15
J Am Acad Audiol 10 : 557-571 (1999) Recognition of One-, Two-, and Three-Pair Dichotic Digits under Free and Directed Recall Anne Strouse* Richard H. Wilson' Abstract A one-, two-, and three-pair dichotic digit test was administered to 180 subjects (20-79 years) with varying degrees of hearing sensitivity . The test was carried out under free- and directed-recall response conditions . The results indicated that recognition performance decreased as a function of increasing age . Statistical analysis using analysis of covariance indicated that differences in performance between age groups were not owing to differences in hearing sensitivity. Thus, with the effects of hearing sensitivity removed, age alone accounted for a significant portion of the variability in the data among age groups . Normative data for the free-recall condition are reported for the two lists of randomly interleaved one-, two-, and three-pair digits included on the Tonal and Speech Materials forAuditory Perceptual Assessment, Disc 2.0 (1998) compact disc . A comparison of performance in free- and directed-recall con- ditions indicated three patterns of results . First, 58 percent of the subjects had normal recognition performance for one-, two-, and three-pair digits under both recall conditions . Second, 39 percent of the subjects had a deficit in the free-recall condition, with normal per- formance in the directed-recall condition, which was interpreted as representing primarily a cognitive problem . Third, the remaining 3 percent of the subjects showed a deficit in both free- and directed-recall conditions, which was attributed primarily to an auditory problem . Key Words : Aging, auditory perception, compact disc, dichotic listening tests, speech per- ception, uncertainty Abbreviations : ANCOVA = analysis of covariance, ANOVA = analysis of variance, CD = compact disc, rau = rationalized arcsine unit W ord recognition performance is poorer among older listeners as compared to younger listeners, particularly when listening in the presence of noise or com- peting sounds (Dubno et al, 1984) . Unfortu- nately, speech perception problems, especially in the elderly population, cannot be predicted solely from the degree of peripheral hearing loss . Rather, speech perception problems reflect both peripheral and central components (Jerger et al, 1989, 1990b). In addition to the well-documented age-related decline in pure-tone sensitivity, the *Veteran Affairs Medical Center, Mountain Home, Tennessee ; 'Departments of Surgery and Communicative Disorders, East Tennessee State University, Johnson City, Tennessee Reprint requests Anne Strouse, VA Medical Center, Audiology (126), Mountain Home, TN 37684 literature suggests that with advancing age, changes occur within the central auditory sys- tem. Since aging is accompanied by a decline in cognitive as well as auditory function, it is impor- tant to examine the extent to which the auditory problems of the elderly can be explained by con- comitant decline in such cognitive functions as memory and attention (CHABA Working Group, 1988). Dichotic listening tasks, in which different auditory stimuli are presented to the two ears simultaneously, can place sufficient auditory and cognitive demands on the auditory system to assess auditory perceptual abilities . Dichotic listening tasks can be made progressively more difficult by lengthening the stimuli (e .g., a hier- archy of one to three digit pairs) (Strouse and Wilson, 1999) and/or by varying the response mode (e .g., free recall vs directed recall) (Jerger et al, 1990c) . In a free-recall task, stimuli are pre- 557

Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

J Am Acad Audiol 10 : 557-571 (1999)

Recognition of One-, Two-, and Three-Pair Dichotic Digits under Free and Directed Recall Anne Strouse* Richard H. Wilson'

Abstract

A one-, two-, and three-pair dichotic digit test was administered to 180 subjects (20-79 years) with varying degrees of hearing sensitivity . The test was carried out under free- and directed-recall response conditions . The results indicated that recognition performance decreased as a function of increasing age . Statistical analysis using analysis of covariance indicated that differences in performance between age groups were not owing to differences in hearing sensitivity. Thus, with the effects of hearing sensitivity removed, age alone accounted for a significant portion of the variability in the data among age groups . Normative data for the free-recall condition are reported for the two lists of randomly interleaved one-, two-, and three-pair digits included on the Tonal and Speech Materials forAuditory Perceptual Assessment, Disc 2.0 (1998) compact disc . A comparison of performance in free- and directed-recall con-ditions indicated three patterns of results . First, 58 percent of the subjects had normal recognition performance for one-, two-, and three-pair digits under both recall conditions . Second, 39 percent of the subjects had a deficit in the free-recall condition, with normal per-formance in the directed-recall condition, which was interpreted as representing primarily a cognitive problem . Third, the remaining 3 percent of the subjects showed a deficit in both free- and directed-recall conditions, which was attributed primarily to an auditory problem .

Key Words : Aging, auditory perception, compact disc, dichotic listening tests, speech per-ception, uncertainty

Abbreviations : ANCOVA = analysis of covariance, ANOVA = analysis of variance, CD = compact disc, rau = rationalized arcsine unit

W ord recognition performance is poorer among older listeners as compared to younger listeners, particularly

when listening in the presence of noise or com-peting sounds (Dubno et al, 1984) . Unfortu-nately, speech perception problems, especially in the elderly population, cannot be predicted solely from the degree of peripheral hearing loss . Rather, speech perception problems reflect both peripheral and central components (Jerger et al, 1989, 1990b). In addition to the well-documented age-related decline in pure-tone sensitivity, the

*Veteran Affairs Medical Center, Mountain Home, Tennessee; 'Departments of Surgery and Communicative Disorders, East Tennessee State University, Johnson City, Tennessee

Reprint requests Anne Strouse, VA Medical Center, Audiology (126), Mountain Home, TN 37684

literature suggests that with advancing age, changes occur within the central auditory sys-tem. Since aging is accompanied by a decline in cognitive as well as auditory function, it is impor-tant to examine the extent to which the auditory problems of the elderly can be explained by con-comitant decline in such cognitive functions as memory and attention (CHABA Working Group, 1988).

Dichotic listening tasks, in which different auditory stimuli are presented to the two ears simultaneously, can place sufficient auditory and cognitive demands on the auditory system to assess auditory perceptual abilities. Dichotic listening tasks can be made progressively more difficult by lengthening the stimuli (e .g., a hier-archy of one to three digit pairs) (Strouse and Wilson, 1999) and/or by varying the response mode (e .g., free recall vs directed recall) (Jerger et al, 1990c) . In a free-recall task, stimuli are pre-

557

Page 2: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Journal of the American Academy of Audiology/Volume 10, Number 10, November/December 1999

sented dichotically, and the listener must mon-itor and report items heard in both ears . In a directed-recall task, stimuli are presented dichot-ically, but the listener is only required to report items heard in one ear. The ear to be reported is cued prior to stimulus presentation . Jerger et al (1990c) proposed that the comparison of per-formance in free- versus directed-recall condi-tions can differentiate between the two principal factors underlying performance on dichotic tasks: (1) an auditory/structural factor deriving from the anatomy of the auditory pathways and (2) a task-related/cognitive factor deriving from the demands placed on cognitive processing such as memory and speed of mental processing . Specifically, Jerger et al (1990) hypothesized that when recognition performance is poor in the free-recall condition, but improves substantially in the directed-recall condition, the problem is primarily in the cognitive domain since memory and attention abilities are insufficient for suc-cessful performance when both ears must be monitored simultaneously (Bayles and Kasz-niak, 1987 ; Jerger et al, 1990b). In contrast, when performance is below normal in both the free- and directed-recall conditions, then the problem is interpreted to be primarily in the auditory domain . Performance does not improve by reducing cognitive demands . Jerger and col-leagues examined this paradigm previously using the Dichotic Sentence Identification Test (Chmiel and Jerger, 1996; Fifer et al, 1983; Jerger et al, 1990a, 1991, 1994).

The current study examined this inter-pretation using a dichotic digit paradigm . Dichotic digit materials are ideal for use with elderly listeners because digits (1) are rela-tively immune to the effects of cochlear hear-ing loss (Musiek, 1983 ; Speaks et al, 1985, Musiek et al, 1991) and (2) have demonstrated high intertest reliability for both young and elderly adult listeners (Strouse and Hall, 1995; Humes et al, 1996 ; Strouse & Wilson, 1999). Moreover, the digit stimuli are generally famil-iar to most listeners.

