BRAIN AND LANGUAGE 28, 86-94 (1986) An Unlearned Foreign “Accent” in a Patient with Aphasia NEILL Division of Behavioral WILLIAM Neurology, R. GRAFF-RADFORD Department of Neurology, University of Iowa E. COOPER AND PATRICIA L. COLSHER Department of Psychology, University of Iowa AND ANTONIO Division of Behavioral Neurology, R. DAMASIO Department of Neurology, University of Iowa Rarely, aphasics may develop what appears to be a foreign accent, as noted by Monrad-Krohn whose Norwegian patient sounded German. We describe a right-handed native American who developed a foreign accent following damage to the left premotor region and white matter anterior to the head of the left caudate nucleus. Her aphasia was of the transcortical motor type. Both she and her parents were born in the USA, she never traveled outside the country and never learned a foreign language. Phonetic analysis of her voice taped prior to the stroke revealed normal speech with a midwestem accent. In contrast, analyses of her current spontaneous speech, repetition, and reading reveal (a) shifts in vowels, e.g., /I/+/i/, /r-e/-+/a/; (b) Increased diphthonigization; and (c) tense speech posture. These features, which were especially frequent in spontaneous speech, probably explain her “accent.” Acoustic analysis of fundamental frequency contours of sentences read in different emotional tones revealed a restricted range and variability of the peaks and valleys. o 1986 Academic PWSS, h. In 1919 Pick reported the case of a 26-year-old Czech butcher whose speech sounded Polish after he suffered a stroke. The patient was still aphasic at the time of examination. His speech was fluent but hesitant and agrammatical with multiple paraphasic errors. He read poorly and The authors thank Ms. Jean Hulme for help in the preparation of this manuscript. Address reprint requests to Dr. Neil1 R. Graff-Radford, Department of Neurology, University of Iowa Hospitals and Clinics, Iowa City, IA 52242. This article was supported by NINCDS Program Project Grant PO NS 19632-02. 86 0093-934X/86 $3.00 Copyright All rights 0 1986 by Academic Press, Inc. of reproduction in any form reserved. AN UNLEARNED FORElGN “ACCENT” 87 could only write his name. He had difficulty naming objects and repeating unfamiliar words. He counted slowly but correctly. Pick thought that the patient’s accent sounded Polish because of a softer pronunciation of the consonants /s/, /z/ and /cl, stress on the penultimate rather than the first syllable, and shortening of all vowels. These are reportedly features of Polish as opposed to Czech. The purity of the case was clouded, however, because prior to his stroke the patient had been on an Army tour of duty in Poland. Monrad-Krohn (1947) reported a similar case in a Norwegian aphasic. The patient, a 30-year-old woman, was hit by a bomb fragment during a German air raid in 1941. A fronto-temporo-parietal lesion was determined by skull radiography and air encephalography. Ironically, the patient’s newly acquired accent was perceived as German and she was denied service by shopkeepers because of their bitterness toward Germany. The patient was apologetic about her accent and could not inhibit it. She was examined two years after her injury at which time she was still moderately aphasic. She had fluent speech, paraphasic errors, mild agrammatism, and also made paragraphic errors when writing. According to MonradKrohn, the acquired accent was characterized by the adoption of German prosodic patterns, including differences in word-level stress and sentencelevel intonation contours. For example, the patient overemphasized utterance-final pronouns and produced them with a rising pitch, violating the prosodic patterns common to her Norwegian dialect. Whitaker (1982) studied a 30-year-old woman who developed a “foreign accent” after recovering from a stroke that initially caused an expressive Broca’s type of aphasia. She was thought to sound Spanish or German. The features of her accent were monophthongization, nonreduced vowels, nonelided word boundaries, lack of control of voice onset and offset, the fronting and raising of non-back vowels, and the fronting of alveolar consonants. The patient made the introspective observation that, when she read “it goes through my mind, I guess like it used to,” implying, as Whitaker pointed out, an “internal acoustic image consistent with her former