AUDITORY AGNOSIA WITH REI:.ATIVE SPARING OF SPEECH .PERCEPTION J. Lambert, F. Eustache, B. Lechevalier, Y. Rossa and F. Viader (Department of Neurology, University Hospital Center of Caen) INTRODUCTION The perception of sounds, music and language may be impaired in a disso­ ciated fashion. Cases of auditory agnosia predominating in music and non verbal sounds perception have been reported (Albert, Sparks, von Stockert and Sax, 1972; Motomura, Yamadori, Mori and Tamaru, 1986). Others have observed cases of pure word deafness with relative or complete sparing of music and non verbal sounds identification (Albert and Bear, 1974; Naeser, 1974; Denes and Semenza, 1975; Saffran, Marin and Yeni-Komshan, 1976; Auerbach, Allard, Naeser, Alexander and Albert, 1982; Coslett, Brashear and Heilman, 1984). Finally, the auditory impairment may involve all three modalities, resulting either in auditory agnosia (Miceli, 1982) or in cortical deafness (Chocholle, Chedru, Botte, Chain and Lhermitte, 1975; Michel, Peronnet and Schott, 1980). We report another case of post-traumatic auditory agnosia, that evolved from ,a global impairment (both verbal and non verbal stimuli) at onset, to a selective agnosia for non verbal sounds and music, and with no brain damage on magnetic resonance imaging (MRI). CASE REPORT I.D., a 28 year old, right-handed female nurse, suffered a head injury on June 23, 1985. She remained in a light coma for six days, with lower right limb paresis; computerized tomographic (en scans showed a ventricular haemorrhage with no obvious parenchy­ matous brain damage. On recovery from the coma, the patient behaved as if deaf. In addition, she complained of aphonia, swallowing difficulties, rectal incontinence, vertical diplopia and upper right limb clumsiness. Examination revealed cerebellar incoordination of the upper right limb associated with ataxia, and hyperactive tendon reflexes on the right, but no motor or sensory deficit and no Babinski's sign. The EEG was normal. ACT scan performed in July was normal. A new CT scan performed two months after the accident demonstrated ventricular enlargement (with no sub or supratentorial lesions) (Figure 1). Cerebral blood flow was bilaterally and symetricilly decreased. Finally, MRI studies performed 5 months post onset again revealed a ventricular enlargement with no visible damage to the brain itself. Cortex, (1989) 25, 71-82 72 J. Lambert, F. Eustache, B. Lechevalier, Y. Rossa and F Viader Fig. 1 - CT scan performed September 1985 showing an en­ largement of the cerebral ventri­ cles without sub- or supertentorial lesions. INVESTIGATIONS: THE INITIAL PHASE: AUGUST 1985 Audiologic Examinations Psychoacoustic Studies The patient behaved as if completely deaf. Otoscopic examination showed the tym­ panic membranes to be normal. Pure tone threshold audiometry revealed a severe, bila­ teral mixed deafness curve, predominant on the left. During audiometry the patient was unable to understand any of the stimuli at any intensity level, whether on the right or on the left. Localization of sounds was severely impaired. Sounds were emitted at a 2 meter distance from the patient, in one of 6localisations: front right, left or straight, behind right, left or straight. The patient was sitting with eyes closed, and had to point to the sound emission with her finger. She was able to distinguish from behind to front, but not from left, straight or right. Objective Studies lmpedancemeter tympanograms were normal bilaterally. Stapedial reflexes were obtained at normal levels on both side with contra- and ipsilateral stimulation. Waves I-VI of the brainstem auditory evoked potentials were obtained for each side at 65 dB HL with a normal 1-V interval of 4.3 ms. Responses to non filtered clicks (fre­ quencies 2000-4000 Hz) were recorded with decreasing intensities down to 0 dB on each side, which supports an intact peripheral auditory function. Middle latency evoked potentials were recorded on both sides after stimulation with a nonfiltered click of 65 dB HL. Late latency auditory evoked potentials were recorded from the vertex and both temporal areas. Clicks of 1000 Hz at 100 dB HL given to the left or right ear. Waves N ~, P2, and N 2, were recorded in all conditions with latencies of 105-115 ms, 200 ms and 270 ms respectively. P 3 was not identified. Auditory agnosia with relative sparing of speech perception 73 N europsychologic Examinations The neuropsychologic examinations performed 40 days after onset only revealed a mild