The new Department of Veterans Affairs compact disc (CD), Tonal and Speech Materials for Auditory Perceptual Assessment, Disc 2.0 (1998), contains two lists of randomly inter-leaved one-, two-, and three-pair digits in a free-recall paradigm . Using this free-recall paradigm, data from our laboratory on young and elderly listeners with hearing impairment confirm that as the complexity of the task increases from easy (one pair) to difficult (three pair), recogni-tion performance decreases systematically (Wil-

son and Jaffe, 1996 ; Strouse and Wilson, 1999). The primary purpose of the current study was to describe normative recognition data for 20- to 79-year-old listeners under the free-recall con-dition on the interleaved one-, two-, and three-pair dichotic digit paradigm. A secondary purpose was to examine the hypothesis proposed by Jerger and colleagues by evaluating recogni-tion performance of the randomly interleaved one-, two-, and three-digit paradigm in both free- and directed-recall conditions .

METHOD

Materials

The preparation of the stimulus materials is detailed in an earlier paper (Strouse and Wil-son, 1999) . Briefly, the digital waveform files (1, 2, 3, 4, 5, 6, 8, 9, and 10, spoken by a male) used on the Tonal and Speech Materials for Audi-tory Perceptual Assessment, Disc 1.0 (Department of Veterans Affairs, 1992) were edited so that the onset of the stimulus coincided with the start of the data file . A silent interval was added to the end of the file to equalize the file lengths to the longest digit (561 msec) . A 500-msec silent inter-val, which served as the interdigit interval for the multipair digit sets, was added to the end of each one-pair dichotic digit file . The two- and three-pair files were made by combining as required two or three of the compiled one-pair dichotic digit files, with an interstimulus inter-val of 4, 5, and 6 sec following the one-, two-, and three-pair stimuli, respectively. The following two rules were used in the compilation of each mul-tipair digit list : (1) no digit was repeated in a stimulus set and (2) each of the 72 one-pair dichotic sets was used once (randomly) in each presentation position . In this manner, a list of 108 items for the free-recall condition was com-piled that contained 36 stimulus sets of the one-, two-, and three-pair digits . Additionally, a 10-item practice list was compiled . Both lists were recorded on CD (Pinnacle, Model RDC-1000) . The 108-item list subsequently was divided into two lists of 54 stimulus sets (18 each of the one-, two-, and three-pair digit sets) and recorded as Tracks 7 and 8 on Version 2.0 of the VA-CD Tonal and Speech Materials for Auditory Perceptual Assessment (1998) .

For the directed-recall condition, the words left and right (recorded by the speaker who recorded the digits) were added as required 500 msec before each digit file . Thus, the word left (or right) was presented binaurally, 500 msec

558

Page 3: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Dichotic Digit Recognition under Free and Directed Recall/Strouse and Wilson

before the simultaneous presentation of differ-ent digit stimulus sets to the right and left ears . Again, a list of 108 stimulus sets was compiled that consisted of 36 of each of the one-, two-, and three-pair digit sets . Half of the stimulus sets (18 sets of each of the one-, two-, and three-pair digits) were recorded on CD as cued to the left ear and the other half of the digit sets cued to the right ear. Different 108-item stimulus sets were used for the free- and directed-recall conditions .

Subjects

A total of 180 right-handed subjects (129 male, 51 female) were evaluated, ranging in age from 20 to 79 years . There were 30 subjects in each of the following six age groups : 20 to 29 years (mean age = 23.5 years), 30 to 39 years (mean age = 34.6 years), 40 to 49 years (mean age = 45.6 years), 50 to 59 years (mean age = 55.2 years), 60 to 69 years (mean age = 64.6 years), and 70 to 79 years (mean age = 72.9 years) . The subjects were recruited from within the audiol-ogy clinic and from the community and were paid for their participation . The mean audiometric thresholds for each group are shown in Table 1 . For each subject, the severity of hearing loss did not exceed the mild-to-moderate range and asym-metry of the pure-tone thresholds never exceeded 15 dB . An analysis of variance (ANOVA) indi-cated a significant difference in pure-tone thresholds for the right (F [5, 1741 = 29.4,

p < .0001) and the left (F [5, 1741 = 31.5, p < .0001) ears . Post hoc analysis using the Tukey test revealed that for both ears, thresh-old data were not significantly different (p < .05) for adjacent age groups (i .e ., 20- vs 30-year group, 30- vs 40-year group, 40- vs 50-year group, etc .) ; however, statistically significant differences (p > .05) were found for comparisons between the remaining age groups . Word recog-nition performance on a recorded version of the Northwestern University Auditory Test No . 6 (NU-6) (Department of Veterans Affairs, 1991) was >_80 percent in each ear and within 10 per-cent between ears for each subject . There was no significant difference in word recognition performance between the six age groups (F [5, 1741 = 1.8, p > .05) .

Procedures

The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through an audiometer (Grason-Stadler, Model 10) to TDH-50 earphones encased in P/N 510CO17-1 cushions . All stimuli were presented at 70 dB HL (ANSI, 1996). For the free-recall con-ditions, the subjects were instructed to recall in any order the digit pairs presented to both ears . For the directed-recall conditions, the subjects were instructed to recall only the digits pre-sented to the ear precued relative to the stimu-lus presentation (Bryden et al, 1983) . For example, the subject first heard right and then

Table l Mean Thresholds (dB HL)* and Standard Deviations for the 30 Subjects in Each of the Age Groupings

Frequency (Hz)

Age Group 250 500 1000 2000 4000 8000

Left Ear 20-29 9 .3 (5 .5) 8.0 (4 .1) 4 .8 (4 .0) 4.5 (4 .6) 5.1 (5 .0) 8.2 (6 .8) 30-39 10 .8 (7 .0) 8 .7 (5 .4) 7 .2 (6 .1) 7 .5 (5 .8) 13 .5 (12 .0) 12.6 (11 .8) 40-49 11 .5 (6 .2) 8.0 (5 .1) 8 .5 (5 .6) 10 .1 (7 .4) 29 .0 (23 .4) 29 .5 (21 .9) 50-59 14 .3 (8 .3) 11 .8 (6 .5) 12 .8 (9 .3) 19 .7 (13 .9) 35 .8 (22 .5) 43 .5 (24 .5) 60-69 14 .8 (9 .6) 15 .2 (9 .0) 17 .7 (10 .7) 24 .3 (14 .5) 50 .5 (23 .6) 60.6 (23 .1) 70-79 18 .5 (6 .6) 15 .8 (7 .4) 16 .0 (9 .2) 23 .7 (11 .7) 45 .0 (18.2) 55 .2 (17.8)

Right Ear 20-29 7 .2 (5 .8) 5 .7 (5 .5) 5 .0 (6 .0) 4 .2 (4 .4) 4 .2 (4 .5) 9 .3 (6 .5) 30-39 10 .0 (8 .1) 9.2 (7 .0) 5 .5 (6 .1) 4.5 (6 .9) 10 .8 (15.4) 14 .3 (12.4) 40-49 10 .7 (8 .0) 8 .5 (5 .6) 7 .2 (6 .5) 8 .2 (8 .0) 26 .3 (24 .1) 31 .7 (20 .1) 50-59 12 .8 (6 .9) 12 .0 (6 .9) 12 .2 (8 .6) 17 .7 (15 .2) 35 .2 (21 .8) 43 .3 (24 .2) 60-69 13 .8 (7 .0) 16 .0 (10 .6) 18 .0 (11 .8) 23 .5 (15 .4) 45 .3 (22.3) 59 .5 (23.8) 70-79 18 .0 (7 .0) 16 .0 (8 .1) 18 .0 (9 .0) 25.8(14 .2) 47 .2 (17 .6) 57.2 (19 .5)

-ANSI (1996) .

559

Page 4: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Journal of the American Academy of Audiology/Volume 10, Number 10, November/December 1999

digits were presented simultaneously to each ear. The task of the subject was to recall only the dig-its presented to the right ear.

All of the subjects were practiced on the dichotic listening task before data collection . Practice consisted of three to five presentations each of the one-, two-, and three-pair digit stim-uli in free- or directed-recall conditions . During the practice sequence, the subjects responded verbally and were queried periodically con-cerning their comfort in performing the listen-ing/response task . When the subject reported feeling comfortable with the task, which typically was after 10 or so total practice items, data col-lection began.

Both free- and directed-recall conditions were evaluated. The order of presentation of the two test conditions was counterbalanced among the subjects who participated in two tri-als during separate test sessions . Thus, during each test session, the subject was administered 108 free-recall items preceded by 10 practice items (test time of -14 minutes) and 108 directed-recall items preceded by 10 practice items (test time of -20 minutes) . All testing was conducted in a double-walled sound booth.

