耳 鼻 臨 床 58: 8 557 研 究 Binaural Fusion Test: A Diagnostic Approach to the Central Auditory Disorders. RITsu Hayashi, M. D. Chapter 1: Review of the Literatures on the Clinical Application of Binaural Sum- mation, Integration or Fusion. Two principal methods utilizing binaural hearing phenomenon have been recently developed for the diagnostics of disturbances in the central auditory system: 1) a directional audiometry and 2) a test of binaural fusion or summation with speech or pure tones. A number of papers have been published as to the directional audiometry but will not be reviewed in this article. Hirshl summarized the research works on binaural summating effect during the last hundred years. According to his review, Seebeck2 first discoveredthe binaural summating effect during the experiment on binaural beat. Since then, many investigators have experimented on that subject and some of them have made attempts to study the central auditory dysfunction by means of binaural hearing test with distorted speech. Bocca3)developed a diagnostic method to examine the efficiency of central mechanism in hear- ing. The Italian PB words were delivered simultaneously to both the ears through two independent channels. The non-distorted speech, of which the average discrimination score did not exceed 30 percent, was presented to one ear through Channel 1. The distorted speech (500 cps low pass) was presented to the other at 45 db above the threshold through Channel 2. The results revealed that the discrimination score in binaural presentation was nearly equal to the addition of each monaural score. Calearo4 applied this technique to the diagnostics of central auditory function. He was the first who used the binaural integration for the diagnostic purpose. Ten patients with lesions of the temporal lobe or of the surrounding regions were examined. In most of the cases, the binaural integration was absent when the signals through the Channel 2 were delivered to the ear opposite to the affected lobe. Several years later, Jerger6) used the same technique and ascertained Calearo's result. Matzker67 used binaural fusion of two coherent distorted speech materials as a test index. In Test 1, one ear received the words filtered through the frequency band of 500-800 cps, and the other received simultaneously the band of 1500-2400 cps of the same verbal message. In normal subjects, each band alone yielded poor discrimination, but binaural presentation of the two bands resulted in a good fusion of them and a high discrimination score. In Test 2, the two bands of verbal message were mixed electrically and then presented binaurally, i. e. each ear received the mixed information of the two bands simultaneously. In Test 3, the same procedure as Test 1 was repeated. Then, the numbers of unidentified words in Test 1 or 3 were compared with the num- bers in Test 2. Normal subjects showed no remarkable difference between these two kinds of pre- sentation, but some of the cases of presbycusis and diseases of the central nervous system showed very poor discrimination in Test 1 or 3 in spite of good discrimination in Test 2. On the basis of more than one thousand cases examined, Matzker considered that the poor binaural fusion is attributable to the impairment of the brain stem, especially the medial geniculatebodies. Modifying the Matzker's method, Lindens used the Swedish spondaic words filtered through the bands of 560-715 cps and of 1800-2200 cps. Comparison of discrimination scores was made between monau- * Department of Otolaryngology Faculty of Medicine Kyoto University (Director: MASANORI MORIMOTO, M. D. Instruction: FUMIHIKO OHTA, M. D.) 57 558 Ritsu Hayashi 耳 鼻 臨 床 58: 8 ral and binaural presentations of each band. In the cases of unilateral temporal lobe legion, monau - ral distorted speech yielded poor discrimination in the contralateral ear. The discrimination score in the binaural presentation usually ranged between each monaural score. His results were different from those of Matzker in that poor binaural fusion was not demonstrated in the cases of expand- ing intracranial lesion. Groen and Hellema9 classified sensorineural deafness into two groups according to the steepness of monaural and binaural speech audiograms; 1) if the binaural curve was steeper than each mo- naural one, the lesion might be located somewhere between the cochlea and the superior olivary nucleus and 2) if the binaural curve was parallel to each monaural one, the lesion might be located centrally beyond the superior olivary nucleus. Cherry and Taylor10) studied on the periodical switching of verbal messages. Normal subjects were able to obtain more than 90 percent of intelligibility when the periodically switched speech materials were delivered alternately to both the ears keeping 50 percent on-off ratio at any frequ- ency of switching. Using this tehnique, Calearo11 obtained a good result in detecting a malinger- ing. Bocca12 suggested