constructional apraxia (in copying Rey's figure). The patient was very aware of her difficulties, she was cooperative and showed no fatigue or slowing down during the examination. She had no aphasia: spontaneous speech and writing were normal as were Picture-naming, definitions, and sentence-building. Comprehension of written language was assessed using the French Aphasia Battery Test (Ducarne, 1976) and was normal. The auditory perception examinations revealed a major disorder. However, whereas the patient stated that she could hear nothing at all on admission to our department (August 2), she noted a slight improvement a few days later. At that time, she said she could hear noises which "resonated", but which she did not find unpleasant, and she managed to identify a few familiar environmental sounds (telephone ringing, bell ringing, paging of nurses over the loudspeaker system). The patient's musical background was poor. She had taken piano lessons from age 7 to 9, but had not played an instrument since that time and had long forgotten her solfeggio. Sound Perception Her ability to perceive familiar noises, animal calls, musical instruments, music and voices was examined, in a silent room, using tape recordings presented binaurally without earphones. The patient was presented with a recording of 48 meaningful sound sequences com­ prising familiar environmental sounds (30), animals calls (11), and musical instruments (7). The items were 10" long and were presented randomly. The patient was unable either to name or to give verbal evidence that she had identified any of the items. In a meaningful sound matching test, each item of the test was presented to the patient together with a 9 picture display. She was instructed to listen to the noise and then to point to the corresponding picture. Of the 9 pictures, one represented the real source of the sound heard (e.g. piano), one belonged to the same semantic category (guitar in the example given) and 7 were unrelated (hammer, motorbike, steps, scissors, saw, vacuum­ cleaner, birds). She only scored 15/48 at this test. Out of the errors, 2 were acoustically related (e.g. saw mistaken for bird), 3 were semantically related and 23 were unrelated to ~e presented sound. She gave no response for 5 items. · In a meaningfull sound discrimination test, using the same tape, the patient was asked to say whether two successive sequences presented at 10 s interval were identical or not. Her performance in this test was close to normal, with 1 error/12. Music Reception The proposed material consisted of musical sequences lasting 10 to 15 seconds and corresponding to an entire musical phrase. The initial phrases of well-known folk or popular music items were used. All the sequences were played on the open diapason of the organ on a frequency range from 220 to 880 hertz. I.D. was completely unable to name these popular melodies, which she mostly iden­ tified as noise. When asked to point to the title of a melody from a choice of 4 written titles presented before and during presentation of the musical item, she gave random answers. Thereafter, in a melodies discrimination test, the patient was asked to say whether two successive melodies presented at 10 s intervals, were identical or not. She also gave random answers. Environmental Sound/ Music/ Language Discrimination Studies The patient failed completely to discriminate sound categories in a test comprising sequences of music, environmental sounds and speech: music was identified as speech or 74 J. Lambert, F. Eustach.e, B. Lecheva/ier, Y. Rossa and F. Viader noise, environmental sounds as speech or music, and speech as music or noise. Voice Perception When presented with a tape recording comprising men's, women's and children's voices, I.D. made identification errors in all three categories. Rythm Perception The patient performed a test of rhythm reproduction successfully. She was able to reproduce the rhythm sequences emitted by the examiner by tapping on the table with a pencil, such as: Ill Ill II III IIIli III I III I Language Perception Without lip-reading, the patient was unable to understand what was being said to her, and speech was most frequently perceived as noise. Moreover, she did not react when spoken to if she was not able to see the speaker. Nonetheless, on the second day of the examination, she managed, without the aid of lip-reading, to repeat four bisyllabic words out of a series of 15 (cafe, chaussure, voiture, rideau), (coffee, shoe, car, curtain). She