dialect. ” Unfortunately the patient did not have a radiological study to localize the lesion. There are other brief reports of unlearned foreign accents after brain injury or stroke (Aronson, 1980; Critchley, 1962; Whitty, 1964). These are best summarized by Aronson (1980). We report here a case of a 56-year-old female patient presenting with transcortical motor aphasia who developed a foreign accent. Unlike the cases reported by Monrad-Krohn, Pick, and Whitaker, however, the accent acquired by the patient is primarily attributable to a shifting of vowel quality. The case is noteworthy because most patients presenting with similar aphasias do not exhibit such an accent. 88 GRAFF-RADFORD ET AL. CASE STUDY A. History and Physical Examination On December 31, 1983, a 56year-old right-handed library worker awoke feeling ill. She had difficulty buttoning clothes with the right hand and found she could not speak. She was admitted to the hospital with the diagnosis of a stroke. There was no history of cardiac disease, hypertension, diabetes, or excessive alcohol use. The patient’s father and his two brothers, however, had myocardial infarctions and she had smoked two packs of cigarettes daily for 20 years. Her general examination was normal. The neurological examination revealed that the patient was alert and aware of her problem and surroundings. The neurologist noted that her speech was sparse, dysarthric, and labored. She named two of three objects and obeyed simple commands. Her writing contained misspellings and was incomprehensible. Her cranial nerves were intact. There was a mild pronator drift and impaired rapid alternating movements of her right arm. Her reflexes were symmetrical and plantar responses were flexor. The sensory examination was normal. A CT scan 1 week later showed a left frontal infarction. The patient began speech therapy with a therapist, born and raised in the midwestern United States, and was discharged 10 days after admission. In January 1984, her family, friends, and strangers began to say that she spoke with a foreign accent. In fact, those who did not know the patient would ask from which country she came. When she was referred to the University of Iowa Hospitals in August, 1984, her examination also showed a mild right hemiparesis. B. Analysis of CT Scan A second, chronic state, fourth generation CT scan revealed an area of lucency in the left frontal lobe (Fig. 1) caused by an infarction. The images were plotted on an appropriate set of templates according to a standard method in our laboratory (H. Damasio, 1983). Brodman’s area 6 on the left, as well as the deep white matter anterior and superior to the head of the left caudate nucleus, were damaged. C. Neuropsychological Tests Neuropsychological testing was performed in two sessions. She was strongly right handed on handedness questionnaire (Oldfield, 1971). She had a 12th grade education and worked as a serials coordinator in a library for 8 years. On the Benton Orientation Task (Benton, Hamsher, Varney, & Spreen, 1983) she was appropriately oriented to time. She scored 11 on the arithmetic and 12 on the block design subtests of the WAIS-R (Wechsler, 1981). She scored in the 85th percentile on the Benton Facial Recognition Test (Benton & Van Allen, 1968) and 33/36 on the copy of the Rey-Osterrieth figure (Osterreith, 1940). With regard AN UNLEARNED FOREIGN “ACCENT” 90 GRAFF-RADFORD ET AL. to language, the patient was fluent and only rarely made paraphasic errors. However, she did produce numerous pauses in her speech. On the Multilingual Aphasia Examination (Benton & Hamsher, 1978) she scored 60 for visual naming (superior), 14 for sentence repetition (superior), 27 for the controlled oral word association (15th percentile), 43/44 for the token test (67th percentile), 9/l 1 on oral spelling (normal). Her spontaneous writing and her writing to copy and dictation were accurate but very slow. Her reading was slow with several errors, but she scored 9/10 on the reading comprehension subtest of the Boston Diagnostic Aphasia Examination (Goodglass & Kaplan, 1983). Her reading deficit was best seen on the Chapman Cook speed reading tests (Chapman, 1924) where she scored 9 of a possible 25 in the allotted time (<5 percentile from available norms). She also said she had tried to go back to her library job but could no longer perform adequately because of slowed reading. D. Analysis