Research

in LkvelqnnentalDhbilities, Vol. 13.pp.533-551.1992 0891-4222192 $5.00 + .OO Printedin the USA.All rightsremvcd. Copyright0 1992PergamoaResl Ltd.

Hearing Abilities of Down Syndrome and Other Mentally Handicapped Adolescents Michael M. Marcell Collegeof Charleston

Stuart Cohen CharlestonSpeech and Hearing Center

Down syndrome(OS) individuals are prone w audiwry processing digCulties in a varietyof audiological.short-termmemory,and language tasks.The present studyexplored: (a) the hearing capabilitiesof DS adolescentsand young adultsrelativew a sample of non-DSmentallyhandicapped(MH) individuals, and (b) the relationshipbetween hearing abilityand performance on several cognitivetaks. Samplesof 26 DS and 26 MH trainablementallyhandicapped in&iduals wen matchedon IQ and chronologicalage (C-4).Audiometrica&a revealedgreater Irearinglossesin DS than in MI individualsat five of the six tested frequencies, more DS conductiveand mixed hearing losses,and particularly high DS losses in the high-fnquency range. Measurementof the speech receptionthresholdrevealed poorer receptionof speech by the DS thanthe MH group. Classificationof tympanogramsindicatedfewer normal ears and twice as manyDS ears withmiddleear problemsreflectingno mobilityor retraction of the tympanicmembrane.Presenceof DS middleear dficulties wasalso confirmed by poorer elicitationof the acoustic (stapedius)rejlex in DS subjects. Correlationof hearing variableswithseven cognitivetasksadministeredon the same day revealedonlyone significantrelationshipafter the statisticalremoval of the effects of CA and IQ: DS subjectswithpoorer hearing identifiedfewer Portions of this paper were presented at the annual meeting of the American Psychological Association in Boston, August, 1990. Special thanks are extended to PatriciaWeathersand Peggy Wiseman of the Charleston Speech and Hearing Center and Pamela Croen and David Sewell of the College of Charleston for their assistance in gathering the audiological and cognitive data, respectively. Financial support for this research was provided by an Academic Research EnhancementAward (#HD25793) tu the fust authorfrom the National Instituteof Child Health and HumanDevelopment Requests for reprints should be sent to Michael M. Marcell, Department of Psychology, College of Charleston,66 George Street,Charleston,SC 29424. 533

534

M. M. Marcel1 and S. Cohen

wonis in a taskin whicha maskingnoise quicklyfolloweda spokenword This resultsuggeststhatgroupfindings of DS auditory-verbal processingdificulties stem, in part,from subjectswithhearing dl@iiultiesneeding more time to identifyspokenwonis.

Considerable evidence has accumulated in three separate literatures - cognitive psychology, language development, and audiology - to suggest that Down syndrome (DS) individuals have difficulty processing auditory stimuli. Their difficulties have been documented in tasks that focus on auditory short-term memory (e.g., Marcel1 & Weeks, 1988; McDade & Adler, 1980; Varnhagen, Das, & Vamhagen, 1987). successive processing ability (e.g., Ashman, 1982; Hartley, 1982, 1985; Snart, O’Grady, & Das, 1982), language comprehension (e.g., Bridges & Smith, 1984; Burr & Rohr, 1978; Marcell. Croen, & Sewell, 1990). and language expression (e.g., Andrews & Andrews. 1977; Chapman, Schwartz, & Kay-Raining Bird, 1989; Cornwell, 1974; Marcell, Harvey, & Cothran, 1987; Miller, 1987). Furthermore, research in the realm of audiology indicates the presence of mild to moderate hearing impairments and conductive (middle ear) problems in a majority of the DS population (Balkany. 1980; Brooks, Wooley, & Kanjilal, 1972; Buchanan, 1990; Cunningham & McArthur, 1981; Davies & Penniceard, 1980; Downs, 1983; Fulton & Lloyd, 1968; Glovsky, 1966; van Gorp & Baker, 1984; Keiser, Montague, Wold, Maune, & Pattison. 1981; Schwartz & Schwartz, 1978; Wilson, Folsom, & Widen, 1983). Few psychologically oriented investigators of DS cognition (including the present authors) have seriously considered hearing loss in relation to their research. This problem may stem from a general lack of familiarity with the terminology and concepts of audiology, an inability to meet the time constraints and economic costs associated with comprehensive audiological evaluations, and a tendency to rely on readily-available school records of simple pure-tone screenings - a procedure that may underestimate DS hearing loss (cf. Cunningham & McArthur, 1981). The present report describes findings from the fast year of a 3-year longitudinal study undertaken to explore relationships among auditory memory, language, and hearing abilities in a population of DS adolescents and young adults. Two questions are addressed in this report. First, are the hearing abilities of the DS group different from those of a non-DS mentally handicapped (MH) comparison group matched on age and intelligence? Use of an MH comparison group allows one to determine whether hearing difficulties are more closely associated with mental retardation in general or Down syndrome in particular. Also, as Dahle and McCollister (1986) noted, the failure of previous studies to control for DS and MH intelligence differences presented an ambiguity of interpretation: Were group differ-

Hearing Abilities of Down Syndrome Adolescents

535

ences due to the type of mental retardation or differential understanding of test procedures? Dahle and McCollister - the only other investigators, to our knowledge, who included an IQ-matched (as well as CA-matched) MI-I group - found more hearing impairment and otologic disorders in their lo-year-old DS sample. The present project attempted both to replicate their findings in an older age group and to expand our understanding of DS hearing difficulties by using additional measures of hearing acuity. The current report also concentrates on a second question: Does hearing ability in DS subjects correlate with performance on measures of language ability, auditory short-term memory, and speed of auditory word identification? Middle ear infections, for example, are frequently associated with delayed language acquisition and lowered academic achievement in nonmentally retarded children (e.g., Holm & Kunze. 1969; Needleman, 1977; Silva, Kirkland, Simpson, Stewart, & Williams, 1986; Zinkus, Gottlieb, 8z Schapiro, 1978). Although it is not yet clear how much of a hearing loss will cause a cognitivelinguistic impairment (Ruben, 1984), or which linguistic abilities will be affected (Teele, Klein, Rosner, & The Greater Boston Otitis Media Study Group, 1984). the relationship between hearing and linguistic abilities has been clearly established (Paul & Quigley, 1987). The following four studies suggest that such a relationship also holds for the DS population: DS adolescents with abnormal tympanograms scored lower in intelligence than those with normal tympanograms (Libb, Dahle. Smith, McCollister, & McLain. 1985); DS adolescents who received early surgical intervention for otitis media had higher language scores than those with otitis who had not been surgically treated (Whiteman, Simpson, h Compton, 1986); and DS children (Davies & Penniceard. 1980) and adults (Nolan, McCartney, McArthur. & Rowson, 1980) with higher pure tone hearing losses had lower receptive vocabularies. The present study attempted to determine whether such hearing-language relationships, if present in DS subjects, are syndrome-specific; that is, would similar patterns be found in an age- and intelligence-matched MH sample? METHOD Subjects

