These commands select one of the corresponding data files in the floppy disc system and set it to a work area of a RAM. After selecting a pattern, a start command is given by the voice of ‹START›, resulting in an acceptance of proportional control signal and readout of the corresponding stimulating data. Thus, multichannel stimulating outputs are applied to the nerves distributed in the paralyzed hand muscles. A voice command ‹HOLD› causes holding of constant stimulating voltages at the state when the command was given. Restart of the proportional control can also be ordered by the voice command ‹START›.
Using this voice-controlled system, easy volitional control of the paralyzed hand in a C 5 quadriplegic could be obtained successfully.

View full abstract

A microcomputer-based pacemaker system for evaluating the pacemaker treatment of arrhythmia is described. This system can function not only as a demand pacemaker but also as a pacemaker for tachycardia control as well as a defibrillator. All stimulations are delivered through a catheter electrode in the right ventricle and a subcutaneous electrode. For the detection of arrhythmia, ECG and blood pressure signals are used. In particular, the blood pressure signal is useful to detect the actual heart rate during burst pulse pacing which is capable of controlling ventricular tachycardia automatically. Moreover, during burst pulse pacing, special attention is paid to detecting hypotension in order not to cause ventricular fibrillation. Pacing failure is quickly sensed from variations in R-wave pulse amplitude and pulse width.
When, after the heart has been stimulated, the EGG shows no response, and hypotension continues without any rhythmic contraction, a battery-operated defibrillator is triggered and defibrillation is accomplished automatically.
This system has been designed using two 16-bit microcomputers, one of which performs signal processing and the other, output control. In animal experiments, examples of the use of the system for demand pacing, tachycardia termination, and defibrillation are given. This system may be used to study tachycardia termination and defibrillation in clinical laboratories.

View full abstract

In Japan, those who have severe hearing loss are not eligible for a car driver's license. As a technical approach to solve this problem, an electronic device, which detects traffic-alarm-sounds, i. e., horns of cars, sirens of emergency cars, and alarm signals of railroad crossings, and then displays them as a light signal to the driver, has been developed. The basic operating principle of the device is that those traffic-alarm-sounds have sharp line spectrum (s) whereas the ambient traffic noise is a wideband random signal. The results of simulation experiments and road tests have demonstrated that the performance of the device is satisfactory except in the case of detecting the alarm signal of the railroad crossing.

View full abstract

Jaw movements provide valuable data both for diagnosis and treatment in dental clinic. So many technics and instruments have been developed and tested, but for most of them measurable points on the jaw have been limited.
In order to measure motions of an arbitrary point on the jaw relatively easily, a method that assumed the jaw to be a rigid body has been used in recent years. On this standpoint, a system using light-emitting diodes (LED), position sensitive detectors (PSD) and a mini-computer has been developed. Measuring of the motions of the jaw assumed to be a rigid body was carried out by measuring motions of the three LEDs attached to it rigidly. Motions of an arbitrary point on the jaw could be easily calculated from LED motions, if the position of the point with respect to the LEDs had been measured.
The LEDs and the PSDs were mounted respectively on a lower jaw face-bow and an upper jaw face-bow that were attached to each dental arch using resin clutches. The position of each LED was detected with a PSD camera at every 10ms. Then the voltage signals of two channels relating to the two dimensional position of light spot on the PSD sensing sarface were obtained.
In order to obtain the three-dimensional position of all LEDs, a method using inter-LED distances has been developed. Using characteristics of the optical system and the PSD, a straight line of E³ in which each LED existed has been obtained. Then using inter-LED distances, it has been possible to determine each LED position.
When all considerable factors that may have been causes measurement errors were accounted for, the overall position accuracy proved to be better than 0.13mm and the position resolution was found to be 0.04mm.

View full abstract

It is well known that through a single determination of serum protein electrophoresis, various important clinical and pathological data can be obtained. Usually the interpretation of a serum protein electrophoretic pattern has been made by a physician on the basis of his personal experience.
This paper describes an automatic interpretative system for serum protein electrophoretic pattern according to the method previously proposed by Dr. Kawai. An electrophoretic pattern of serum protein is mainly composed of five componens-albumin, α₁, α₂, β and γ. The data obtained from densitometry of serum protein electrophoresis are automatically put into a microcomputer through a 12-bit A/D converter, and there the data are analysed sequentially as following: Initially, the density of each component is measured. Next, each density is ranked +2, +1, 0, -1, -2. Finally, the ranked values are compared with the diagnostic library of Dr. Kawai's. The patterns are then classified into 12 categories on the basis of Kawai's method.
This system consists of two tables. One is the HL library and the other is the diagnostic discriminant library. The former is the table which shows the threshold levels for determining the ranked values, and the plus and minus values correspond to increase and decrease of respective components. The latter is the table which shows the characteristic patterns of various kinds of disease by means of the ranked values. Furthermore, in order to detect the most frequently encountered monoclonal gammaglobulinemia in the extra band, relative mobility and sharpness of the peak of the serum protein pattern are defined.
As a result, electrophoretic patterns of serum protein have been classified into 12 categories according to the modified method of Kawai's diagnostic table and detection of an extra band has been confirmed in more than 90% of the cases.

View full abstract

Conventional X-ray tomography is a useful method for obtaining a cross-sectional image at a certain depth of the body. But the problem is blurring of the tomogram by the shadows of the organs at other depths. This paper presents a procedure for making the tomogram clear by removing the obstructive shadows.
A tomographic image, which is produced by linear movement of a pair of X-ray source and film, is considered to be formed by two elements, one being the clear tomogram of the specific layer of the target and the other being linear shadows caused by organs existing on the upper and lower sides of the layer. So it is necessary to suppress the shadows in order to obtain a sharp image. The improving procedure is composed of three steps. The first is to digitize the tomogram and transfer it to a computer; the second is to prepare a shadow image by averaging up the tomogram which is shifted pixel by pixel along the linear shadows, and the last is to subtract the shadow image from the original tomogram. This procedure is also shown to be equivalent to the high-pass filtering process in the frequency domain.
This computer-assisted technique presented in this paper is very simple and, therefore, can be applied to any conventional X-ray tomography.

View full abstract

A method for systematically storing the knowledge of drug-to-drug interactions on the basis of formal representation is described. The hierarchy is defined on the set C of drug-categories and drugs by a partial ordering relation induced by category inclusion. Possible interactions between the elements of C are represented by another relation γ called the interacting relation. A partial ordering is further defined on the solution set of γ, and the set of maximal elements is defined as the maximal interacting relation. The elements of C and the maximal interacting relation are represented by nodes and links in a computer. The links joining two nodes are superimposed on all the subordinate nodes. That is to say, the general rules of interactions are applied to the individual drugs. The structured representation makes the database free of duplication and redundancy, which in turn, helps greatly to prevent updating anomalies in future modifications and expansions.

View full abstract