of her Foreign “Accent” Background. The main deficit regarding the patient’s speech was her foreign accent, which sounded Nordic to her family and friends. (However, when we played her taped voice to a Swedish person he said that she did not have a real Nordic accent). She was born in Jacksonville, Illinois, and until age 20 lived in Winchester, Indiana, and in Chicago. At age 20 she moved to Moline, Illinois, and has lived there ever since. Both her parents were born and raised in the U.S.A., as were her paternal grandparents and maternal grandmother. Her maternal grandfather was from England but died before the patient was born. The patient lives in a community consisting mostly of native speaking Americans. She speaks no foreign language and has not travelled outside the United States. We were able to obtain a brief recording of the patient in conversation with her grandson 6 years prior to the stroke. She had no foreign accent at that time, confirming the patient’s and her family’s report that the accent was newly acquired. Prosodic tests. The patient was given a battery of prosodic tests. In comprehension of sentences spoken by an actress to sound angry, happy, sad, questioning, or neutral, she scored 20/20; on comprehension of prosodic expressions which have meaning (e.g., ah ha! meaning “now I see”) she scored 12/12, and in discriminating same from different prosodic expressions she scored 1 l/12. Two speech production tests were also administered. The first, a mood production test, required the subject to read the four sentences “Tomorrow I’m leaving for Chicago,” “The bird flew away,” “We sold our cottage last month,” and “My horse jumped over the fence,” in an angry, happy, sad, questioning, and neutral manner. There were thus a total of 20 items, presented in a random order. The second test was intended to induce the subject to place tonic accent on certain words of sentences AN UNLEARNED FOREIGN “ACCENT” 91 by asking her specific questions. For example, consider the sentence, “The boy took the candy.” The subject might be asked “Who took the candy?” or “What did the boy take. 3” in order to induce emphasis on respectively. The base sentences for the accent “boy” and “candy,” test consisted of four assertions with three tonic accents each and eight questions with one or two tonic accents each, for a total of 25 items. The subject’s utterances were recorded on magnetic tape (Ampex 632) and submitted to acoustical analysis using the FPRD (fundamental period) computer program (Cooper & Sorensen, 1981). Of particular interest were the overall duration of each utterance and the range and standard deviations of peaks and valleys of the speaker’s fundamental frequency. In addition, on the tonic accent test, the peak fundamental frequency and duration of all words in both accented and unaccented circumstances were measured. The acoustical analysis of the mood production test indicated that, in general, the overall duration of the subject’s utterances tended to be longer than those of normals by almost 30%. Although this effect was largely due to the insertion of pauses between words, the subject also had a tendency to lengthen certain words, particularly utterance final words, by as much as 11%. The ranges and standard deviations of peak and valley fundamental frequencies were normalized by dividing them by their respective means. All four of these measures were considerably below those obtained in normals with the valley measures relatively more depressed than the peak measures. The normalized standard deviations of the peak and valley fundamental frequencies were 70 and 45% of normal, respectively, and the normalized peak and valley ranges were 68 and 46%, respectively, of normal. Further, the terminal peak fundamental frequency rise usually associated with questions was approximately 60% of the magnitude of rise that we normally obtain. Not surprisingly, the above production abnormalities coincided with a moderate degree of difficulty experienced by listeners in correctly identifying the intended emotion. W.E.C., a trained phonetician, perceived that the patient’s rendition of the sentences sounded correct for only 10 of 20 sentences. The incorrect sentences sounded neutral. The subject’s performance on the tonic accent test was also abnormal. Of the 25 items on this test, the patient was judged not to accent properly the target