RESULTS

T he dichotic digit data were scored by pre-sentation position (i .e ., each digit in the pair

was scored separately) . Thus, for the three-pair digits, there were six possible numbers to be scored as correct or incorrect. For the analysis of the two- and three-pair conditions, the data from the presentation positions were averaged for each subject. Prior to statistical analyses, all percentage scores were transformed into ratio-nalized arcsine units (raus) (Studebaker, 1985). This had the effect of minimizing the relation-ship between mean score and variance that is characteristic of percentage scores, while pro-viding a scoring unit that is similar to percent-ages . The data expressed in raus were used for statistical analysis ; however, figures and tables in the following discussion of results are expressed in percent correct.

Trial Effects

The mean data from the six age groups by trial are shown in Table 2. Although an ANOVA revealed that the main effect for trials was sig-nificant (F [1, 1741 = 23 .8, p < .0001), inspec-tion of the mean data in Table 2 indicates that the largest percent difference between trials for

Table 2 Means and Standard Deviations for the Right and Left Ears of Trials 1 and 2 in Free- and Directed-Recall Conditions

Trial 1 Trial 2

Left Ear Right Ear Left Ear Right Ear

Free Recall 20-29 years One pair 99 .6 (1 .2) 99 .7 (1 .1) 99 .5 (1 .3) 99 .9 (0 .5) Two pair 96 .9 (4 .7) 97 .6 (3 .7) 98 .1 (3 .2) 99 .0 (1 .3) Three pair 89 .8 (8 .5) 92 .7 (7 .1) 93 .1 (7 .2) 95 .4 (3 .6) 30-39 years One pair 96 .5 (9 .1) 98 .3 (6 .1) 98 .7 (3 .3) 99 .5 (1 .3) Two pair 86.5 (16 .4) 95 .1 (5 .4) 89 .2 (15 .3) 94 .8 (5 .9) Three pair 76.0 (16 .2) 85 .2 (8 .6) 80 .1 (15 .3) 88 .2 (8 .8) 40-49 years One pair 96 .6 (7 .4) 98 .9 (4 .1) 97 .9 (3 .7) 98 .3 (5 .6) Two pair 88.3 (13 .0) 94 .3 (8 .1) 89 .8 (14.6) 95 .4 (4 .8) Three pair 77.3 (16 .3) 83.3 (13 .2) 78 .6 (16.1) 86 .8 (8 .6) 50-59 years One pair 93 .5 (8 .6) 96 .9 (5 .1) 95 .6 (9.9) 98 .1 (4 .1) Two pair 78.4 (16 .2) 90 .7 (8 .3) 82 .8 (17.4) 91 .2 (9 .2) Three pair 68.2 (15 .8) 77.0 (13 .2) 73 .8 (15.5) 79 .9 (13 .2) 60-69 years One pair 93.6 (14 .3) 96 .9 (6 .3) 96 .5 (6.0) 99 .0 (2 .1) Two pair 83.8 (16 .3) 91 .9 (9 .4) 88 .0 (10.7) 94 .0 (7 .0) Three pair 72 .0 (16 .2) 80.2 (13 .8) 76 .6 (12.9) 84 .1 (13 .6) 70-79 years One pair 90.3 (13 .0) 95 .2 (9 .9) 93.7 (9.3) 97 .8 (3 .8) Two pair 74 .3 (16 .7) 90.6 (10 .9) 78 .8 (17.0) 90 .6 (11 .5) Three pair 64.5(18 .4) 78.3(14 .0) 70 .2 (18.4) 80 .6 (11 .7)

Directed Recall 20-29 years One pair 98 .5 (2 .9) 98 .9 (3 .4) 96 .5 (5.2) 98 .9 (2.7) Two pair 96 .1 (4 .9) 97 .8 (2 .7) 96 .9 (4.0) 98 .1 (2 .4) Three pair 95 .3 (5 .7) 98 .4 (2 .8) 96 .3 (3.8) 98.6(l .9) 30-39 years One pair 96 .3 (7 .7) 97 .8 (4 .5) 94 .4 (11 .5) 98 .3 (3 .0) Two pair 94 .5 (8 .3) 95 .5 (8 .0) 91 .7 (10 .5) 95 .7 (7 .2) Three pair 94 .2 (9 .8) 96 .2 (8 .4) 91 .8 (11 .7) 95 .9 (7 .9) 40-49 years One pair 93 .9 (11 .0) 97 .8 (4 .3) 95 .7 (7 .8) 97 .2 (8 .9) Two pair 94 .0 (9 .2) 96 .0 (6 .2) 94 .6 (8 .6) 97 .2 (6 .0) Three pair 92 .6 (11 .1) 95 .4 (7 .2) 94 .4 (7 .0) 97.5 (4 .1) 50-59 years One pair 95.0 (8 .6) 96 .5 (7 .1) 94 .8 (9 .0) 97 .2 (6 .2) Two pair 94 .5 (5 .8) 96 .5 (6 .4) 95 .5 (6 .3) 95.5 (7 .6) Three pair 92 .5 (8 .2) 96 .1 (6 .2) 93 .2 (9 .4) 96 .1 (6 .1) 60-69 years One pair 95.6 (9 .5) 98 .5 (4 .6) 98 .4 (3 .5) 99 .8 (3 .9) Two pair 94 .5 (8 .4) 97 .8 (4 .9) 95 .7 (4 .6) 94 .2 (6 .7) Three pair 94.3 (9 .8) 97 .9 (4 .2) 95 .3 (4 .9) 98 .6 (2 .3) 70-79 years One pair 90 .4 (16 .6) 97.0 (8 .4) 94 .2 (11 .5) 98 .0 (3 .9) Two pair 90.9 (9 .5) 95 .7 (9 .4) 92 .0 (11 .8) 96 .6 (6 .8) Three pair 89 .8 (13 .3) 96.2 (9 .6) 92 .6 (12 .5) 97 .6 (5 .1)

the free-recall condition was for the 70- to 79-year group. For the three-pair digit condition, there was a difference of 5.7 percent between

560

Page 5: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Dichotic Digit Recognition under Free and Directed Recall/Strouse and Wilson

trial 1 and trial 2, with recognition perfor-mance on trial 2 (70.2%) better than recogni-tion performance on trial 1 (64.5%) . As each token in the free-recall condition equaled 2.8 percent (100%/36 items possible for each of the one-, two-, and three-pair digit sets), the 5.7 per-cent difference was equivalent to two tokens . In the directed-recall condition, the largest percent difference between trials was for the 70-to 79-year group, where there was a 3.8 percent improvement on trial 2 (94.2%) compared to trial 1 (90.4%) for the left ear in the one-pair digits condition . With each token in the directed-recall condition equivalent to 5.6 per-cent (100%/18 items possible for the right and left ears in each of the one-, two-, and three-pair digit sets), this trial difference was less than one token . Because of the small between-trial dif-ferences, in addition to the fact that one trial typically would be used for clinical purposes due to the test length, the dichotic digit data from trial 1 only were used for the remainder of the analyses . Accordingly, statistical analysis using the trial 1 data was performed.

For the dependent variable (percent correct recognition), the influence of (1) age group, (2) stimulus condition, (3) ear, (4) stimulus pair, and (5) interactions of these factors was assessed using a mixed model ANOVA with condition, ear, and stimulus pair as within-subject factors and age group as the between-subject factor. The results indicated that the main effects for the three independent variables were signifi-cant (condition : F 11, 1741= 210.9, p < .0001 ; ear : F [1, 1741 = 73 .3, p < .0001; stimulus pair : F [2, 3481= 465.6, p < .0001) . The main effect of age group also was significant (F [5, 1741 = 7.8,

p < .0001) . To examine the effects of hearing loss, the three-frequency pure-tone average (500, 1000, and 2000 Hz) and the four-frequency pure-tone average (500, 1000, 2000, and 4000 Hz) for the right and left ears were used as covari-ates in separate analyses of covariance (ANCOVA) that examined the same variables . The results of the ANCOVA using the three-(F [5, 1721 = 3.6, p < .01) and four-frequency (F [5, 1721 = 3 .2, p < .01) pure-tone average were not different than the ANOVA results, indi-cating that the differences in performance between groups were not owing to differences in hearing sensitivity. Thus, with the effects of hearing sensitivity removed, age alone accounted for a significant portion of the variability in the data between age groups . The main effects and significant interaction effects for these data are described in detail in subsequent sections .

