The author developed the new system of discrimination test and the results of test were processed by digital computer. The processing was based on the acoustical feature of the speech sounds. The data of the normal hearing and the deaf children (6-10 years old, mostly sensori-neural hearing loss and partially mixed hearing loss) were analysed. The following results were obtained. 1. It was found that the order of articulation performance among the consonant groups had a relation with the type of audiogram and its tendency depended on the acoustical feature of speech sounds. It was thought that the order of articulation performance could be used effectively as the criterion of hearing aid fitting and auditory trainning program. 2. For the full automatic system of the speech discrimination test, it is necessary to develop a method that any subjects can easily respond. 3. This new system can save many manpowers on making confusion matrix etc., and the well formed pictures can be displayed.
The authors developed a new type of speech audiometer by which we could automatically measure speech reception and discrimination according to the testing program based on the standard manual for speech audiometric procedure established by the Japan Audiological Society. The characteristics of the equipment were as follows: 1) The equipment consisted of a micro-computer which performs the testing program, controller of input-output data, specific memories, cassette tape recorder, character display panel and console key terminal. 2) The examinees responded to the audible or inaudible test speech signals recorded in a cassette tape, by pushing the terminal keys which were arranged in order of the Japanese alphabet. 3) The information from the key terminal of the examinees and the processed scores were shown on the character display panel. 4) The equipment could be used manually as a conventional speech audiometer.
The authors developed an equipment which automatically proceeds a program of auditory screening test for preschool children. The characteristics of this equipment were as follows: 1) This equipment consisted of touch switches operated by finger of children, micro-computer, data memories, signal controller and character display panel. 2) The response signal was derived from the touch switches and each of them was covered with picture card. 3) The test procedure was automatically performed by one tester. 4) All obtained data were punched on paper tape. 5) One tester could examine two children at one time. 6) The testing time per one child was within two minutes. 7) The storage and retrieval of obtained data was performed by another' large-scale computer.
In the Hitachi General Hospital, the Hitachi General Health Testing Center was established a few years ago. The various examination systems in this center are computerlized with “On Line System”, and about 50 subjects are examined in. The automatic audiometer was developed for the purpose of screening test, and the test frequencies consist of 250, 500, 1, 000, 2, 000, 4, 000 and 8, 000Hz. These frequencies are presented at of 20dB and 40dB (JIS standard). The audiometer is housed in a cabinet with a taperecorder for explanation of the test to the testees. The instruction for each procedure is displayed on the front panel of the cabinet. The test takes about 270 seconds starting from the time testee sits in front of the cabinet and put an ID card into the card reader to the end. The examination data are transmitted to the computer system at the moment the test was completed, and recorded in the system. These procedures are automatically performed according to the instructions from the computer system without manual operation. The Center was opened in August, 1974 and more than 8, 000 subjects were examined up to January, 1976.
A system for automatic data storge and retrieval of audiograms by digital computer was developed. This system consists of a full-automatic computerised audiometer, Békésy type audiometer, storage type CRT graphic display which is attached to a hard copy and digital computer. The operation of this system was mainly performed on the basis of on-line real time processing. The data of airbone-conduction and Békésy audiometry were stored on a disk in the computer. Then, various commands of requirements for retrieval and output as specified by physicians were given to the computer, and the output response to them were promptly shown on the CRT graphic display. This system had several advantages as follows: 1) The displayed patterns of both audiograms were just the same as the designs and symbols of the standard audiogram. 2) A key-puncher to make punch tapes was unnecessary. 3) The measured data of both audiometries for about 1100 subjects could be stored in one magnetic disk. 4) It was observed that the data were retrieved by the computer without any faults in about 10 seconds.
A system was made to process the data of patients with hard of hearing, and to make statistical process by combining the multi-card selector and electronic calculator which is capable of program calculation. This is to report the fact that the system was useful to obtain correlations between certain diseases and the results of hearing test. This system is characterized as the followings: 1) Operation is easy and requires no specialist. 2) Many ready-made programs are available for calculator. 3) Program can be altered easily according to the purpose of study. 4) Low price of apparatus (lower than ¥2, 000, 000). 5) By using the system, studies on various objects and methods can be performed simultaneously and easily. A disadvantage of this system is that the complicated numeral data must be turned to simple ones (1 or 2 fingers) to make the process easier.
A newly developed programming language “MUMPS” was employed to process data derived from audiological examinations. A short test program was written, and the steps of data acquisition, retrieval and processing were executed on limited materials. It was shown that MUMPS is useful particularly because of its simplicity. The ease of program modification also seemed to be a meaningful feature. It was pointed out on the other hand, however, that this language is quite incompetent for complex mathematical processing. It was suggested that an appropriate programming language should be chosen according to the nature of the desired analysis, before the input structure of the data is designed.