Fifty-nine trainable mentally handicapped subjects were recruited in a three-county area surrounding Charleston, South Carolina, for participation in the first year of a 3-year longitudinal study. An informal attempt was made, through initial letter and telephone communication with parents, to recruit participants with understandable speech and knowledge of the numbers l-9. However, four individuals (3 DS and 1 MH) with diffkult-tounderstand speech (subjectively determined by the experimenters) and two

536

M. M. Marcel1 and S. Cohen

individuals (both MH) unable to recognize, by sight, all nine digits were included in the final samples. These individuals were kept in the study because of their ability to perform most of the tasks and to speak clearly enough to be understood by the experimenters. Seven other participants were excluded from the study because of no speech (2 MH), an inability to understand (or failure to cooperate during) 40% or more of the tasks (3 DS and 1 MH), or later discovery of educable (rather than trainable) school placement (1 MI-I). The final samples consisted of 26 DS (14 male and 12 female) and 26 MI-I (18 male and 8 female) adolescents and young adults. There were 21 white and 5 black DS participants and 15 white and 11 black MH participants. The samples were matched on Stanford-Binet IQ [DS mean = 39.7 (SD = 7.3); MH mean = 40.9 (SD = 6.2)] and chronological age, [DS mean = 18 years, 10.1 months (SD = 3 years, 4.3 months); MH mean = 18 years, 8.5 months (SD = 3 years, 4.5 months)] by the frequency distribution control method (Christensen, 1991). Subjects were recruited from the following locations: 11 public schools (N = 44). 1 residential institution (N = 2), and 2 community programs for mentally handicapped adult citizens (N = 6 public school graduates). School records, discussions with parents, and medical histories (when available) were used in an attempt to determine etiologies of mental retardation in the MH group. The “causes” (or relevant associated medical conditions) were as follows: unknown (12). seizure disorders (5). early concussion (1 with and 1 without seizures), microcephaly (1 with early malnutrition and 1 with early high fever), cerebral palsy (1 with and 1 without seizures), early oxygen deprivation (1). early meningitis, seizures, and blood clot (1). and Williams syndrome (1). Apparatus, Tasks, and Dependent Variables Audiological tests. Audiological testing was accomplished at the Charles-

ton Speech and Hearing Center. Stimuli were presented in a double-walled Industrial Acoustics Controlled Acoustical Environment. Audiometric and impedance measurements were made possible by a Maico Model MA32 Audiometer and a Madsen Model 2073 Electroacoustic Impedance Bridge. The following audiological assessments were made: 1. Pure Tone Audiometry. Pure tone air conduction measurements of auditory sensitivity were made in the frequency range of 250-8000 Hz. Forty-five of the subjects responded by hand raising to the presence of the tone, and seven (3 DS and 4 MH) were conditioned to respond with standard play audiometry techniques (e.g., dropping a colored token in a box whenever a tone was heard). Decibel hearing

Hearing Abilities of Down Syndrome Adolescents

537

level (dbHL) thresholds were obtained for each ear at each of the six frequencies. Audiogram data were analyzed by both the mean dbHL scores for left and right ears (Coren, 1989) and by the dbHL scores for the better ear only (e.g., Buchanan, 1990). The better ear was determined by both the audiological classification for that ear and the mean dbHL score (averaged over all frequencies) for that ear. One overall measure of pure tone hearing level was calculated for use in correlational analyses; this binaural composite measure was created by averaging hearing thresholds across all six frequencies and across the left and right ears (Coren, 1989). Bone conduction measurements were made in the 250-4000 Hz range and were used in conjunction with air conduction measurements to produce, for each ear, an auditory classification of type of hearing loss as conductive, sensorineural, mixed, or none. l 2. Speech Audiometry. The mean speech reception threshold (SRT) of left and right ears provided an index of the lowest intensity level (in db) at which common, two-syllable spondaic words (e.g., “cowboy”, “hot dog”) were just intelligible. Live voice was used to present the words over headphones at 5 db increments; the subject responded by repeating the word. After SRT was established, the audiometer was reset at 40 dbSL (i.e., 40 db above SRT) and a list of monosyllabic words was spoken in a quiet background for a measure of speech discrimination. Subjects responded either by repeating the word (N = 47) or by pointing to a picture of the word on a plate with five distractor pictures (Ross & Lerman, 1971) (N = 3 DS and 2 MH participants). The mean speech discrimination score of left and right ears reflected the percentage of words correctly recognized out of 50 (25 per ear). 3. Impedance Measurement. Impedance techniques provided a relatively objective assessment of the integrity of middle ear functioning. Compliance tympanograms were plotted for each ear at different air pressure loads ranging from +200 to -200 rnmH+~Oor below. Each tympanogram was categorized as representing a middle ear that was either functioning normally or showing reduced mobility, no mobility, retraction, or hyperflaccidity. Acoustic (stapedial) reflex thresholds were elicited in each ear by contralateral acoustic stimulation at 500, 1,000, kMeria for classifying types of hearing loss were as follows: (1) Norm41- air and bone conduction scores at 20 dbHL or better with no more than 5 db difference between the two; (2) Conductive- bone conduction scores at 20 dbHL or betterand air condoctioo scxn-espoaxx than 20 dbHL. with at least a 15 db difference between tbe two; (3) Sensorineur41- air and bone conduction scores poorertban 20 dbHL and no more than a 5 db difference between tbe two; and (4) Mixed - bone conduction scores poorer than 20 dbHL, air conduction scores poorer than 20 dbHL, and at least a 15 db differencebetween the two.