Condition Effects

Figure 1 shows percent correct recognition by ear and presentation position for the one-, two-, and three-pair dichotic digits presented to the six subject groups in the free- (upper panels) and directed-recall (lower panels) conditions . The means and standard deviations for the left and right ears for each condition are listed in Table 3 . The data in the table are shown by presen-tation position for the one-, two-, and three-pair digit conditions and are averaged across pre-sentation position for two- and three-pair digit conditions in the right two columns . The data for all subject groups indicated that the recognition performance in the free-recall condition was poorer than recognition in the directed-recall condition . For example, in the free-recall con-dition, the mean recognition performance for the three-pair dichotic digits presented to the left ear ranged from 89.8 percent (20- to 29-year group) to 64.5 percent (70- to 79-year group) . In contrast, performance on the three-digit pairs presented to the left ear in the directed-recall condition was better, with a smaller intergroup range (95 .3% and 89 .8% for the 20- to 29-year and 70- to 79-year groups, respectively) . For the most part, the data in Figure 1 are orderly and systematic . Generally, as the age of the group increased, there was a decrease in recognition performance. There was, however, one exception . The performances on the two- and three-pair dig-its in the free-recall condition by the 60- to 69-year group were better than the performances by the 50- to 59-year group . We are unable to account for this discrepancy.

Ear Effects

The ANOVA indicated a significant differ-ence between scores on the materials presented to the right and left ears (F [1, 1741 = 73 .3 ; p < .0001) . As previously reported by Kimura (1961), when the dichotic task involved little difficulty as with the one-pair stimuli, there was only a small, if any, right-ear effect . When the listening task became more difficult, as with the three-pair stimuli, the difference between ear preference increased.

Consider first the percent correct recognition data from the free-recall conditions displayed in the upper panels of Figure 2. The data are dis-played separately for one-, two-, and three-pair digits within each condition. The right-ear scores were higher than the left-ear scores for all sub-ject groups . The ear differences for the youngest

561

Page 6: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Journal of the American Academy of Audiology/Volume 10, Number 10, November/December 1999

ONE-PAIR TWO-PAIR THREE-PAIR

100

90

80

70

60

50

100

90

80

70

60

50

0

0 g a 0

0__-0 0-0

~fl 20-29

0-.0 30-39 40-49 0-9

50-59

70-79

FREE RECALL

o o

DIRECTED RECALL

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 Left Ear Right Ear Left Ear Right Ear Left Ear Right Ear

PRESENTATION POSITION

Figure 1 The mean percent correct recognition for the one-, two-, and three-pair digit materials presented to the right and left ears by presentation position for the six subject age groups in free- (upper panels) and directed-recall condi-tions (lower panels).

age group (20-29 years) were small for one-, two-, and three-pair digits (0 .1%, 0.7%, and 2.9%, respectively) and were not statistically different by paired t-test (p > .05) . Because of the near ceiling performance, there was little dif-ference between right and left ears. With the exception of the 20- to 29-year group, however, performance on materials presented to the right ear was significantly (p < .05) better than per-formance on materials presented to the left ear.

Analysis of the data for the free-recall con-dition revealed a significant ear by group inter-action (F [5, 174] = 2.5 ; p < .05) . As can be seen in Figure 2, scores obtained for the right ear were better than scores obtained for the left ear and the right-ear advantage increased as age increased. The linear equations used to fit the mean data for

the left ear (Xs) and right ear (Os) are shown at the bottom of each panel in the figure . In the equations, the coefficient of x defines the slope of the function . The slope of the function describes the change in performance as a function of age and is expressed as percent/decade . All slopes of the linear functions for the free-recall conditions are negative, which indicates a decrease in dichotic digit recognition as a function of age. For one-pair digits in the free-recall condition (upper left panel), the slopes are -0.91 percent/decade and -1.85 percent/decade for the right and left ears, respectively, indicating that the performance on the materials presented to the left ear decreased faster across age than the performance on the materials presented to the right ear. This result was statistically significant by paired t-test

562

Page 7: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Dichotic Digit Recognition under Free and Directed Recall/Strouse and Wilson

Table 3 Means and Standard Deviations for Digit Materials Presented to the Right and Left Ears for Six Subject Groups

Left-Ear Presentation Position

1 2 3

Right-Ear Presentation Position

1 2 3 Left-Ear Average

Right-Ear Average

Free Recall 20-29 years One pair 996 (1 .2) 99 .7 (1 .1) 99 .6 (1 .2) 99 .7 (1 .1) Two pair 96.9 (48) 96 .8 (5 .1) 97 .8 (3 .7) 97.5 (4 .5) 96 .9 (4 .7) 97 .6 (3 .7) Three pair 94 .9 (55) 89 .4 (9 .6) 85 .2 (13.7) 95 .8 (4 .5) 90 .2 (11 .2) 91 .9 (9 .6) 89 .8 (8 .5) 92 .7 (7 .1) 30-39 years One pair 96.5 (1 .9) 98 .3 (6 .1) 96 .5 (9 .1) 98 .3 (6 .1) Two pair 86 .6 (17.4) 86 .5 (16 .4) 95 .8 (4 .8) 94 .4 (6 .7) 86 .5 (16.4) 95 .1 (5 .4) Three pair 82 .3 (167) 74 .0 (17.4) 71 .7 (196) 92.8 (5 .0) 81 1 (11 .8) 81 .7 (13.2) 76 .0 (16.2) 85 .2 (8 .6)

40-49 years One pair 96.6 (7 .4) 98 .9 (4 .1) 96 .6 (7 .4) 98 .9 (4 .1) Two pair 87 .7 (13A) 88.9 (14 .8) 92.9 (11 .3) 95 .7 (6 .0) 88 .3 (13 .0) 94.3 (8 .1) Three pair 83 .9 (14.0) 75 .6 (18.6) 72,4 (22.6) 86 .2 (14.2) 78 .3 (19.7) 85 .5 (11 .6) 77 .3 (16.3) 83 .3 (13.2) 50-59 years One pair 93 .5 (8 .6) 96 .9 (5 .1) 93 .5 (8 .6) 96 .9 (5 .1) Two pair 79 .0 (16 .1) 77 .9 (20.5) 90 .0 (8 .2) 91 3 (10.6) 78 .4 (16.2) 90 .7 (8 .3) Three pair 74 .1 (163) 64.0 (19 .2) 66.5 (18 .6) 82.0 (13.0) 71 .2 (18.1) 77 .9 (14.7) 68 .2 (15.8) 77 .0 (13.2) 60-69 years One pair 93 .6 (14 .3) 96.9 (6 .3) 93 .6 (14 .3) 96.9 (6 .3) Two pair 833 (16 .6) 84.2 (17 , 3) 92 .1 (7 .4) 91 .6 (13 .3) 83.8 (16 .3) 91 .9 (9 .4) Three pair 77 .8 (19 .8) 69 .8 (18.6) 68 .3 (17.9) 86 .9 (11 .7) 74 .6 (17.1) 791 (18 .0) 72 .0 (16.2) 80 .2 (13.8) 70-79 years One pair 90 .3 (13 .0) 95.2 (9 .9) 90.3 (13 .0) 95.2 (9 .9) Two pair 74 .5 (20.5) 74 .2 (19.0) 89 .2 (13.0) 92 .0 (10.5) 74 .3 (16.7) 90 .6 (10.9) Three pair 72 .0 (21 .6) 60 .7 (20 .1) 60 .8 (22 .8) 84 .3 (13 .4) 70.9 (18 .2) 79.9 (14 .5) 64.5 (18 .4) 78 .4 (14 .0)