also succeeded in pointing to the three presented French oral vowels /i,a,o/ uttered by the speech therapist in a list of oral vowels written before her. Although we were unable to investigate her speech comprehension more thoroughly at that time, a dissociation was already discernible between a profound disorder of environmental sounds and music perception, on one hand, and an impaired speech perception which was already improv­ ing, on the other hand. INVESTIGATIONS: SECOND PERIOD: SEPTEMBER TO NOVEMBER 1985 Audiologic Examinations Psychoacoustic Studies Standard-pure tone threshold audiometry showed a mild bilateral sensorineural hear­ ing loss (15-20 dB) for frequencies ranging from 250 to 8000Hz. Speech audiometry, using the phonetic lists of Lafon (1958) presented binaurally with earphones, revealed distorsions which increased proportional to intensity or with super­ imposed white noise. Tone-decay test: The procedure in this test is to emit a sound at the threshold level at a given frequency without interruption for one minute. If the subject ceases to perceive the sound during this period, its intensity is increased in increments of 5 dB. I. D.'s perception at frequencies of 1000 and 500Hz was normal. However, 15 dB increments were required at 2000-3000 Hz, and these reached 25 dB at 4000Hz. At that frequency, after the one minute test, further testing of tonal threshold necessitated an increment of about 10 dB relative to the threshold at onset. A repeat of the one-minute test at the same frequency required an additional 25 dB increment. Click counting: nonfiltered clicks (squared pulses of 100 11s) were delivered binaurally with earphones at an intensity of 60 dB HL, and at rates increasing from I to 100 clicks per second. Our patient's performance was the same as that of control subjects, and her counting was accurate for 7-8 clicks per second. We were unable to administer a click fusion test as our system does not emit pulses at less than 10 ms interval, whereas fusion Auditory agnosia with relative sparing of speech perception 75 thresholds for normal subjects have been given at 1 to 3 ms (Hirsch, 1975, and Chocholle et al., 1985). However, our patient was easily able to distinguish two clicks emitted at a 10 ms interval. Discrimination of small variations in tonal intensity (Liisher's Test). This loudness discrimination test was performed with a Madsen OB 802 Audiometer at frequencies of 500 and 1000 Hz and a tone of supraliminal intensity of 40 and 50 dB. Maximal modu­ lations in intensity of 5 dB were produced, decreasing in 9 stages to zero. Control subjects perceived intensity variations from 0.7 to 2 dB. The patient perceived a variation of 5 dB at the first modulations only, that is, at the onset of the signal. She could perceive no variation below 5 dB. These results were reproduced several times. Discrimination of small frequency variations: In this test, which follows the same principle as the above, we used frequencies of 500 and 1000 Hz. Maximal modulations of 5% decreases in 9 stages to zero were produced. The patient's results were comparable to those of control subjects; she perceived frequency variations down to the low levels of 0.6 to 0.74%. These examinations thus showed that the patient's ability to discriminate frequency variations was not significantly different to that of normal subjects. However, her appre­ ciation of the differential threshold of intensity was impaired; a deterioration of tone threshold was also observed on certain frequencies. A verbal dichotic listening test was administered to the patient using the procedure of Michel and Peronnet (1975). This test revealed partial left ear extinction. Sound localisation: The patient made the same errors as previously. In conclusion: the audiologic examination, carried out at the first and second phases, show that the auditory pathways from the middle ear to the brainstem are virtually intact. Neuropsychological Examinations The patient was no longer aphonic. Laryngoscopy was normal; however, her voice remained deep and she presented a mild dysarthric disorder suggestive of cerebellar damage. She had great difficulty in singing a melody with or without lyrics, either on verbal command or on imitation. The auditory perception study was performed under the same conditions as the first examination and using the same material. Sound Perception When asked to name recorded sequences of environmental sounds, animal calls and musical instruments, the patient gave 11 