538

hf. h-l.Marcel1 and S. Cohen

2,000, and 4,000 Hz. A reflex was considered absent if it could not be elicited within the 85-115 db range. A summary acoustic reflex score represented the number of times out of eight that the reflex was elicited in the two ears across the four frequencies. Cognirive tests. All cognitive tasks were presented in a suite of soundattenuated observation rooms at the College of Charleston. Assessments of short-term memory and speed of word identification occurred in a room containing an IBM PS/2 Model 30 computer, an Animated Voice Corporation Professional System Loudspeaker, and a Realistic 33-992B microphone. Digitized stimuli for the computer-based tasks were created and edited with Ani-Vox Authoring Tools software produced by the Animated Voice Corporation (P.O. Box 819, San Marcos, CA). Language testing occurred in a second room containing a Panasonic RQ-2103 tape recorder and a Realistic 33-2001 microphone. The following paragraphs briefly describe the cognitive assessments; additional details concerning the tasks are available elsewhere (Marcell, Croen, & Sewell, 1990; Marcell, Sewell, & Croen, 1990; Marcell. Croen, & Sewell, 1991): 1. Auditory Digit Span. In this traditional measure of auditory short-term memory, increasingly long digitized random number sequences (two examples of each length) were spoken freefield over the computer at the rate of 1 item every 1.5 s. Testing began with 2-digit sequences and was discontinued when the subject incorrectly repeated two consecutive sequences of the same length. The subject was awarded 1 point for each item correctly reported in its proper location in the sequence and 1 point for each pair of items recalled in their correct relationship in the sequence (Marcel1 & Weeks, 1988). 2. Receptive Vocabulary. One measure of language comprehension was the Picture Vocabulary Subtest of the Test of Language Development2 Primary (TOLD-2P; Newcomer & Hammill, 1988). The subject pointed to one of four drawings that best represented the word spoken by the examiner. Testing began with item one and continued until five consecutive items were missed. The subject was awarded 1 point for each correctly identified item out of 35 possible. 3. Grammatical Understanding. A second measure of language comprehension, the Miller-Yoder Language Comprehension Test (Miller & Yoder, 1984), required the subject to point to one of four drawings that best represented the sentence spoken by the examiner (e.g., “Mother is kissed by father.“). Each item was a 4- or 5-word sentence that assessed understanding of 1 of 10 basic grammatical forms (e.g., pro-

Hearing Abilities of Down Syndrome Adolescents

539

nouns, possession, verb inflection, passive voice). The subject was awarded 1 point for each pair of sentences correctly identified out of the 42 sentence pairs presented. 4. Expressive Vocabulary. One measure of language expression required the subject to defme the first eight items (common words such as “cow” or “rest”) on the TOLD-2P Gral Vocabulary Subtest. The subject’s taperecorded responses earned 1 point per item for either a precise deftition or two descriptive characteristics (Newcomer & Hammill, 1988). The subject’s responses were also scored for oral response time, the time lapse between the beginning of the examiner’s pronunciation of the vocabulary word and the beginning of the subject’s response. 5. Sentence Imitation. A second measure of language expression, the first eight items of the TOLD-2P Sentence Imitation Subtest, required the subject to repeat a sentence spoken by the examiner (e.g., “Yesterday my aunt forgot her lunch.“). The subject’s tape-recorded responses earned 1 point per item for a correct imitation of the entire sentence; misarticulations were ignored (Newcomer & Hammill, 1988). The subject’s responses were also scored for oral response time, the duration between the beginning of the examiner’s last spoken word and the beginning of the subject’s response. 6. Backward Masking. In this measure of auditory word identification speed, a concrete noun spoken from the computer was followed at varying brief intervals by a 500 ms burst of white noise. The purpose of the masking noise was to interfere with the subject’s attempt to identify and repeat the word. The interval between word and mask was systematically varied (40, 80, 160, or 320 ms) in order to determine the amount of time needed to identify words. The dependent variable was the number of words identified (out of seven) in each masking interval condition and in a control condition with no mask. 7. Gating. A second measure of auditory word identification involved presentation of progressively larger amounts of a spoken word for identification (Grosjean, 1980); the subject attempted to guess what word the computer was saying (the words were “bread,” “plate,” “hat,” “red,” and “knob”). Testing always began with the briefest gate (the fust 32% of the word) and proceeded to four increasingly longer gates (each representing an additional 17% of the word).2 The dependent hUthough subjects were given a third word identification task - speed of judging whether two spoken items were the same or different- tbis task is not reportedbecause of the large numberof participants (13) who were unable to make reliable same-different judgments during a pretesthiniug procedure.Because the 13 excluded subjects had a significantly lower mean IQ thau the remaiuing subjects, any correlationsbetween audiological measures and same-different wormauce data would be attenuatedby restrictionof range.

M. M. Marcel1 and S. Cohen

variable was the mean earliest gate at which the subject first offered a meaningful word phonetically compatible to the sound spoken by the computer. General Testing Procedure

Each subject spent one day at the College in an individualized program that included the above-mentioned comprehensive audiological assessment, standardized language subtests, and computer-based memory and wordidentification experiments. (Only one subject received the audiological and cognitive assessments on different days.) The day generally began at 9 a.m., ended at 230 p.m., and included a variety of nonresearch activities (e.g., computer games, picnic lunch, crafts project) scheduled at the discretion of the experimenters to provide breaks from the testing routine. In addition to the two previously described computer- and language-testing rooms at the College, a third brightly decorated room was used as a rest, play, and crafts area. Each room had an adjacent observation area equipped with one-way mirror and microphones; parents were invited to observe at any time during the day-long session. The audiological assessment was the last test administered; the other tests were administered in a random order. Forty-eight of the subjects visited the College in pairs; the other four subjects, due to scheduling diffrculties, visited individually. Three well-practiced experimenters shared cognitive testing duties and three clinical audiologists experienced in working with mentally handicapped individuals shared audiological testing duties. All tests were individually administered, and all visits were made during 1989 in the months of June, July, or August, thus reducing the likelihood of audiological scores being influenced by upper respiratory and middle ear infections (Keiser et al., 1981). RESULTS Analyses of Audiological Measures Pure tone audiometry.