Directed Recall 20-29 years One pair 98 .5 (29) 98 .9 (3 .4) 98 .5 (2 .9) 98 .9 (3 .4) Two pair 94 .6 (76) 97 .6 (4 .0) 97 .2 (4 .1) 98 .3 (3 .0) 96 .1 (4 .9) 97 .8 (2 .7) Three pair 93 .5 (8 .8) 95 .6 (6 .8) 96 .7 (5 .0) 97 .2 (4 .1) 98 .7 (3 .2) 99 .3 (2 .4) 95 .3 (5 .7) 98 .4 (2 .8) 30-39 years One pair 963 (7 .7) 97 .8 (4 .5) 96 .3 (7 .7) 97 .8 (4 .5) Two pair 93 .1 (8 .0) 95 .9 (10 .4) 95 .2 (7 .3) 95 .7 (9 .4) 94 .5 (8 .3) 95 .5 (8,0) Three pair 93 .7 (10.5) 95 .4 (7 .9) 93 .5 (13.3) 96 .3 (5 .5) 96 .1 (12.2) 96 .1 (11 .2) 94 .2 (9 .8) 96 .2 (8 .4) 40-49 years One pair 93 .9 (10.9) 97 .8 (4 .3) 93 .9 (11 .0) 97 .8 (4 .3) Two pair 93 .3 (9 .3) 94 .6 (9 .8) 94 .8 (7 .7) 97 .2 (5 .0) 94 .0 (9 .2) 96 .0 (6 .2) Three pair 91 1 (12.3) 92 .2 (13.1) 94 .4 (11 .3) 93 .7 (10 .1) 95 .6 (7 .5) 97 .0 (4 .8) 92 .6 (11 .1) 95 .4 (7 .2) 50-59 years One pair 95 .0 (86) 96.5 (7 .1) 95 .0 (8 .6) 96 .5 (7 .1) Two pair 91 .2 (8 .9) 94 .6 (9 .8) 95 .6 (8 .9) 97 .4 (5 .0) 94 .5 (5 .8) 96 .5 (6 .4) Three pair 90.9 (10 .5) 93.7 (8 .7) 93 .0 (10 .3) 94.8 (9 .2) 96 .5 (6 .4) 97 .0 (5 .2) 92 .5 (8 .2) 96 .1 (6 .2) 60-69 years One pair 95 .6 (9 .5) 98.5 (4 .6) 95 .6 (9 .5) 98 .5 (4 .6) Two pair 94 .8 (7 .7) 94 .3 (10 .3) 96.5 (6 .4) 99 .1 (4 .2) 94 .5 (8 .4) 97 .8 (4 .9) Three pair 94 .6 (102) 93.5 (13.5) 94 .8 (8 .0) 97 .4 (5 .4) 98 .2 (3 .7) 98 .2 (4 .5) 94 .3 (9 .8) 97 .9 (4 .2) 70-79 years One pair 90 .4 (16.6) 97 .0 (8 .4) 90.4 (16 .6) 97 .0 (8 .4) Two pair 89 .3 (15.0) 92 .6 (12.0) 94 .1 (11 .0) 97 .4 (8 .4) 90 .9 (9 .5) 95 .7 (9 .4) Three pair 887 (12 .8) 89.6 (18 .4) 91 .1 (12,2) 94 .8 (12 .5) 97.0 (8 .7) 96 .9 (8 .7) 89 .8 (13 .3) 96 .2 (9 .6)

The data are shown by presentation position for the one-, two-, and three-pair digits in the first two columns and are averaged across presentation position for two- and three-pair digits in the last two columns of the table .

(p < .05) . For two- and three-pair digits in the free-recall condition, again the slope of the func-tion for the left ear (-3 .94 and -4.48

percent/decade, respectively) is about twice as steep as the slope of the function for the right ear (-1.35 and -2.91 percent/decade, respec-

563

Page 8: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Journal of the American Academy of Audiology/Volume 10, Number 10, November/December 1999

100

80

60

40

20

100

80 60

40

20

0

ONE-PAIR TWO-PAIR THREE-PAIR

1 1 T

1 lX FREE

1 1 RECALL

YHE =100.98 - 0.91 X ypE = 98.05 - 1 .35X y1,E - 93 .42 - 2.91x

yLE =101.83 - 1.85x yLE = 98 .87 - 3 .94x yLE = 90 .84 - 4.48x

1 1 1 1 1 1

t 1 } 1

y 1

DIRECTED RECALL

y.E= 97.67 - 0.08x yqE= 95 .60 + 0.14x yqE = 95.96 + 0.08x

yLE = 97.61 - 0.91 x yLE = 95 .00 - 0.43x yLE = 94.31 - 0 .50x

V°' 1b Oq 0°' 9o°~ 1°~ ;L°' S°' V h°' ;cP 'l°' do 40 44o ,~o IV 40 yo 60

AGE GROUPS

~oti ~o5°'~oop~oh~4o6 1101~

Figure 2 The mean percent correct recognition and standard deviations for the right (Os) and left (Xs) ears as a func-tion of age group in free- (upper panels) and directed-recall (lower panels) conditions . Data are displayed separately for one-, two-, and three-pair digits within each condition. The linear equations used to fit the data for the left and right ears are shown at the bottom of each panel.

tively). The results from the right and left ears for the two- and three-pair digits also were sig-nificantly different (p < .05) .

The mean data for the directed-recall con-dition are shown in the lower three panels of Fig-ure 2. As with the free-recall conditions, recognition performance in the directed-recall conditions was better on the materials presented to the right ear than on the materials presented to the left ear for one-, two-, and three-pair dig-its . Only the scores in the three-pair condition, however, reached statistical significance by paired t-test (p < .05) . The largest difference between right- and left-ear scores in the three-pair, directed-recall condition was 6.4 percent for the 70- to 79-year group (see Table 3) . The slopes of the linear functions for the right ear are essen-tially flat (-0.08, 0 .14, and 0.08%/decade for one-, two-, and three-pair digits, respectively), indicating consistent performance across age. In comparison to the slope data for the right ear, the slope data for the left ear show a slightly greater decrease in recognition performance

with age (-0.91, -0.43, and -0.50%/decade for the one-, two-, and three-pair digits, respectively) .

Stimulus Pair Effects

The results from the ANOVA on the data from the six subject groups indicate that in the directed-recall condition, recognition perfor-mance among the one-, two-, and three-pair digits was not significantly different (F [2, 348] = 1 .9 ; p > .05) . For the free-recall condi-tion, however, as the complexity of the listen-ing task increased from easy (one pair) to difficult (three pair), there was a correspond-ing significant decrease in recognition perfor-mance (F [2, 348] = 705.0 ; p < .0001) . In addition, there was a direct relation between difficulty of the task (one, two, or three pair) and age . As age increased, there was an increase in the difference in performance on the easy and difficult tasks . There was also a sig-nificant interaction effect between stimulus pair and ear (F [2, 348] = 16.0 ; p < .0001) . As

564

Page 9: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Dichotic Digit Recognition under Free and Directed RecalUStrouse and Wilson

the stimulus difficulty increased, recognition performance on the materials presented to both the right and left ears decreased with the larger decrease in performance associated with the materials presented to the left ear.

Figure 3 shows mean percent correct recog-nition performance for one-, two-, and three-pair digits for each age group in the free-recall condition . The linear equations used to fit the data for the left (Xs) and right ears (0s) are shown at the bottom of each panel . The differ-ences between the functions for the right and left ears represent the right-ear advantage or, con-versely, the left-ear disadvantage . All slopes of the linear functions (%/digit pair) are negative with the slopes for the left-ear data steeper than the slopes for the right-ear data . For the right ear, the slope of the function increased from -3.5 percent/pair for the 20- to 29-year group to -8.4 percent/pair for the 70- to 79-year group. Likewise, for the left ear, the slope of the func-tion increased from -4 .9 percent/pair for the 20- to 29-year group to -12.9 percent/pair for the 70- to 79-year group . Thus, the slope data indi-cated that (1) the performance on the materials

100

90

80

70

60

50

100

90

80

70

60

50

presented to the left ear decreased faster as a function of task difficulty than performance on materials presented to the right ear and (2) that subject groups above the 20- to 29-year group exhibited an increase in the difference in per-formance on the easy and difficult tasks .

The percent correct performance for the three-pair condition in the 50- to 59-, 60- to 69-, and 70- to 79-year groups illustrated in Fig-ure 3 was near chance, especially in the left ear. If the responses for the three-pair digits were based strictly on chance performance, then the likelihood of correctly guessing six of the nine possible digits would be 67 percent . In Figure 3, the percent correct recognition for the three-pair condition was 68.2 percent, 72 percent, and 64.5 percent in the left ear for the 50- to 59-, 60- to 69-, and 70- to 79-year groups, respectively. Based on chance, the results suggest that the subjects in the 50- to 59-, and 60- to 69-year groups performed just above chance and subjects in the 70- to 79-year group performed slightly below chance . Because the response task of the subjects did not require a response to all stim-uli, however, often the number of responses was

20-29 YEARS 30-39 YEARS 40-49 YEARS

0

YAE =103 .67 - 3.60x yaE = 105.33 - 5.65x ypE = 107.77 - 7.80x

yLE = 105.23 - 4.90x yLE = 108.33 - 9.75x yLE

= 106.70 - 9.65x

50-59 YEARS 60-69 YEARS 70-79 YEARS

o

x

°

1Z

o

yqE = 108 .10 - 9.95x yqE = 106.37 - 8.35x yqE =104.87 - 8 .40x

yLE = 107.73 - 13 .55x yLE = 104.73 - 10.80x yLE = 102.17 . 12 .90x

1 2 3 1 2 3 NUMBER OF DIGIT PAIRS

1 2 3

Figure 3 The mean percent correct recognition for the right (Os) and left (Xs) ears for one-, two-, and three-pair dig-its for each age group in the free-recall condition. The linear equations used to fit the data for the left and right ears are shown at the bottom of each panel.