stimulus six times. In addition, on three occasions she placed tonic accent on two words in a sentence when she was required to accentuate only one, e.g., for the sentence “I ate the apple” she was asked “What did you do with the apple”? She replied “I ate the apple, emphasizing both “ate” and “apple.” The accent on apple was preceded by a pause. Unnecessary pauses were frequent in this test, as in the other production tasks. The acoustic analysis supports the above findings in that, when the tonic accent was identified as incorrect, the target word did not have the same 92 GRAFF-RADFORD ET AL. magnitude of increased duration and/or peak fundamental frequency, characteristically found with stressed words. The patient was recorded in free conversation, performing repetition of sentences and reading a passage. These recordings were analyzed by WEC. The patient’s speech was characterized by sporadic, consistent but not invariable, vowel shifts. The most common shift involved /I/-+/i/, as in “teal” for “till,” “deed” for “did, ” “deeference” for “difference,” “treep” for “trip,” “speet” for “spit.” Other vowel shifts included the following: /se/+/a/ /u/+/U/ /a/+/o/ /a/+/o/ as in /hav/ for “have”, “bock” as in “ruhler” for “ruler” as in “hosepital” for “hospital” as in “soam” for “some.” for “back” In most instances, the vowel shifts involved movement toward the cardinal vowels, representing those vowels that are articulatorily and acoustically most distinct in standard two-dimensional vowel space (Abercrombie, 1967). In addition, vowel shifts involved movement toward longer vowels, (e.g., /I/+/i/), and some vowels which retailed their phonetic quality were prolonged by the speaker. The patient also produced diphthongs in environments where a single vowel would ordinarily be appropriate. It is difficult to provide a uniform explanation for these various vowel shifts. One possibility is that the patient’s vowel shifts were attributable to a generally tense speech posture, arising from a spastic dysarthria associated with damage to the upper motor neuron system. It must be noted, however, that the upper motor neurons to the speech masculature were probably not involved by the lesion (Fig. 1). In physiological terms, the tensity attribute generally refers to an increase in isometric muscle tension, particularly manifest in the pharynx and vocal cords (Hardcastle, 1973; Perkell, 1969; Raphael, 1971). In the absence of physiological recordings from the patient’s speech musculature, the perceptual impression of tensity must be viewed as a guiding hypothesis. Tensity does appear to provide a possible organizing principle to account for otherwise disparate results of our study, including the prominence of the /I/+/i/ vowel shift and the relatively restricted range of peaks and valleys in the speaker’s fundamental frequency. Alternatively, the vowel shifts involving nonlow vowels might be attributable to an advancing of the tongue root (e.g., Abercrombie, 1967; Chomsky & Halle, 1968). Otherwise, the spontaneous speech of this patient was quite fluent and contained few pat-aphasic errors (e.g., “smoke” for “spoke,” subsequently corrected). There were also a few grammatical errors, e.g., “I born and raised in Illinois,” “Then I got improved,” and “I not drive.” Oral reading was somewhat slow but fluent and typified by fewer vowel shifts, producing the percept of reduced foreign accent. In repetition, the patient’s AN UNLEARNED FOREIGN “ACCENT” 93 speech included long pauses, usually appearing at phrase boundaries. As with oral reading, accent was reduced during repetition. COMMENT This case study is different from those previously reported because (a) a prior recording of the patient’s voice was available, (b) the patient’s lesion was localized on CT, (c) speech errors and acoustical patterns of speech were analyzed on various tasks and occasions, on the basis of a tape recording. While some of the patient’s most salient speech characteristics can be accounted for by her adoption of a tense speech posture, it is unclear why the speaker maintains this posture in the first place. We compared the patient’s chronic CT to the chronic CTs of 15 other patients with transcortical motor aphasia, none of whom had an acquired “foreign accent.” As noted in previous studies, the anatomical correlate of transcortical motor aphasia (TMA) is varied, including instances of