Pure tone air conduction measurements for the mean dbHL thresholds of the combined ears (see Fig. 1) were submitted to a 2 (group) x 6 (frequency) ANOVA. The analysis revealed a significant main effect of group (F (1 SO) = 7.51, p = .008), in which the DS group showed a higher overall hearing threshold (20.8 dbHL) (i.e., poorer hearing sensitivity) than the MH group (12.7 dbHL). The analysis also revealed significant effects of frequency (F (5,250) = 22.92, p < .OOOl) and group x frequency interaction (F (5,250) = 3.60, p = .004). Newman-

Hearing Abilities of Down Syndrome Adolescents

541

Keuls post hoc comparisons (at alpha = .OS>on the interaction effect indicated that the DS group had higher hearing thresholds than the MH group at each frequency except 2,000 Hz. Furthermore, in the DS group, the dbHL threshold was higher at the 8.000 Hz frequency than at every other frequency, and the dbHL threshold at the 4,000 Hz frequency was higher than every frequency below it. [In contrast, the hearing sensitivities of MH subjects in the current study did not differ at 2,000.4,000. and 8,000 Hz]. When Widen, Folsom, Thompson, & Wilson (1987; Wilson et al., 1983) classified DS subjects by hearing losses of 10 db or greater, they found that 13 of their 15 subjects (87%) had 8,000 Hz hearing losses. A similar classification of DS subjects in the current study revealed 23 of 26 (89%) with 8,000 Hz hearing losses. (In contrast, 17 of 26 MH subjects (65%) showed such a loss.) The current data not only support the conclusions of Buchanan (1990), Davies (1985). Keiser et al. (1981), and Widen et al. (1987) that high frequency hearing loss is prevalent in DS adolescents and adults, but also demonstrate that high frequency losses are more severe in DS individuals than in their IQ- and CA-matched non-DS mentally handicapped peers. Air conduction thresholds for the subjects’ better ear (see Fig. 2) were also submitted to a 2 (group) x 6 (frequency) ANOVA. As before, the analysis revealed significant effects of group (F (1.50) = 4.89, p = .032) frequency (F (5,250) = 17.37, p < .OOOl) and group x frequency interaction (F (5,250) = 3.25, p = .007). The results of the better ear analysis

FIGURE 1. Pore

tonehearinglevelsfor compositemeasure (both ears).

542

M. M. Marcel1 and 8. Cohen

on-=0

O-O

zm

Nan-OS u-Ha

Pprd -P Down!&ndmnu (OSV= Ofuu~

soo1omzmo4oooaoao

,

FIGURE 2. Pure tone hearing levels for better car.

closely paralleled those of the binaural analysis, with the exception that the DS-MH hearing difference at 1,000 Hz was not significant when data from the better ear were used. Auditory classification of a subject’s hearing ability for each ear using pure tone and bone conduction scores revealed that there were fewer normal DS than MH ears (52% vs 79%) and more DS ears showing hearing losses of the conductive (19% vs 2%) or mixed (17% vs 4%) type, X2(3) = 14.99, p = .0023. The percentages of DS (12%) and MH (15%) ears showing a sensorineural hearing loss did not differ. Classification of the audiograms of mentally handicapped individuals in the present study extended Dahle & McCollister’s (1986) findings with DS and MH children by showing that DS adolescents and young adults also show a higher percentage of abnormal ears, conductive hearing loss, and mixed hearing loss than their IQ- and CA-matched mentally retarded peers. Speech az.diometry. The DS group had a higher mean SRT (i.e., poorer

reception of speech) than the MI-I group (15.0 db vs 8.4 db), t(50) = 2.68,~ = .Ol , supporting Keiser et al’s (1981) finding of poor speech reception thresholds in DS adults. Analysis of speech discrimination scores indicated that the two groups did not differ in accuracy (IX mean = 92.0% and MH mean = 94.1%). t(50) = 1.109. This finding is not unexpected given previous findings of good word recognition ability in hearing-impaired children

Hearing Abilities of Down SyndromeAdolescents

543

with moderate and severe losses (e.g., Erber, 1974) and the fact that words in the present study were presented at sufficient intensities (40 db above each subject’s threshold). Impedance measurement. Relative to the MH group, the DS group had fewer normal tympanograms (50% vs 100%) and more middle ear problems indicative of no mobility (21% vs 0%) or retraction (21% vs 0%) of the tympanic membrane, X2(4) = 34.67, p < .OOOOl.The overall incidence rates were similar to those reported by Davies and Penniceard (1980), who found that 55% of their DS children and 100% of their severely retarded children (not matched on IQ) had normal tympanograms, and Keiser et al. (1981). who found that 49% of their DS adults had normal tympanograms. DS middle ear problems tended to be bilateral: 12 subjects had abnormal tympanograms in both ears and 2 subjects had an abnormal tympanogram in one ear. Acoustic reflexes were less likely to be elicited in the DS (mean = 3.8) than the MH (mean = 6.9) group, t(47) = 3.59, p < .OOl, supporting Schwartz and Schwartz’s (1978) finding of absent acoustic reflexes in a large number of DS children’s ears. In general, the impedance findings reinforced those of researchers who used only DS subjects and reported a relatively low incidence of normal ears and high prevalence of conductive (e.g., Balkany, 1980; Cunningham & McArthur, 1981; Downs, 1983) and mixed (e.g., Fulton & Lloyd, 1968; Keiser et al., 1981) hearing losses. Analyses of Cognitive Measures The results of analyses for each of the cognitive tasks are summarized in Table 1. As expected on the basis of previous research (e.g., Marcell, Harvey, & Cothran, 1988; Rondal, Lambert, & Sohier, 1981), DS subjects showed more limited auditory recall of both digit sequences and meaningful sentences than MH subjects; furthermore, they were slower to initiate their sentence imitations. The DS group also identified fewer words when they were rapidly (40 ms) masked by noise, and needed more acoustic information (on the average, 65% of the word vs 53% for MH subjects) before guessing the identity of a gated word. The backward masking and gating findings suggested that as a group, DS individuals were slower to identify spoken words than other mentally handicapped individuals - a finding that extends previous reports of slow auditoryverbal processing by mentally retarded individuals in general (e.g., Merrill & Mar, 1987). Although the two groups did not differ in their ability to comprehend or define single words or in their general ability to comprehend sentences, a more detailed analysis of sentence comprehension revealed that the DS group showed poorer understanding of grammatically difficult sentences.