correct responses out of 48. She often made comments which were correct, about the pitch, eg: flute: high voices; hens: lower voices, cow: it's low; bird: it's high. She gave no response to 9 items, 13 responses were acous­ tically close, 3 were semantically related and 12 were unrelated to the stimulus. On the meaningful sound matching test, she still scored poorly (28/48). Out of 20 errors she gave no response to 2 items, 11 were unrelated, 5 were acoustically related and 2 were seman­ tically related to the stimulus. Music Perception The patient was still unable to name popular melodies (0120). When asked to point to the written title from an array of 4, she scored only 8/20. When asked to discriminate between two popular melodies heard successively, she showed an improvement, with 2 erroneous responses out of 10; in a similar test involving unknown melodies, she made 4 errors out of 10. In a musical categories identification test (cf Lherrnitte et al., 1971), composed of 10 different types of music (sacred music, symphony, nursery rhymes, jazz, waltz, opera tunes, military marches, music with accordions, ... etc.), she completely failed to identify the different types of music. 76 J. Lambert, F. Eustache, B. Lechevalier, Y. Rossa and F. Viader Environmental Sound! Music/ Language Discrimination Study When asked to distinguish between music, environmental sound and speech, she made errors in the discrimination of music and environmental sounds from each other or from speech. Musical instruments were mistaken for speech or noise, and environmental sounds for speech or music. Only speech was correctly identified. Voice Perception The patient was able to identify children's, men's and women's voices. However, errors persisted concerning the number of voices participating in a conversation and, for exam­ ple, she said that she heard the voice of one child only, when listening to a conversation between two children. Language Perception At this stage of the clinical course, the patient's speech comprehension had improved, and she was able to hold a conversation. However, she said that it was more difficult for her to understand over the telephone or when listening to a tape recorder. Her language perception was investigated using the examiner's voice (female) without a tape recorder or earphones. The examiner sat facing the patient with her lips hidden by a screen. Sentence Comprehension The patient was presented with a verbal comprehension test (Khomsi, 1985). This test enabled to study the following linguistic structures; interrogative, declarative, negative sentences, passivation singular, plural, definite/non definite, future/past, spatial prepo­ sitions, pronouns, relative and completive sentences. The patient was asked to select the picture corresponding to the sentence spoken by the examiner from a choice of 4. She scored 34/36. Moreover she was able to repeat the sentences she heard even when these were long, e.g., "Je mange les cerises que maman cueille" (I eat the cherries, mother is picking). Single Word Comprehension Picture matching test: the patient was given a word orally and asked to point to the corresponding picture, from an array of 6. Within these, one was semantically and another was phonemically related to the target. Eg: target =ruche /rys/ (hive). Multiple choice of pictures: ruche /rys/ (hive), abeille /abej/ (bee), bfiche /bys/ (log), car /karl (bus), chausettes /soset/ (sock), toupie /tupi/ (spinning-top). Eight series were given and the patient scored 7/8. Writtenword-matching test: The patient was asked to point to the corresponding written word in a 4-way multiple choice-test, each series of 4 items consisting of words which differed by one phonetic feature only: voicing (port, bord, fort, sort) /p~r, b:>r, f9r, s xI or place of articulation (pont, thon, long, son) /p.), n, 1'>, s5/. Only two erroneous answers were given for the 21 series presented (84 items). Word repetition test (from the French Aphasia Battery, Ducarne, 1976). Words were of different length, phonemic complexity, and frequency. I.D. correctly repeated 67/91 words. She performed better with plurisyllabic {15% errors) than with monosyllabic (28% errors) or bisyllabic (31% errors) words. Only once did the patient give a non-word as a response: scolaire /skJlET/ spo... ? However, she perfectly realized that her response was meaningless. In 18/20 errors the number of syllabes in the word was unchanged. 