the possibility that this test could be effective for the differential diagnosis of brain stem pathology vs. corticel or peripheral lesions. Kirikae et a113 tested the aged persons by using a similar procedure and reported that the aged subjects showed poor discrimination in monaural and binaurally switched presentations of interrupted speech. They attributed the results to senile changes in the auditory nervous system from the level of the spiral ganglion to the audi- tory cortex. Many researchers listed above have contributed to the establishment of the diagnostics of im- pairments in the central auditory nervous system with their own methods which were elaborated by utilizing binaural interaction. As to their methods, Calearo used the binaural presentation of distorted speech and non-distorted one of low intensity. In the studies of Matzker and Linden, the speech was filtered through two frequency bands and presented binaurally. The comparison of discrimination score was made by Matzker between the diotic presentation of the signals through the two bands and the dichotic presentation of the signals through the high band and those through the low band, but Linden compared the discrimination scores between the binaural presentation and the monaural presentation of the signals through each frequency band. Groen and Hellema compa- red the steepness of binaural and monaural speech audiograms. In the study of Kirikae et al, the interrupted speech was presented in binaural switching. As to the diseases which were treated by them, Calearo and Linden tested lesions of the brain tumor, especially of the temporal lobe. Matzker, and Groen and Hellema examined the cases of some diseases of the central nervous system and deafness due to many kinds of origin, and the aged persons were tested by Kirikae et al. There are some divergencies among their opinions on the site of binaural interaction, but most of them supposed that the site might be the brain stem at which both auditory pathways crossed each other and that a message might be integrated in the brain stem before being transmitted to the auditory cortices. However, they appear to have no sufficient evidences to make such an ex- planation and it is doubtful that a binaural fusion test could reveal so many disorders as shown by Matzker. The purpose of this paper is to examine the feasibility of the Matzker's method as a diagnostic procedure for the central auditory dysfunction, to modify the method if necessary, and to obtain some information which may give the cue regarding the site of binaural fusion. Chapter 2 Speech as Test Material. What kind of speech is best to be used as the material of binaural test? The language, of 58 耳 鼻 臨 床 58: 8 Binaural Fusion Test: ADiagnostic ApProach to the Central Auditory Disorders. 559 course, must be native or familiar to the test subjects. Bocca, Calearo, Matzker and Linden used disyllabic words in their tests, but Jerger used PB words. Short sentences were employed by Calearo in his study on switched speech. For the binaural speech audiometry with non-distorted material, monosyllables were used by Groen and Hellema. Most of these investigators used meaningful ma- terials besides some of the Bocca's works. On the other hand, Kirikae et al used Japanese nonsense monosyllables as the test material. Horiguti14) made a comparative study of materials in speech audiometry between Japanese and other languages, and emphasized that nonsense monosyllables were best to be used in Japan. Non- sense monosyllables, as he pointed out, are not so unfamiliar to Japanese people. Listener's know- Table 1. The phonetic symbols of Japanese ledge of vocabularies and dialectic difference have 67 monosyllables. less influence on the discrimination if nonsense monosyllables are used, than when meaningful speech materials are used. This shows the reason why they have been actually used as the standard material of the conventional speech audiometry in Japan. They consist of 100 monosyllables, of which 5 are vowels, 62 are consonant-vowel (CV) type and 33 are consonant-semivowel-vowel (CJV) type. Vowels and CV syllables, of which phonetic sym- bols are shown in Table 1, are more familiar and more frequently used than CJV syllables. The 67 monosyllables (5 vowels and 62 CV syllables) cover 90.2 percent in the rate of occurence in con- versation. 14 For these reasons, the 67 monosylla- bles were used as the test material in the author's study. Chapter 3 Selection of Frequency Bands for the Binaural Fusion Test. Matzker's method is based on the binaural fusion of two coherent distorted speech signals of which each alone is unintelligible, but of which simultaneous hearing obtains a good discrimination. The purpose of study in this chapter is to select the frequency bands and the test-syllables which are appropriate to such distortion. The vowels play very important roles in the phonetic structure of the Japanese 67 nonsense monosyllables.