565

Page 10: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Journal of the American Academy of Audiology/Volume 10, Number 10, November/December 1999

100

90

80

70

~O1P 9g A°' 4~9 6°'

�~O QO y0 60

AGE GROUPS

9 101

Figure 4 The percent of subject responses per presentation as a function of age for one- (squares), two-(triangles), and three-pair (circles) dichotic digits .

fewer than the number of stimulus presenta-tions. It became instructive then to examine the relation between the number of responses and the number of stimulus presentations, which, in effect, adjusted the level of chance performance. Figure 4 shows the percent of responses per presentation as a function of age. The data represent how many digits the subjects within each age group reported in each condition, regardless of correctness . Recall that if the sub-ject always reported six digits in the three-pair condition, chance would be 67 percent. The data in Figure 4, however, indicate that in the three-pair condition for the 70- to 79-year group, the subjects actually responded to only 69.3 per-cent of the digits presented, which would reduce chance performance to 46 percent ([6 x 0.693]/9) . A similar pattern was observed for the one- and two-pair digits with 94 percent and 80 percent responses, respectively. When viewed from this perspective, all correct responses were above chance performance. The data in Figure 4 also show that as the task increased in difficulty from one- to three-pair digits, the percentage of responses per presentation decreased for all age groups . In the 20- to 29-year group, subjects responded to 100 percent of the one-pair digits and 96 percent of three-pair digits . For com-parison, in the 70- to 79-year group, subjects responded to 94 percent of one-pair digits and only 69 .3 percent of three-pair digits . These results suggest that as the response task increased in difficulty, the subjects became more conservative in their responses and that this effect became more pronounced as a function of age.

Free Recall vs Directed Recall

To examine the hypothesis proposed by Jerger et al (1990a and 1996), the recognition performance of each subject for the free-recall dichotic digit condition was scored according to the criteria in Figure 5 and Table 4. The data in the figure show mean percent correct recognition for one-, two-, and three-pair data for the left and right ears . The solid lines above and below the mean data represent the 99.7 percent confi-dence interval that was calculated as the mean percent correct recognition score ±3 standard errors . The 99.7 percent confidence intervals for the free-recall condition are listed in Table 4 for each age group. The 95 percent confidence intervals (representing ±2 standard errors) also are provided . For the current study, the lower cutoff of the 99.7 percent confidence interval was used to categorize "normal" and "below-normal" performance for the one-, two-, and three-pair digits within each age group .

On the basis of these normative findings, overall performance for each subject was cat-egorized as normal or below normal according to criteria delineated by Jerger et al (1990c). According to this criteria, if subject perfor-mance was normal for both free-recall and directed-recall conditions, then overall perfor-mance was categorized as normal . A total of 104 of the 180 subjects (58%) were in this normal category. If recognition performance was below normal in the free-recall condition and normal in the directed-recall condition, then overall performance was again categorized as normal since the dichotic deficit in the free-recall con-dition was eliminated by lessening the cogni-tive demands of the task in the directed-recall condition. According to the hypothesis, the reduced performance for the free-recall condi-tion could be attributed to cognitive rather than auditory-specific factors . There were 70 subjects (39%) in this second normal group. Finally, if recognition performance was below normal in both free- and directed-recall condi-tions, then overall performance was catego-rized as below normal, interpreted as an auditory-specific deficit. The reduced memory and attention requirements in the directed-recall task did not result in increased recogni-tion performance. Before categorizing a subject as having an auditory-specific deficit, the sub-ject scored below the normal cutoff for a least one ear in both the free- and directed-recall con-ditions. There were six subjects (3%) in this cat-egory. Of the six subjects, one was from the

566

Page 11: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Dichotic Digit Recognition under Free and Directed Recall/Strouse and Wilson

100

90

80 70

60

100

90

80

70

60

100

90

80 70

60

LEFT EAR RIGHT EAR

x x x x

x

° ° o 0 o

ONE-PAIR

0 0

x x x ° ° o

x x

TWO-PAIR

x

x x

xxx - THREE-PAIR

25 35 45 55 65 75 25 35 45 55 65 75

AGE IN YEARS

Figure 5 The mean percent correct recognition for one-, two-, and three-pair data in the free-recall condition for the left and right ears as a function of age. The solid lines above and below the mean data represent the 99.7 percent con-fidence interval calculated as the mean percent correct recognition score +3 standard errors .

30- to 39-year group, one was from the 40- to 49-year group, one was from the 60- to 69-year group, and three were from the 70- to 79-year group . The deficit, in both free- and directed-recall conditions, was on the right ear for two subjects and on the left ear for four subjects . The below-normal score was present in all three digit pair conditions for one subject and was below normal in the two- and three-pair con-ditions for the remaining five subjects . For these five subjects, recognition performance for one-pair digits was normal .

DISCUSSION

Normative Data

The normative data presented in this study are based on a hierarchy of randomly inter-leaved one-, two-, three-, and four-pair dichotic digits in which the subjects did not know a priori the number of digit pairs to be presented

on a given trial . The results indicated good intertest reliability for the test materials . Although statistically different, the two trials differed by only two tokens (5.7%), a differ-ence that was deemed insignificant for clinical purposes . The overall consistency of recognition performance across trials is in agreement with previous studies of dichotic listening using dig-its and consonant-vowels as stimuli (Ryan and McNeil, 1974; Speaks et al, 1982 ; Musiek et al, 1991 ; Strouse and Hall, 1995 ; Wilson and Leigh, 1996) .

Normative data are reported for the free-recall condition of the dichotic digits test as a function of age group. Statistical analysis using ANCOVA indicated that with the effects of hear-ing sensitivity removed, age alone accounted for a significant portion of the variability in the data between age groups . Given that the dichotic digit stimuli were presented at 70 dB HL, it is possible that the decreased recognition perfor-mance in older listeners was a result of the pre-sentation of digit materials at a lower sensation

567

Page 12: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Journal of the American Academy of Audiology/Volume 10, Number 10, November/December 1999

Table 4 99.7% Confidence Intervals (± 3 Standard Errors) and 95% Confidence Intervals (± 2 Standard Errors) Calculated from the Mean Percent Correct Recognition Data for One-,

Two-, and Three-Pair digits in the Free-Recall Condition for Each Age Group

99.7%

Left Ear

Confidence Interval

Right Ear

95%

Left Ear

Confidence Interval

Right Ear

20-29 years One pair 98.9-100 .3 99.1-100 .3 99.2-100 .1 99.3-100 .1 Two pair 94.3-99 .5 95.0-100 .2 95.2-98 .5 96.3-99 .0 Three pair 85.1-94 .5 88.8-96 .6 86.8-92 .9 90.1-95 .2

30-39 years One pair 91 .5-100 .5 94.9-101 .7 93.2-99 .7 96.1-100 .5 Two pair 77 .5-95 .5 92.1-98 .1 80.7-92 .4 93.2-97 .1 Three pair 67 .1-84.9 80.5-89 .9 70.2-81 .8 82.1-88 .3

40-49 years One pair 92 .6-100 .6 96.7-101 .2 93.9-99 .2 97.4-100 .4 Two pair 81 .2-95 .4 89.9-98 .7 83.6-92 .9 91 .4-97 .2 Three pair 69.4-86 .2 76.1-90 .5 71 .5-83 .1 78.6-88 .1

50-59 years One pair 88.8-98 .2 94.1-99 .7 90.5-96 .6 95.0-98 .7 Two pair 69.6-87 .3 86.1-95 .3 72.6-84 .2 87.7-93 .6 Three pair 59.6-76 .8 69.8-84 .2 62.5-73 .8 72.3-81 .7

60-69 years One pair 85.8-101 .4 93.5-100 .3 88.5-98 .7 94.7-99 .2 Two pair 74.9-92 .7 86.7-97 .1 77.9-89 .6 88.5-95 .2 Three pair 63.1-80 .9 72.7-87 .7 66.2-77 .8 75.3-85 .1