15/20 erroneous responses still contained at least 50% of the phonemes of the target: eg: fil-fine /fil, fin/, maigre-mettre /megr, metr/, exceptionnellement-exceptionnel /eksepsjondma, eksepsjond/. The remaining contained less than 50% of correct phonemes: eg: grand-trop /gra, tro/, encore-porte /iik.)r, p.xt/. Phonemic Identification Identification of vowels, stop consonants and fricatives was tested. The consonants 77 Auditory agnosia with relative sparing of speech perception were studies in consonant-vowel combina,tions, with 1m initial consonant paired with the vowel /a/. Consonants The test procedure was as follows: the examiner read aloud a list of 20 syllabes at 4 second intervals with her lips hidden by a screen. The patient was presented with 2-way and 3-way multiple choice tests, and asked to point to the written syllable corresponding to the one she heard. The first trial involved the identification of consonants differing by a single phonetic feature. Thus, for each category (stops and fricatives), we were able to study our patient's ability to identify consonants differing either by the feature of voicing or by the feature of place. The second trial, presented under the same conditions, involved identification of consonants differing by two features: both voicing and place /pa/ da/. Results. Fricative consonants were identified 100 percent correctly. Within stop con­ sonants: when they differed by the feature place, the patient made 25% errors for voiceless (pa/ta/ka) and 2,5% for voiced ones (/ba/,/da/,/ga/). When the consonants differed by feature voice the patient made 1,6% errors (pa/ba; ta/da; ka/ga). When the consonants differed by two features: place + voice (eg: pa/da; ba/ka; ta/ga), the patient scored 100%. Results for stop consonants are summed up in Table I. TABLE I Number of Errors and Error Rate on Stop Consonants Identification Tests Distinctive feature Voicing Place of articulation Voiceless consonants %errors pa/ba Stop consonants ta/da ka/ga pa/ka pa/ta ta/ka 0/20 5/20 1120 3120 0/20 7/20 1,6 25 pa/ta/ka 5/20 ba/ga Voiced consonants 0120 ba/da 1/20 da/ga .1/20 2,5 ba/da/ga Voicing + Place of articulation 0/20 pa/da 0/20 ba/ka 0/20 taiga 0/20 0 Vowels These studies were conducted under the same conditions as the consonant identifi­ cation test. The examiner uttered one vowel every 4 s, and the patient was asked to point to the corresponding written oral or nasal vowel before her. Results. Identification of oral vowels was 100% correct. Within nasal vowels, /a/ was always identified, IF>/ gave rise to 1/20 errors and/£/ to 5120 errors. The patients always discriminated nasal from oral vowels. These tests thus demonstrated a relative sparing of language comprehension. However, the phonemic identification test revealed 1) preserved oral vowel identification, but some 78 J. Lambert, F. Eustache, B. Lechevalier, Y. Rossa and F. Viader errors concerning the nasal vowels; 2) preserved fricative identification; 3) an almost selectively impaired perception of stop consonants, and within this category, a clear predominance of errors concerning the articulatory feature of place, with a 25% error rate for voiceless and 2.5% error rate for voiced stops. The error rate for the feature of voicing was only 1.6%; 4) the performances were better when the tested pairs differed by two rather than by one single phonetic geature. DISCUSSION Our patient had a severe impairment in sound, music and language identif­ ication, which finally strongly predominated in music and sound. However, she still failed to accurately discriminate categories: music was sometimes mistaken for either sound or speech. In that sense, she had a first level musical desinte­ gration, as described by Lechevalier, Eustache and Rossa (1985). Audiometry and brain stem auditory evoked responses ruled out significant damage in audi­ tory pathways. On the other hand, this case cannot be classified as cortical deafness (Michel et al., 1980; Chocholle et al., 1975), since normal cortical auditory evoked responses have been obtained. The patient was able to discriminate between 2 meaningful sounds. On the contrary, she was unable to name them, and scored only 20/48 at a multiple choice matching test. It is worth stressing that erroneous responses were most often acoustically related to the target. Her performances were similar with musical material. She was unable to name well-known melodies and scored 8/20 at a multiple choice-test, but she made only 2/10 errors at a melody discrimi­ nation test. This loss of identification contrasting with