damage to left dorsolateral cortex, generally involving areas 6 and 8, to a greater or lesser extent, and also variably involving underlying white matter (H. Damasio, 1981; Damasio & Geschwind, 1984; Freedman, Alexander, & Naeser, 1984). Involvement of Broca’s area is generally absent or is partial only. In some instances, the lesion is exclusively subcortical in white matter (H. Damasio, 1981). In two of our cases the lesions were fairly close to the one seen in our patient but neither the size nor the precise location were equal. For this reason, the role that the specific locus of the patient’s lesion might have played in producing the accent cannot be determined at this time. Additional patients need to be studied in order to answer this question. Nonetheless, the frontal lobe involvement may account for the patient’s greater difficulty in spontaneous speech than in reading and repetition, based on the notion that dorsolateral frontal lobe damage is likely to impair tasks that require motor programming (A. Damasio, 1979). The frontal lobe site might account for the patient’s apparent deficit in speech planning at the sentence level, evidenced by numerous speech pauses and improperly located stresses in the emphasis test. REFERENCES Abercrombie, D. 1967. Elements of general phonetics. Chicago: Aldine. Aronson, A. E. 1980. Clinical voice disorders: An interdisciplinary approach. New York: Thieme-Stratton. Pp. 119-122. Benton, A. L., & Hamsher, K. 1978. Multilingual aphasia. Iowa City: Department of Neurology, University Hospitals. Benton, A. L., Hamsher, K., Varney, N., & Spreen, 0. 1983. Contributions to neuropsychological assessment. New York: Oxford Univ. Press. Benton, A. L., & Van Allen, M. W. 1968. Impairment in facial recognition in patients with cerebral disease. Cortex 4, 344-358. Chapman, J. C. 1924. Educational test bureau. Minneapolis: Educational Publishers. 94 GRAFF-RADFORD ET AL. Chomsky, N., & Halle, M. 1968. The sound pattern of English. New York: Harper & Row. Cooper, W. E., & Sorensen, J. M. 1981. Fundamental frequency in sentence production. New York: Springer-Verlag. Critchley, M. 1962. Regional accent, demotic speech and aphasia. Antwerp: Livre Jubilaire de Ludo van Bogaert. Pp. 182-191. Damasio, A. R. 1979. The frontal lobes. In K. Heilman and E. Valenstein (Eds.), CIinical neuropsychology. New York: Oxford Univ. Press. Pp. 360-412. Damasio, A. R., & Geschwind, N. 1984. The anatomical basis of language: Recent advances. Annual Review of Neurological Science 7, 127-147. Damasio, H. 1981. Cerebral localization of the aphasias. In M. T. Sarno (Ed.), Acquired aphasi. New York: Academic Press. Pp. 27-50. Damasio, H. 1983. A computed tomographic guide to the identification of cerebral vascular territories. Archives of Neurology 40, 138-142. Freedman, M., Alexander, M. P., & Naeser, M. A. 1984. Anatomic basis of transcortical motor aphasia. Neurology 34, 409-417. Goodglass, H., & Kaplan, E. 1983. The assessment of aphasia and related disorders. Philadelphia: Lea & Febiger. Hardcastle, W. J. 1973. Some observations of the fense-lax distinction in initial stops in Korean. Journal of Phonetics 1, 263-272. Monrad-Krohn, G. H. 1947. Dysprosody or altered “melody of language.” Brain 70, 405415. Oldfield, R. C. 1971. The assessment of analysis of handedness. The Edinburgh inventory. Neuropsychologia 9, 97-l 13. Osterreith, P. A. 1940. Le test de copie dune figure complexe. Archives de Psychologie 30, 206-356. Perkell, J. S. 1969. Physiology of speech production: Results and implicaiions of a cineradiographic study. Cambridge, MA.: MIT Press. Pick, A. 1919. Uber Anerungen des Sprach-characters als Begleiterscheinung aphasischer Storungen. Zeitschrif fuer de gesamte Neurun. Psychologie 54, 230-241. Raphael, L. J. 1971. An electromyographic investigation of the tense-lax feature in some English vowels. Status report on speech research, Nos. 25/26. Haskins Laboratories. Pp. 131-140. Wechsler, D. A. 1981. Wechsler adult intelligence scale. Revised ed. New York: Psychological Corporation. Whitaker, H. 1982. Levels of impairment in disorders of speech. In R. N. Malatesha & L. C. Hartlage (Eds.), Neuropsychology and cognition, Vol. 1. The Hague: Nijhoff (Nato Advanced Study Institutes Series D, No. 9). Pp. 168-207. Whitty, C. W. M. 1964. Cortical dysarthria and dysprosody of speech. Journal of Neurology, Neurosurgery & Psychiatry 27, 507-510.