70-79 years One pair 83.2-97 .4 89.9-100 .6 85.6-94.9 91 .6-98.8 Two pair 65.3-83 .5 84 .6-96 .6 68.4-80.3 86 .7-94 .5 Three pair 54.4-74.6 70 .8-86 .1 57 .9-71 .1 73 .4-83.4

level since pure-tone audiometric thresholds differed between age groups . Several findings argue against the likelihood that the findings are a result of an increase in pure-tone thresholds as a function of increasing age. First, if reduced pure-tone sensitivity was a factor affecting per-formance, then we would expect reduced recog-nition performance on the easy one-pair items as well as the more difficult two- and three-pair items. The results indicated, however, that recog-nition performance was within the range of nor-mal for one-pair dichotic digits in all age groups. Similarly, if hearing loss had an effect on dichotic digit performance, then we would expect also that subjects would demonstrate reduced recog-nition performance on directed-recall items. This finding was not supported by the present data . Second, the observed differences in ear performance across tasks do not suggest a gen-eral inability to hear the dichotic stimuli. Third, monosyllabic word recognition performance was evaluated in each subject at 70 dB HL, which was the same presentation level used for dichotic digit test stimuli. As indicated previously, there was no significant difference in monaural word recognition performance between groups . Finally,

previous investigations using dichotic digit mate-rials support the notion that digits are rela-tively immune to the effects of cochlear hearing loss (Musiek, 1983; Speaks et al, 1985 ; Musiek et al, 1991). These results support the conclusion that there was an overall decline in dichotic digit performance with age that cannot be eas-ily attributed to the effects of peripheral hear-ing loss. The results of the ANCOVA support this conclusion .

As the normative data illustrate, there was a difference between left and right ear recogni-tion performance for one-, two-, and three-pair digits . Previous investigations on dichotic tasks report that subjects, especially right-handed subjects, tend to exhibit a right-ear advantage more often than a left-ear advantage (Lake and Bryden, 1976 ; McKeever et al, 1984 ; Bryden, 1988 ; Wilson and Leigh, 1996). The present data support this finding. The recognition perfor-mance on materials presented to the right ear was better than the recognition performance on materials presented to the left ear for free- and directed-recall conditions . Additionally, as the complexity of the listening task increased from easy to difficult, there was a corresponding

568

Page 13: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Dichotic Digit Recognition under Free and Directed Recall/Strouse and Wilson

decrease in recognition performance for all age groups .

In the free-recall condition, recognition of one-, two-, and three-pair digits presented to the left and right ears decreased as a function of age, with the decrease faster for the left ear than for the right ear. Consequently, the right-ear advan-tage exhibited by the older adult groups was larger than the right-ear advantage of the

younger adult groups . When results were viewed in terms of the slopes of the functions, the data for the free-recall condition indicated that (1)

recognition of one-, two-, and three-pair digits presented to the left and right ears decreased as a function of age ; (2) scores obtained for the left ear decreased as a function of age faster than scores obtained for the right ear ; and (3) the

slope of the functions increased as the diffi-

culty of the task increased . These findings are in agreement with previous studies reporting a

progressively larger right-ear advantage (or left-ear disadvantage) in dichotic listening with

advancing age (Johnson et al, 1979; Jerger et al, 1990c ; Wilson and Jaffe, 1996) . The analy-

sis of ear effects also indicated that when the

dichotic task involved little difficulty as with the one-pair stimuli, there was only a small, if any, right-ear effect. When the listening task became more difficult, as with the three-pair stimuli, the difference between ear preference increased .

For the protocol used in the current study, 108 stimulus items from the free-recall condi-tion were presented . These stimulus items were

subsequently recorded as Tracks 7 and 8 on the VA CD Tonal and Speech Materials forAudi-tory Perceptual Assessment, Disc 2.0 . Of the 108 stimulus items, 54 items were recorded as Track 7 and 54 items were recorded as Track 8 . We compared recognition scores from the first half of the test (Track 7) to the second half of the test (Track 8) in an effort to determine whether significant practice effects were present . Results

of an ANOVA indicated a significant main effect

of half (F [1, 1741 = 33.8, p < .0001), indicating that recognition performance on the second 54

test items was better than recognition perfor-mance on the first 54 items of the test . Percent differences between first and second half scores across age groups ranged from 0.1 percent to 6.3 percent . The largest improvement (6.3%) was in

the left ear for one-pair digits in the 70- to

79-year group . As each token in the free-recall

condition equaled 2.8 percent, this difference

amounted to only two tokens . Thus, it is feasi-ble that the interleaved dichotic digits test could

be administered using one list of 54 items as

opposed to 108 items, so that Tracks 7 and 8 on

the VA CD could be used as independent mea-sures .

Free Recall vs Directed Recall

A secondary objective of this investigation was to compare recognition performance between free- and directed-recall response con-

ditions in an effort to differentiate between

auditory/structural and cognitive/task-related deficits in dichotic listening . According to pre-vious investigators (Bayles and Kaszniak, 1987 ;

Jerger et al, 1990, 1994 ; Chmiel and Jerger, 1996), the free-recall condition places greater stress on attentional resources than the directed-recall condition . In the free-recall con-dition, the subject has to attend to and report digits heard in both ears . In the directed-recall condition, the subject hears digits presented to

each ear but can focus on the digits heard in the

cued ear while disregarding the digits presented to the noncued ear, resulting in a reduced demand on cognitive processing . As Jerger et al (1990c) described, when recognition performance

is below normal in the free-recall condition but

improves substantially in the directed-recall

condition, then the problem is primarily in the cognitive domain since memory and attention

abilities are insufficient for successful perfor-

mance when both ears must be monitored simul-

taneously. When performance is below normal

in both the free- and directed-recall conditions,

then the problem is interpreted to be primar-

ily in the auditory domain since performance does not improve by reducing cognitive demands . Results of the present study indicate

that of the 180 subjects, 58 percent showed no

deficit in either ear under either response con-dition, 39 percent showed below-normal per-

formance in one ear in the free-recall condition and a normal response in the directed-recall con-

dition, and 3 percent scored below normal in

both response conditions . Thus, the overall recognition performance of 97 percent of the

subjects (174 of 180) was categorized as normal

and the recognition performance of just 3 per-cent of the subjects (6 of 180) was categorized

as below normal (suggesting an auditory-spe-cific deficit) . Interestingly, all subjects in this

group complained of substantial listening dif-

ficulties, especially in the presence of compet-

ing noise ; however, this complaint was not

exclusive to this group of listeners . The number of subjects identified in this

study as possibly having an auditory-specific

569

Page 14: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Journal of the American Academy of Audiology/Volume 10, Number 10, November/December 1999

deficit is considerably less than that reported in previous studies using dichotic sentence mate-rials in a similar paradigm (Chmiel and Jerger, 1996 ; Jerger et al, 1990b, 1994) For example, Chmiel and Jerger evaluated 115 elderly subjects over the age of 60 years with high-frequency sensorineural hearing loss . Using the free- and directed-recall conditions for the Dichotic Sen-tence Identification Test, 33 of the 115 subjects (29%) were categorized as having an auditory-specific deficit. No specific information was offered characterizing the remaining 61 percent of subjects who were categorized as normal . Using the same paradigm, Jerger et al (1990) administered dichotic sentences to 172 elderly subjects (60-90 years) with various degrees of sensorineural hearing loss . Of the total group of 172 subjects, 19 percent showed no deficit in either ear under either condition, 58 percent showed below-normal performance in one ear in the free-recall condition and a significant improvement in the directed-recall condition, and 23 percent showed a deficit in both response conditions .

The smaller percentage of subjects identified in the current study as having an auditory-specific deficit as compared to previous work may be attributed to the age range of the sub-jects evaluated. All subjects in the dichotic sen-tence investigations were over the age of 60 years, whereas the current study evaluated sub-jects ranging in age from 20 to 79 years. Because the incidence of central auditory deficit increases with age (Stach et al, 1990; Jerger, 1992), it is not surprising that a smaller percentage of sub-jects was identified from the population under study in this investigation. For a better com-parison to previous work, a smaller sample was formed by eliminating from the original group of 180 subjects all individuals under the age of 60 years. This resulted in a group of 60 subjects in the age range of 60 to 79 years. Within this group, 52 percent showed no deficit in either ear under either the free- or directed-recall condition, 42 percent showed below-normal performance in one ear in the free-recall condition and a sig-nificant improvement in the directed-recall con-dition, and 7 percent showed a deficit in both response conditions . The greater percentage of results consistent with cognitive deficit as a function of age is not unexpected given the well-documented age-related decline in various dimensions of cognitive function (CHABA Work-ing Group, 1988). Because subjects in the cur-rent study were not screened to determine cognitive functioning, this finding must be con-

sidered speculative. In support of the explana-tion based on cognitive status, however, it has been suggested that the presence of a cognitive deficit should affect results in the free-recall condition more than in the directed-recall con-dition since the demands on attentional resources are presumably greater in the free-recall condition (Jerger et al, 1990b, 1994). This finding is supported by the current data . The data for all subject groups indicated that the recognition performance in the free-recall con-dition was poorer than recognition in the directed-recall condition.