sparing of dicrimination ability is consistent with an asemantic-associative agnosia, as described by Faglioni, Spinnler and Vignolo ( 1969) and Vignolo ( 1982). According to them, difficulty in recognition of meaningful non-verbal material is not due "to a perceptual deficit, but it consists, rather, of the inability to associate the auditory percept to a definite source or event". The same authors found that the results of left brain damaged subjects in such associative-tasks were correlated across auditory and visual modalities lending them to support the view that this impairment might be "due to one basic cognitive-associative disorder". A major difference between these and our own patient is that her cognitive impairment involved only auditory material. Although language comprehension was relatively spared, she had difficulties in phonemic identification. The test we used has been given different names: phonemic discrimination test by (Motomura et al., 1986; Chocholle et al., 1975), identification/discrimination test (Auerbach et al., 1982) and phonemic identif­ ication test (Basso, Casati and Vignolo, 1977; Saffran et al., 1976; Blumstein, Cooper, Zurif and Caramazza, 1977; Miceli, 1982). Our patient had difficulties in stop-consonant identification, and mostly made articulatory place confusions with the feature place. Studdert-Kennedy and Shankweiler ( 1970), using dichotic listening in normal subjects, demonstrated a significant right-ear advantage both for overall iden­ tification of stop consonants, and for the articulatory features of voicing and Auditory agnosia with relative sparing of speech perception 79 place in stop consonants. They concluded that auditory processing of a verbal sign may be processed by two systems: "a general auditory system may be equipped to extract the auditory parameters of a speech signal" whereas "the dominant hemisphere is specialized for the extraction of linguistic features from those parameters". Oscar Berman, Zurif and Blumstein (1975) stated that two levels, auditory and phonetic, should be distinguished in speech processing. They suggested that difficulty to perceive feature place of articulation is attributable to its acoustic properties: "while feature voice is closely tied to acoustic invariance, feature place is not invariant and necessites acoustic restructuring of the speech signal for the conversion of the acoustic signal into invariant phonetic paramet­ ers". In this sense, feature place is more encoded than is feature voice and depends on the integrity of both the auditory and phonetic ·levels of speech processing. Discrimination errors on place contrasts are more common than on voicing contrasts (Blumstein et al., 1977b; Blumstein, Baker and Goodglass, 1977a; Miceli, Caltagirone, Gainotti and Payer-Rigo, 1978). Moreover there is a con­ sistent relation between the ability to label and discriminate place of articulation. Studies by Blumstein et al. (1977a), Blumstein, Tarter, Nigro and Statlender (1984), Riedel and Studdert-Kennedy (1985) support the view that the ability to discriminate categories of speech underlies the ability to use them linguistically for labelling and that aphasics demonstrate better discrimination than identifi­ cation. However there is no correlation between phonemic perception ability and auditory language comprehension (Basso et al., 1977; Blumstein et al., 1977a, 1984; Riedel and Studdert-Kennedy, 1985). Patients with pure word deafness have shown heretogeneous performances on phonemic perception tests. Naeser's (1974) and Auerbach et al.'s (1982) patients e?<:perienced more difficulties to perceive feature place, whereas Saffran et al.'s Q1976) had greater difficulties with feature voicing, and Denes and Semenza's (1975) made errors in both. Miceli (1982) reported predominant impairment of feature place identification in a patient with auditory agnosia for language, non verbal sounds and music, as did Chocholle et al. (1975) in a case of cortical deafness. Unfortunately, in cases close to ours, like Albert et al.'s (1972) and Motomura et al.'s (1986), phonemic perception was not investigated thoroughly enough to disclose any specific impairment. The feature place of a articulation thus seem to be highly vulnerable, both in aphasia and in pure verbal deafness, and more generally, in auditory agnosia. This was stressed by Oscar Berman et al. (1975) and by Blumstein et al. (1984) who noted that "the place of articulation contrasts are especially difficult because of the