M. M. Marcell and S. Cohen

544

TABLE 1 Major Results From Cogdtive Tests Cognitive Test

DS Meau (SD)

MH Mean (SD)

statistic @F)

Auditory digit span* Receptive vocabuhy

16.7 (7.8) 12.2 (4.5)

23.2 (15.3) 13.5 (6.0)

t (SO)= 1.92. f (50) = 0.89

Number of sezttence & ideutifti hopc&onofsentencepairs identified by difficulty levelb &Y Modgate DitEcult Expressive vocabulary l%nnh of words &fined Respouse the (s)C Senteke imitation Numbe~ofse~~tencesrepeated

15.9 (6.6)

19.0 (9.1)

.71 (.33) 44 (.17) 20 (.16)

.71 (34)

3.5 (2.1) 3.6 (1.3)

4.2 (2.3) 3.6 (1.6)

r(50)=1.12 f (49) = 0.16

1.1 (1.4) 2.0 (0.9)

2.9 (2.3) 1.5 (0.6)

t (50) = 3.38*** t (49) = 2.05’

5.3 (1.1) 4.8 (1.3) 5.4 (1.3) 5.1 i1.4, 5.5 (1.5) 2.9 (l.ti)

5.7 (1.2) 5.7 (1.1) 5.7 (1.2) 5.7 (1.2) 5.7 (1.1) 2.3 (0.7)

r(50)=1.27 t (49) = 2.65.’ t (49) = 68 t (49j = 1.67 t (49) = .70 t (49j = 2.82+*

B%TS$ unmaskedumditioo 4Omsmaskingc4mdition 8olll.3lMSkillRconditiOn

160 ms mask&g condition 320 ms masking condition

Gatinge

-

t(50)=1.40

F (3.48) = 3.80**

Note. Additional details about the cognitive analyses are available in Marcell, Cmen, & Sewell (1990). Marcell, Sewell, & Croen (1990). and Marcell, Croen. & Sewell(l991). Unless othenvise noted, N = 26 for the Down syndrome (DS) group and N = 26 for the non-DS mentally handicaDped (MH) ~lroug. *A one&led t-test was brfonued to assess the prediction of better MH thau DS performance. bThe three difficulty levels were entered into a MANOVA with group as the indeuendent variable. The effect of group~membership (W&s’ Lambda F) was due s&ielyio the DS g&p identifying a lower propordon of grammatically difticult sentences than the MH group. cN = 25 for tbe DS group (one subject’s responses were not tape recorded due to experimenter error). dA prioricomparisons were performed to assess the prediction that the DS group would perform less accurately on only the fastest masking conditions. N = 25 for the MH group (one subject was unable to follow task instructions during the mast&a conditions). oue subject was not giventhe task due to experimenter error).

.

oo1

.

Correlational Analyses of Audiological and Cognitive Measures

In this analysis, the three numerical audiological measures possessing the greatest variability - overall pure tone hearing loss, SRT, and acoustic reflex score - were correlated with summary scores from each of the previously described language, memory, and word identification measures. Due to the known relationships between age and cognitive performance (e.g., Case, Kurland, & Goldberg, 1982) and intelligence and cognitive performance (e.g., Hunt, Lunneborg, & Lewis, 1975), partial correlations were

Hearing Abilities of Down Syndrome Adolescents

545

used to remove statistically the influences of CA and IQ. As can be seen in Table 2. the analysis yielded significant correlations in only one task, the backward masking experiment: DS subjects, but not MH subjects, who had poorer hearing tended to identify fewer words across the combined masking conditions. To better understand this finding, correlations were performed between the audiological measures and each of the four masking conditions. These data, presented in Table 3, reveal that DS (but not MH) subjects with greater hearing losses, higher (worse) SRTs, and fewer instances of elicited acoustic reflexes identified fewer words when the items were followed quickly (40 or 80 ms) by masking noises. Neither DS nor MH subjects showed this pattern when there was a longer pause (160 or 320 ms) between the word and the mask (or when there was no mask). An attempt was made to replicate findings of three of the four previously described studies of DS individuals that reported relationships between tympanogram type and IQ (Libb et al., 1985) and pure tone hearing loss and receptive vocabulary (Davies & Penniceard, 1980; Nolan et al., 1980). [The fourth study (Whiteman et al., 1986) could not be evaluated due to lack of access to detailed records of our subjects’ early medical histories.] A partial correlation between number of abnormal ears (as revealed by tympanograms) and Stanford-Binet IQ (with the effect of age partialled out, as performed by Libb et al., 1985) was in the predicted direction for TABLE 2 Relationships Between Audiological and Cognitive Audiological Measures* Cognitive Measures

DS Group Pure Tone

Auditory digit span -.003 Receptive vocabulary -.085 Grammaticalunderstandingb -.084 Expressive vocabulary -.134 Response time -.014 Sentence imitation -2.40 Response time -.I32 Backward ma&it@ -.392 Gating -.122

MH Group

SRT

Reflex

PureTone

SRT

RefkX

.058 -.233 -.167 -.291 .055 -.173 -241 -.427’ .064

-.088 .277 .025 -.009 .087 .229 .142 .449* .054

.Oll -.136 -.150 -.052 -.182 ,304 .084 -.216 -.023

-.130 -.146 -.132 -.168 -.I62 -248 .062 -.313 .059

-.189 .229 .165 .012 .122

.328

-.356 .235 .094

Note. Partialcorrelationswith effects of IQ and CA removed were used. DS = Down syndrome, MH = non-DS mentally bandiiped. Vme Tone = Binaural composite pure tone hearing level; SRT = Speech reception threshold; Reflex = Acoustic reflex score. bThedependentvariablewas the total numberof sentence pairs corfrcdy identifd. “he dependent variable was the total number of words correctly identified across tbe four masking conditions. l p < .05.

hi. M. Marcel1 and S. Cohen

546

TABLE 3 RelatIomkipr Between Audiologkal Mesrurea and Number of Words Identlfkd h Backward Masking lbsk

Maskinginterval (ms) Audidogicd Measures

DSGroup(N=26) 40

80

160

Puretoaeheariuglevel

-.485* -.469*

[email protected]&old

-.441* -.52A** -.149

Acuustic reflex sun+

.455*

.3n

MHGlWp(N=26)

-0%

324I

40

80

,146

,199

-094

,065

-361

-.182

-.317

-.087

-3Og

-.328

247

.129

.179

.360

.292

160

320

.lOO

Note. Partial correlations with the effects of IQ and CA removed were used. DS = Down syndromeand MH = nonDS mentally handicapped. *N = 25 for DS group and N = 24 for MH group due to inability of audidogist to obtain or maintainseakononearbodlears. *p < .05. l p < .Ol.