The smaller number of subjects in the 60-to-79 age range identified with an auditory-specific deficit in the current study (7%) as com-pared to previous work may be attributed to differences in test materials and response tasks. One might speculate that for elderly subjects, the response task associated with the Dichotic Sen-tence Identification Test is more complicated than the response task associated with dichotic digits . Specifically, the response task of (1) iden-tifying the sentence(s) heard from a list of 10 sen-tences, (2) identifying the number corresponding to the sentence, and (3) reporting the number of the sentence(s) heard involves more cognitive steps than the simple recall of digits, which are generally familiar to most listeners.

CONCLUSION

F rom these data, we propose that the ran-domly interleaved one-, two-, and three-pair digits included on the Tonal and Speech Mate-rials forAuditory Perceptual Assessment, Disc 2.0 (1998) are a useful tool for evaluating dichotic digit performance. The use of this hierarchical set of interleaved dichotic digits enables a wide range of recognition performances to be investi-gated. Moreover, modifying the materials to include both free- and directed-recall conditions may better delineate between auditory-specific deficits and those involving cognitive aspects of cognition.

Acknowledgment. This work was supported by Rehabili-tation, Research and Development Service and by Medical Research Service, Department of Veterans Affairs. The first author is on a Career Development Award sponsored by the Rehabilitation, Research and Development Service. Appreciation is expressed to Nicole Brush, M.S ., and Sarah Seehorn, M.S ., for assistance with data collection and to John Kalbfleisch, Ph.D., for assistance with sta-tistical analysis . This paper was presented at the Annual Convention of the American Speech-Language-Hearing Association, San Antonio, TX .

570

Page 15: Recognition of One-, Two-, and Three-Pair Dichotic …...(F [5, 1741 = 1.8, p > .05). Procedures The dichotic digits were reproduced by a CD player (Sony, Model CDP-497) and fed through

Dichotic Digit Recognition under Free and Directed Recall/Strouse and Wilson

REFERENCES

American National Standards Institute. (1996). American National Standard Specification forAudiometers . (ANSI S3-1996) . New York : ANSI .

Bayles K, Kaszniak A . (1987) . Communication and Cognition in Normal Aging and Dementia . Boston : College Hill .

Jerger J, Oliver T, Pirozzolo F. (1990b). Impact of central auditory processing disorder and cognitive deficit on the self-assessment of hearing handicap in the elderly. J Am Acad Audiol 1:75-80 .

Jerger J, Stach B, Johnson K, Loiselle L, Jerger S. (1990c). Patterns of below normality in dichotic listening in the elderly. In : Jensen J, eds. Presbyacusis and Other Age Related Aspects. Copenhagen : Stougaard Jensen,143-150 .

Bryden MP. (1988) . An overview of the dichotic listening procedure and its relation to cerebral organization . In : Hugdahl K, ed. Handbook of Dichotic Listening: Theory, Methods, and Research. New York : John Wiley and Sons, 1-43 .

Bryden MP, Munhall K, Allard F. (1983) . Attentional biases and the right-ear effect in dichotic listening. Brain Lang 18:236-248 .

Chmiel R, Jerger J. (1996) . Hearing aid use, central audi-tory disorder, and hearing handicap in elderly persons. J Am Acad Audiol 7:190-202 .

Committee on Hearing, Bioacoustics, and Biomechanics (CHABA). (1988) . Speech understanding and aging. J Acoust Soc Am 83:859-893 .

Department of Veterans Affairs . (1991) . Speech Recognition and Identification Materials, Disc 1.1 . Long Beach, CA : VA Medical Center.

Department of Veterans Affairs. (1992). Tonal and Speech Materials for Auditory Perceptual Assessment, Disc 1 .0. Long Beach, CA : VA Medical Center .

Department of Veterans Affairs. (1998) . Tonal and Speech Materials for Auditory Perceptual Assessment, Disc 2.0. Mountain Home, TN : VA Medical Center.

Dubno JR, Dirks DD, Morgan DE . (1984) . Effects of age and mild hearing loss on speech recognition in noise. J Acoust Soc Am 76:87-96 .

Fifer R, Jerger J, Berlin C, Tobey E, Campbell J. (1983) . Development of a dichotic sentence identification test for hearing impaired adults . Ear Hear 4:300-305 .

Humes LE, Coughlin M, Talley L. (1996) . Evaluation of the use of a new compact disc for auditory perceptual assessment in the elderly. JAmAcadAudiol 7 :419-427 .

Jerger J. (1992) . Can age-related decline in speech under-standing be explained by peripheral hearing loss? JAm Acad Audiol 3:33-38 .

Jerger J, Chmiel R, Allen J, Wilson A. (1994) . Effects of age and gender on dichotic sentence identification . Ear

Hear 15:274-287 .

Jerger J, Jerger S, Oliver T, Pirozzolo F. (1989) . Speech understanding in the elderly. Ear Hear 10:79-89 .

Jerger J, Johnson K, Jerger S, Coker N, Pirozzolo F, Gray L . (1991) . Central auditory processing disorder : a case study. J Am Acad Audiol 2:36-54.

Jerger J, Mahurin R, Pirozzolo F. (1990a). The repara-bility of central auditory and cognitive deficits : implica-tions for the elderly. J Am Acad Audiol 1:116-119 .

Johnson RC, Cole RE, Bowers JK, Foiles SV, Patrick JW, Woliver RE . (1979) . Hemisphere efficiency in middle and later adulthood. Cortex 15:109-119 .

Kimura D. (1961) . Cerebral dominance and the percep-tion of verbal stimuli. Can J Psychiatry 15:156-165 .

Lake DA, Bryden MP. (1976) . Handedness and sex differences in hemispheric asymmetry. Brain Lang 3:266-282 .

McKeever WF, Nolan DR, Diehi JA, Seitz KS . (1984) . Handedness and language laterality : discrimination of handedness groups on the dichotic consonant-vowel task. Cortex 20:509-523 .

Musiek FE . (1983) . Assessment of central auditory dysfunction: the dichotic digit test revisited . Ear Hear 4:79-83 .

Musiek FE, Gollegly KM, Mbbe KS, Verkest-Lenz SB . (1991) . Proposed screening test for central auditory dis-orders : follow-up on the dichotic digits test . Am J Otol 12:109-113 .

Ryan W, McNeil M. (1974) . Listener reliability for a dichotic task . JAcoust Soc Am 56:1922-1923 .

Speaks C Niccum N, Carney E. (1982) . Statistical prop-erties of responses to dichotic listening with CV nonsense syllables. JAcoust Soc Am 72:1185-1194 .

Speaks C Niccum N, Van Tasell D. (1985) . Effects of stim-ulus material on the dichotic listening performance of patients with sensorineural hearing loss . J Speech Hear Res28:16-25 .

Stach BA, Spretnjak ML, Jerger J. (1990) . The preva-lence of central presbycusis in a clinical population . J Am Acad Audiol 1:109-115 .

Strouse AL, Hall JWH. (1995) . Test-retest reliability of a dichotic digits test for assessing central auditory func-tion in Alzheimer's disease. Audiology 34:85-90 .

Strouse A, Wilson RH . (1999) . Stimulus length uncer-tainty with dichotic digit recognition. JAm Acad Audiol

10:219-229 .

Studebaker J. (1985) . A "rationalized" arcsin transfor-mation . J Speech Hear Res 28:455-462 .

Wilson RH, Jaffe MS . (1996). Interactions of age, ear, and stimulus complexity on dichotic digit recognition . J Am

Acad Audiol 7:358-364 .

Wilson RH, Leigh ED . (1996) . Identification performance by right- and left-handed listeners on dichotic CV mate-rial. J Am Acad Audiol 7:1-6 .