transience of the acoustic cues signaling the phonemic dimension". Auditory difficulties giving rise to a prelinguistic impairment have been observed in auditory agnosia or in pure verbal deafness. For instance the patient of Albert et al. (1972) failed to discriminate pitch, intensity rhythm and duration of auditory stimuli. They attributed this disorder to "an asynchrony of interac­ tion of the auditory impulses, with a delay of auditory impulses" more marked on the right than on the left. Albert and Bear (1974) suggest difficulties in the temporal resolution of auditory stimuli. Their patient had in fact an abnormal 80 J. Lambert, F. Eustache, B. Lechevalier, Y. Rossa and F. Viader fusion threshold for auditory stimuli, and a comprehension deficit than improved at slow presentation rates. Chocholle et al. (1975) found deficits in duration discrimination in their patient, with great variability in fusion thresholds. Auer­ bach et al. (1982) also observed a severe deficit in temporal acuity, demonstrated by click fusion and click counting test. After reviewing previous case studies, they suggested that two distinct types of pure word deafness could be described: type 1, resulting from bilateral temporal damage, in which the deficit is prephonemic and related to a temporal acuity disorder. In this type, the phonemic deficiency predominates in place of articulation. Type 2, resulting from left temporal lesions, with comparable impairment for place of articulation and voicing· per­ ception. Similarly, Motomura et al. (1986) noted abnormal fusion thresholds for clicks and abnormal loudness discrimination, and suggested that, in their case, "temporal summation in auditory processing plays an important role". The tests undertaken in our patient failed to demonstrate any temporal acuity disorder. Auditory testing, however, disclosed deficits during the tone decay and loudness discrimination test. All these results suggest that a variety of psychoacoustic deficits may be involved in impaired perception of sounds and music, as well as in disturbances of the phonemic perception of language and speech. It is worth noting, however, that dissociations may exist in auditory processing. Indeed, whereas Auerbach et al. (1982) explain the prephonemic type of pure word deafness by a temporal acuity deficit, this very deficit is said to be responsible for sound and music agnosia with language sparing in the cases of Albert et al. (1972) and of Moto­ mura et al. ( 1986). Thus, the same disorder may result in distinct clinical patterns, and it may be hypothesized that psychoacoustic abnormalities demonstrated so far in most cases, including ours, are but one element of a larger auditory impairment, to be further investigated by means of more sophisticated auditory tests. In our opinion, the double dissociation: 1 -impaired perception of sounds and music vs relative sparing of language perception and 2 - impaired identif­ ication vs good discrimination of sounds and music, is consistent with the hypothesis of Studdert-Kennedy and Shankweiler (1970), Oscar Berman et al. (1975), Blumstein et al. (1977a) and Riedel and Studdert-Kennedy (1985) that auditory information could be processed in two distinct ways, an auditory and a linguistic one. The linguistic processing might be spared in our patient, thereby allowing a relatively good language comprenhension. In contrast, her difficulties with music and non-verbal sounds, as well as in phonemic identification tasks, could be due to an impairment of the auditory processing mechanism. Actually, our patient's evolution and particularly her improvement in dis­ criminating non-verbal sounds, suggest that are the pathways carrying auditory messages to the site of linguistic processing rather than the auditory processing mechanism itself to be damaged. ABSTRACT We report a case of auditory agnosia in which the initial clinical picture began with generalized auditory agnosia for verbal and non verbal sounds, but rapidly changed to a selective auditory agnosia confined to the perception of non verbal sounds. CT scanning Auditory agnosia with relative sparing of speech perception 81 and MRI did not demonstrate cortical or subcortical damage, except for bilateral ven­ tricular enlargement. The patient was submitted to audiological investigations including physical and psychoacoustic studies. Deficits were revealed during the decay and loudness discrimination test, but no temporal auditory acuity deficit was observed. The results of these