DS subjects, r(23) = -.373, although not significant (critical value at alpha = .05 is .396). The relationship could not be explored in MH subjects because of the absence of variability in the tympanometry data. To evaluate the Davies and Penniceard (1980) and Nolan et al. (1980) studies, a correlation was performed between overall puretone hearing loss and receptive vocabulary score. Although this correlation was not significant, other correlates of receptive vocabulary did achieve or approach significance in the predicted direction for DS (but not MI-l) subjects: number of abnormal ears (tympanogram), ~(24) = -.454, and acoustic reflex score, 423) = .373. However, these correlations (like others) disappeared when the influences of IQ and CA were removed. Because Davies and Penniceard and Nolan et al. did not control for the effects of well-known correlates of receptive vocabulary, such as intelligence (cf., Dunn & Dunn, 1981), the possibility that their findings were an artifact of uncontrolled concurrent variation due to age and intelligence should be raised. DISCUSSION Our audiological findings replicated and extended those of the only other study, to our knowledge, that includes a mentally handicapped control group equated on overall intelligence (Dahle & McCollister, 1986). Like Dahle and McCollister, we found that DS individuals had worse hearing than their mentally handicapped peers on virtually all measures. They showed lower sensitivity across most pure tone frequencies, poorer speech reception thresholds, and a stronger tendency toward conductive/middle ear difficulties. New audiological fmdings from our study include the follow-

Hearing Abilities ofDown Syndrome Adolescents

547

ing: (a) audiological description of older (19 vs 10 years) IQ- and CAmatched DS and MH samples, (b) confirmation of a stronger tendency toward high frequency loss in DS than matched MH individuals, and (c) descriptions of greater DS than MH difficulties with both the reception threshold for familiar speech and elicitation of the acoustic reflex. Interestingly, the results of correlational analyses of audiological and cognitive measures indicated that hearing difliculty did not predict performance in complex cognitive tasks, such as sentence imitation and sentence comprehension. It is possible that factors such as memory limitations and poor articulation may be more relevant to performance on such global tasks than hearing difficulties. The major result of the correlational analyses was that DS individuals with hearing difftculties need more time to identify spoken words than DS individuals without hearing difficulties. Because the hearing ability - speed of word processing relationship did not hold for MH subjects, it may be tentatively suggested that hearing difficulty interacts with etiology; that is, hearing loss appears to be more detrimental to speed of word processing in mentally retarded individuals with Down syndrome. “Speed of word processing” is used here in a global manner; that is, it refers generally to the time needed to identify successfully a spoken word. The particular information-processing locus of this difficulty - for example, phonological encoding, lexical access and retrieval, response execution (Maisto & Baumeister, 1984; Merrill & Mar, 1987; Salasoo & Pisoni, 1985) - cannot be determined from the present data. It is also unclear whether slow processing of stimuli is confined to the auditory modality; it is possible that visual stimuli might be processed more slowly by DS individuals with hearing problems, suggesting a more global and central source of difficulty. [Evidence from traditional reaction time tasks, however, indicates slower DS responses to auditory than visual stimuli (see Anwar, 198 1, for a review)]. Similarly, it is unclear whether a DS auditory processing difficulty would also occur with nonverbal auditory stimuli. If so, their difficulty would be with acoustic stimuli in general and not tied specifically to their language problems. Investigations by Tallal and her colleagues (e.g., Tallal, 1980; Tallal & Piercy, 1974; Tallal, Stark Kallman, & Mellits, 1980) have documented that language-impaired nonretarded children show difficulties in discriminating rapidly changing nonverbal as well as verbal acoustic stimuli. Such a finding in the DS population would imply that slower auditory stimuli, or stimuli in which the acoustic features necessary for identification are made more discriminable, might facilitate comprehension. Finally, it should be remembered that the inference of a speed of processing problem in DS individuals is based hugely on the results of only one task (backward masking). Thus, the interpretation of slow DS word processing is only a working hypothesis; it is possible that another interpretation of

548

M. M. Marcel1 and S. Cohen

poor performance on the backward masking task might prove more appro priate. Given the newness of this task, it would be prudent to suggest that the results be considered tentative until they have been replicated. In addition to a potential DS problem in speed of auditory word processing, there is strong evidence of DS difficulties in hearing acuity, middle ear functioning, and reception threshold for speech. The possibility should be considered, then, that in addition to processing acoustic signals more slowly, DS individuals with poor hearing also tend to perceive them less clearly. Raz, Willerman, & Yama (1987) suggested that individual differences in intelligence among nonretarded adults might reflect differing levels of precision in resolution of perceptual signals as well as differing speeds of information processing. Raz et al. found that IQ correlated significantly with simple frequency discrimination in a task in which college students indicated which was the higher of two 20 ms tones. The researchers suggested that higher-IQ subjects may have faster feature extraction and/or higher fidelity sensory representations than lower-IQ subjects and concluded that the perceptual resolution of the auditory system is “intimately related” to its rate of information processing: “Under time constraints the system with lower need for external signal redundancy will respond faster than a noisy system...” (p. 208). Applied to the backward masking task of the present study, it is possible that DS subjects with poor hearing and, presumably, poor resolution are sometimes unable to “figure out” the stimulus before the rapidly appearing auditory mask. When the mask comes later (160 or 320 ms), the poor resolution of the initial stimulus is unchanged, but the time constraints are not as severe and the subject has sufficient time to decipher the signal. It is possible, then, that poor DS performance in a rapid spoken word identification task may represent an interaction between slower processing speed and poorer quality of perceptual resolution. REFERENCES Andrews, R J., & Andrews, J. G. (1977). A study of the spontaneousoral language of Down’s syndromechildren.ExceptionalChild,2486-94. Anwar,F. (1981). Motor function in Down’s syndrome. In N.R. Ellis (FZd.),International review ofresearch in mentalrefardation(Vol. 10, pp. 107-138). New York:Academic Press. Ashman, A. (1982). Cognitive processes and perceivedlanguage pexformanceof retardedpersons. Journal of MentalDeficiency Research,26,131-141. Balkany, T. J. (1980). Otologic aspects of Down’s syndrome. Seminars in Speech, Iimguage and Hearing, 1.39-48. Bridges, A., & Smith, J. V. E. (1984). Syntactic comprehension in Down’s syndrome children. BritishJournal of Psychology,75187-196. Bmuks, D. N., Wooley, H., & KanjilaI, G. C. (1972). Hearing loss and middle ear disorders in patients with Down’s Syndrome (Mongolism). Journal of Menral Deficiency Research, 16,21-29.