studies are discussed in relation to the clinical picture. Also the dissociation between verbal perception and non verbal perception is discussed. REFERENCES ALBERT, M.L., and BEAR, D. Time to understand: a case study of word deafness with reference to the role of time in auditory comprehension. Brain, 97: 373-384, 1974. ALBERT, M.L., SPARKS, R., VON STOCKERT, T., and SAX, D. A case study of auditory agnosia. linguistic and non-linguistic processing. Cortex, 8: 427-443, 1972. AUERBACH, S.H., ALLARD, T., NAESER, M., ALEXANDER, M.P., and ALBERT, M.L. Pure word deaf­ ness: Analysis of a case with bilateral lesions and a defect at the prephonemic level. Brain, 105: 271-300, 1982. BASso, A., CASATI, G., and VIGNOW, L.A. Phonemic identification defect in aphasia. Cortex, 13: 85-95, 1977. BLUMSTEIN, S.E., BAKER, E., and GooDGLASS, H. Phonological factors in auditory comprehension in aphasia. Neuropsychologia, 15: 19-30, 1977a. BLUMSTEIN, S.E., CooPER, W.E., ZURIF, E.B., and CARAMAZZA, A. The perception and production of voice-onset time in aphasia. Neuropsychologia, 15: 371-383, 1977b. BLUMSTEIN, S.E., TARTER, V.C., NIGRO, G., and STATLENDER, S. Acoustic cues for perception of place of articulation in aphasia. Brain and Language, 22: 128-149, 1984. CHOCHOLLE, R., CHEDRU, F., BOTTE, M.C., CHAIN, F., and LHERMITTE, F. Etude psychoacoustique d'un cas de "surdite corticale". Neuropsychologia, 13: 163-172, 1975. COSLETT, H.B., BRASHEAR, H.R., and HEILMAN, K.M. Pure word deafness after bilateral primary auditory cortex infartcs. Neurology, 34: 347-352, 1984. DENES, G., and SEMENZA, C. Auditory modality-specific anomia: evidence from a case of pure word I deafness. Cortex, 11: 401-411, 1975. DucARNE, B. Examen de I'Aphasie. Paris: Centre de Psychologie Appliquee, 1976. FAGLIONI, P., SPINNLER, H., and VIGNOW, L.A. Contrasting behavior of right and left hemisphere­ damaged patients on a discriminative and a semantic task of auditory recognition. Cortex, 5: 366-369, 1969. HIRSH, I.J. Temporal aspects of hearing. In D.B. Tower (Ed.), The Nervous System vol. 3 - Human . Communication and Its Disorders. New York: Raven Press, 1975, pp.·l57-162. KHOMSI, A. Comprendre un enonce, strategies de comprehension du jeune enfant - Travaux de linguistique n° 1; Department de psychologie- Universite de Nantes 1, 1985. l:JsciffivAr.IER, B., EUSTACHE, F., and ROSSA, Y. Les troubles de Ia perception de Ia musique d'origine "Beurologique. Rapport de Neurologie. Paris: Masson, 1985. LHERMITTE, F., CHAIN, F., EscouROLLE, R., DuCARNE, B., PILLON, B., and CHEDRU, F. Etude des troubles perceptifs auditifs dans les lesion temporales bilaterales. (A propos de 3 observations dont 2 anatomocliniques). Revue Neurologique, 124: 329-351, 1971. MICELI, G. The processing of speech sounds in a patient with cortical auditory disorder. Neuropsy­ chologia, 20: 5-20, 1982. MICELI, G., CALTAGIRONE, C., GAINOTII, G., and PAYER-RIGO, P. Discrimination of voice versus place contrasts in aphasia. Brain and Language, 6: 47-51, 1978. MICHEL, F., and PERONNET, F. Test dichotique auditif verbal (communication personnelle), 1975. MICHEL, F., PERONNET, F., and ScHorr, B. A case of cortical deafness: clinical and electrophysio­ logical data. Brain and Language, 10: 367-377, 1980. MOTOMURA, N., YAMADORI, A., MoRI, E., and TAMARU, F. Auditory agnosia: analysis of a case with subcortical lesions. Brain, 109: 379-391, 1986. NAESER, M. The relationship between phoneme discrimination, phoneme picture/perception and language comprehension in aphasia. Presented at 12th Annual Academy of Aphasia Meetings, Warrenton, Virginia, October 1974. OSCAR-BERMAN, M., ZURIF, E.B., and BLUMSTEIN, S. Effects of unilateral brain damage on the processing of speech sounds. Brain and Language: 2, 345-355, 1975. J. Lambert, F. Eustache, B. Lechevalier, Y. Rossa and F. Viader 82 RIEDEL, K., and STUDDERT-KENNEDY, M. Extending formant transitions may not improve Aphasic's perception of stop consonant place of articulation. Brain and Language, 24: 223-232, 1985. SAFFRAN, E.M., MARIN, 0., and YENI-KOMSHAN, G. An analysis of speech perception in word deafness. Brain and Language, 3: 209-228, 1976. STUDDERT-KENNEDY, M., and SHANKWEILER, D. Hemispheric specialization for speech perception. Journal of Acoustical Society of America, 48: 579-594, 1970. VIGNOLO, L.A. Auditory agnosia. Philosophical Transections of the Royal Society of Landon: B 298, 49-57, 1982. J. Lambert, Service de Neuro1ogie; Pr. B. Lechevalier, CHU de Caen, 14031 Caen, Cedex, France.