Hearing Abilities of Down Syndrome Adolescents

549

Buchanan, L. H. (1990). Early onset of presbyacusis in Down syndrome. Scundinavinn Audiology,19.103-110. Burr, D. B., & Rob, A. (1978). Pat&us of psycholinguistic development in the severely mentally retarded’ A hypothesis. Social Biology, 25.15-22. Case, R., Kurland, M., & Goldberg, J. (1982). OpuaIional effiiency and rhe growth of short-term memory span. Journal of Eqmimental ChildPsychology,33.386&)4. C&mum, R S., Schwartz, S. E, & Kay-Raiuing Bird, E. (1989, November). Are children wi#h Down syndkme language delayed? Paperpresentedat the meeting of the Americaa SpeechLanguage-Hearing Association, St. Louis. cwstensen. L. B. (1991). Erpcrheti tnerhodology. Bostoo: AByu and Bacon. Coren, S. (1989). Summarizing piue-tooe. hearing duesholds: ‘Ibe equipollence of components of tbe aUogram. Bulletinof the PsychonomicSociety,27.42-44. Cornwell A. C. (1974). Devel~nt of laaguage.. abstractkm, and numeficat amcept fcmnation iu Dowa’s syudraue children. Amcria~ Journal of MentalD@cimcy, 79,179-190. Cunningham, C., & McArthur, K. (1981). Hearing loss aad treatment in young Down’s syudmme cbildrea. Child: Care, Healthand Development,7,357-374. Da&. A. J., & McCoUister, E l? (1986). Hearing aad otologic disorders in children with Down syndrome. American Journal of MentalDeficiency,90.636642. Davies, B. (1985). Hearing problems. In D. Lane & B. Stratford (E!ds.), Current appnmches to Down’s syruhme (pp. 85-102). London: FIaeger. Davies, B., & Parni R. M. (1980). Auditory fuaction and receptive vocabulary iu Down’s Syndrome childnz In I. G. ‘l?+ylorL A. Markides (Ed-s.),Disordersof auditoryfwrcrionIll @p. 51-58). Londoz Academic Press. DownsM.P.(1983).Is~bearinghelpforDown’sSyndrrme?InG.T.M~&S.EGerber @Is.). Themuh@yhandicoppulheuring impoiredchiki(pp. 301-316). New Yak t&ax! & Saanan. Duna. L. M.. & Duan, L. M. (1981). Peaho@ PictureW.xbulary Test-Revised. Circle Pines, MN: Amuican Guidance Service. Ekber, N. I? (1974). Pure-tone tlu&olds and word-recognition abilities of hearing impaired children. Journal of Speech and HearingResearch, 17.194-202. F&on. R. T., & Lloyd, L. L. (1968). Hearing impairment in a population of children with Down’s syndrome. AmericanJournal of MentalDeficiency,73,298-302. Glovsky, L. (1966). Audiological assessment of a Mongoloid population. I&e IFain@ School Bulletin,63,27-36. van Gorp. E.. & Baker, R. (1984). The incidence of hearing impairment in a sample of Down’s syndmme scboolchikkat. InternationalJournal of Rehabilitation Research,7,198-200. Grosjean, F. (1980). Spoken word recognition processes and the Bating pamdigm. Perceptionand Psychophysics,28,267-283. Hartley,X. Y. (1982). Receptive language processing of Down’s syndrome children. Journal of MentalDeficiencyResearch,26.263-269. Hartley, X. Y. (1985). Receptive language processing and ear advantage of Down’s syndrome children. Journul of MentalDeficiencyResearch,29,197-205. Helm, V. A., & Kunze. L. H. (1%9). Efkct of chronic otitis media on language aad speech development. Pediatrics,43,833-839. Hunt, E., Lmmeborg, C., & Lewis, J. (1975). What dces it mean to be bigb verbal? Cognifive Psychology,7.194-227. Keiser, H., Montague, J., Wold, D., Maune, S., & FUison. D. (1981). Hearing loss of Down syndrome adults. AmericanJournal of MentalDeficiency,85,467-472. Libb, J. W., Jhhle, A., Smith, K., McCoIJister, E P.. & McJ..ain, C. (1985). Heatiug di.u&er and cognitive function of individuals with Down syndrome. American Journal of Mental DeJiciency,90,353-356. Maisto, A. A., & Baumeister, A. A. (1984). Diion of component processes iu rapid information processing tasks: Comparison of retarded and nonretarded people. In I? H. Books, R.

550

M. M. Marcell and S. Cohen

Sperber, & C. McCauley (Eds.), Learning Q?tdcognifion in the mentally retarded (pp. 165-188). H&dale, NJ: Erlbaum. Marcell, M. M., Croen, P S., & Sewell, D. H. (1990, August). Jirngaage comprehension in Down syndrome and other trainable mentally handicapped individuals. Paper presented at the Cwferwrce cn Humau Development, Richmond, VA. Marcell. M. M., Ctoen. P. S.. & Sewell, D. H. (1991, April). Spokenwordidentificmion by Down syndrome & other mentQllyh~mRcQpped adolescents.Wper presentedat the biemtial meeting of the Society far Research in Chikl Developmen& Seattle. Marcell. M. M., Harvey, C. F., & Cothran, L. P. (1987, April). Profileof the languageabilitiesof Down’s syndrome adolescents. Paper presented at the bienniai meeting of the Society for Research in Child Development, Baltimore. Marcell, M. M., Harvey, C. F.. & Cothran, L. P. (1988). An attempt to improve auditory shortterm memory in Down’s syndrome individuals through reducing distractions. Research in .Det&JnnetUol DisQbilities, 9,405-417. Marcell,M. M., Sewell, D. H.. & Croen, P. S. (March, 1990). Expressivelo.nguQgein Down syndrome end other trainablemenuallyhandicQppedindividuQls. Paperpresented at the meeting of the Conference on Human Development, Richmond, VA. Marcell, M. M.. & Weeks, S. L. (1988). Short-term memory difflcultles and Down’s syndrome. Jourrmlof hientnlDeficiencyResearch,32.153-162. McDude.,H. L.. & Adler, S. (1980). Down syndrome and short-term memay impairment: A storage or retrieval deficit? American Jourmzlof MentalDeficiency,04.561-567. McrriU,E. C.. & Mar, H. H. (1987). Differences between mentally mtarded and mmmta&d persons’ e5ciency of auditay sentence prcce&ng. AmericQn Journalof Men&dDe$ciency,91,406-t14. Miller, J. F. (1987). Language and communication characteristics of children with Down syndrome. In S. M. Pueschel, C. Tmgey, J. E. Rynders, A. C. Crccker, & D. M. Crutdxz @ds.), New perspectiveson Downsyndrome(pp. 233-262). Baltimore: Bmokes. Miier, J. F., & Yoder, D. E. (1984). Miller-YnIerLQnglrageComprehensionTest (clinical ed). Baltimore: University Park Press. Needleman, H. (1977). Effects of hearing loss from early recurrent odds media on speech and language development. In B. F. Jaffe (Ed.), Hearing loss in children: A comprehensivetext (pp. 640-649). Baltimore: University Park Press. Newcomer, l? L., & Hammill, D. D. (1988). Testof Language Development-2Primary. Austin, TX Pro-Ed. Nolan, M., McCartney, E., McArthur, K.. 8c Rowson, V. J. (1980). A study of the hearing and receptive vocabulary of the tminees of an adult training centre. Journal of Men&l Deficiency Reseamh,24.271-286. Paul, P. V., & Quigley, S. P (1987). Some effects of early hearing impairment on English language development In F. N. Martin (Ed.). Hearing disordersin children(pp. 49-80). Austin TX Pro-Ed. Rax, N., Willerman, L., & Yama, M. (1987). On sense and senses: Intelligence and auditory informatlon processing. Persormlityend IndividuQl Difirmces, 8,2Ol-210. Rondal, J. A., Lambert, J. L., & Sohier, C. (1981). Elicited verbal and nonverbal imitation in Down’s syndrome and other mentally retarded children: A replication and extension of Beny. Lengauge and Speech, 24,245-254. Ross, M., & Lerman, J. (1971). Wordintelligibility by picture identificafion.Pittsburgh: Stauwix House Inc.. Ruben, R. J. (1984). An inquiry into the minimal amount of auditory deprivation which results in a cognitive effect in man. Acta Otolaryngology (Supp1.I414.157-164. Salusoo,A., & Pisoni, D. B. (1985). Interztion of knowledge sources in spoken word identitlcation. Jourrmlof Memoryand Language,24.210-231. Schwartx, D. M., & Schwartz. R. H. (1978). Acoustic impedance and otoscopic findings in young children with Down’s Syndrome. Archivesof Otolaryngology,104,652-656.

Hearing Abilities of Down Syndrome Adolescents

551

Silva, P. A., Kirkland, C., Simpson, A., Stewart, I. A., & Williams, S. hi. (1986). Some developmeutal end behavioral problems assochd with bilateral otitis media with effusion Journal of L&am& Disabilities,15.417-421. Smut, F.. O’Gtady, M., & Das, J. P. (1982). Cognitive processing by subgroups of moderately mentally chlkku. AmericanJoumal of MentalDqiciency, sa.465-472. ‘hllal, P. (1980). hcqtual requisites for language. In R L. Schiefelbuscb (Fxl.),hguugc intervention series hlume N: Nonqeech hguage and commuaicationanalysisand intervention (pp. 449-467). Baltimme: Uuiversity ParkPress. ThllaLI!. & Kucy, M. (1974). DevelqYneattalapbask Bate of auditoq processing aud selective lmpahentof cousommtpexepth. Neump~chologia, 12.83-93. ‘hllal, P., Stark, R E.. Kallmau, C., & Mellits, D. (1960). Develw dyspbask Relatiou Neuropsycholoia, l&273-274. behveulacoustkpKk%slugdeficitspndvefbalpluax&g. Teek D. W., Klelu, J. 0.. Rosner,B. A., & the GreatezBoston Otitis Media Study Group.(1984). Otitismedia~itheffusion~gthefvstthreeyearsoflifeanddevelopmentofspeecband language. Pediamics,74.282287. Varnhagen. C. K., Das. J. P., & Varuhagen. S. (1987). Auditory and visual memory span: Cognitive pocessiug by TMR iudividuals with Down syudromeor other etiologies. American Journal of MentalDejiciem, 91,398-405. Whiteman, B. C., Simpson. G. B.. & Chmptou, W. C. (1986). Reladouship of otitis media and language imphment in adolescents with Down syndrome.Mental Retamktion,24.353-356. Widen, J. E., Folsom, R. C.. Thompson, G.. & Wilson, W. R. (1987). Auditory brainstem responses in young adults with Down syndrome. American Journal of Mental Deficiency, 91.472479. Wilsm. W. R., Folsom, R C.. & Widen, J. E. (1983). Hearing impairmentin Down’s syndrome children. Iu G. T. Menck & S. E. Geher @Is.), The multiplyhandicappedhearing iqoaired child (pp. 259-299). New Yc& Cime & Soaaon. Znkus, I? W., Gottlieb, M. I., & Schapiro. M. (1978). Developmental and psychoeducational sequelae of chronic otitismedia. American Journal of Disabilities in Childhood, 132, 1100-1104.