Auditory evoked response, clicks, notch noise bandwidth, frequency

Transcription

1 1~14fYl~t~lilliill ~~~~l~lll ~ I i?~l~i i J Am Acad Audiol 3 : (1992) Effects of Notch Noise Bandwidth on the Auditory Brainstem Response to Clicks Randall C. Beattie* Debra L. Franzone* Kristen M. Thielen* Abstract The effects of notch width of notch masking noise on the auditory brainstem response (ABR) were investigated with ten normal-hearing subjects. Wave V latencies and amplitudes were measured to a click in quiet and in the presence of noise with notches centered around 1000 Hz and 4000 Hz. Notch width was either 0.5, 1.0, 1.5, 2.0, or 2.5 octaves. A 95 db SPL broadband noise was necessary to mask a 65 db nhl click at an effective level of 73 db nhl. All ten subjects at both 1000 Hz and 4000 Hz yielded identifiable responses for the click in quiet and for the 2.5 octave and 2.0 octave conditions, and eight or nine subjects responded to the 1.5 octave notch noise. In contrast, no responses were observed to the 0.5 octave condition, and only three or four subjects responded to the 1.0 octave notch noise. These findings suggest that when ABRs are obtained to 65 db nhl clicks in notch noise, the notch width should exceed 1.0 octave. Moreover, because of the relatively low amplitudes elicited with the 1.5 octave bandwidth, it appears preferable to select notches that are 2.0 to 2.5 octaves wide. Considering the wide bandwidth required, the high noise levels necessary to mask clicks, high ABR thresholds, and the difficulty setting the signal-to-noise ratio, it appears that tonepips are more promising than clicks in notch noise for assessing frequency specific ABRs. Key Words : specificity Auditory evoked response, clicks, notch noise bandwidth, frequency C licks frequently are used as stimuli in auditory brainstem response (ABR) audiometry because they optimize response identifiability. A disadvantage of clicks, however, is that they have a broad frequency spectrum, which conforms closely to the frequency response of the earphone. One solution to this problem is to mix notch noise with the clicks. Limited research has been conducted using this procedure and the results have been inconsistent (Picton et al, 1979 ; Pratt and Bleich, 1982 ; van Zanten and Brocaar,1984 ; Stapells et al, 1985). One variable that may account for the differences among these studies is the bandwidth ofthe notch. Pratt and Bleich (1982) used *Department of Communicative Disorders, California State University, Long Beach, California Reprint requests : Randall C. Beattie, Department of Communicative Disorders, California State University, Long Beach, CA octave notches, whereas van Zanten and Brocaar (1984) employed a bandwidth of 1.67 octaves. Although frequency specificity improves as notch width is decreased, the ABR amplitude also is expected to decrease. In contrast, although a relatively wide notch may not be as frequency-specific as desired, the wider notch may be required to yield an ABR of sufficient magnitude for identifiability. The present study was designed to investigate the effects of notch width (0.5,1.0,1.5, 2.0, and 2.5 octaves) on the identifiability, latency, and amplitude of the ABR (wave V). Several authors suggest that an ABR evaluation for estimating auditory sensitivity should include both high-frequency and low-frequency stimuli (Davis and Hirsh, 1979 ; Jerger et al, 1985). Beattie and Spence (1991) suggest that using notch noises centered around 1000 Hz and 4000 Hz may provide a useful estimate of audiometric configuration. These authors did not recommend stimuli below 1000 Hz because

2 Journal of the American Academy of Audiology/Volume 3, Number 4, July 1992 they observed few ABRs and high electrophysiologic thresholds when a click was mixed with a notch noise centered around 500 Hz. In view ofthese considerations, we selected notches centered around 1000 Hz and 4000 Hz for study. Subjects METHOD Two groups of ten normal-hearing subjects participated in the study. One group (1000 Hz) consisted of ten men who ranged in age from 18 to 29 (mean = 23 years). The second group (4000 Hz) was composed of five women and five men ranging in age from 21 to 35 years (mean = 25 years). The right ear was selected for testing. Instrumentation and Calibration Responses were recorded with three electrodes using a vertex-neck-neck montage (Beattie et al, 1986). The output from the electrodes was directed to a physiologic amplifier (Grass, Model P51 1K) with a bandpass of 30 to 3000 Hz. The signal was directed from the physiologic amplifier to the signal averager (RC Electronics). The test stimulus consisted of rarefaction clicks, which were produced by directing a 100- microsecond rectangular pulse from a signal generator (RC Electronics, Model 200) to an audiometer (Grason-Stadler, Model GSI-16). The output of the audiometer was directed to an earphone (Telephonics, Model TDH-50P) encased in a cushion (Telephonics, Model P/N 51). The clicks were presented at a repetition rate of 25.6/sec. The notch masking noise was produced by directing white noise (Grass, Model S10) to two cascaded highpass filters (Krohn-Hite, Models 31) and to two cascaded lowpass filters (Krohn- Hite, Models 30). Two filters of each type were used to increase rejection rates from a nominal 115 db/octave to 230 db/octave. The filters were adjusted to achieve 0.5, 1.0, 1.5, 2.0, and 2.5 octave notches centered around 1000 Hz and 4000 Hz. Behavioral click thresholds in quiet were obtained on a group of 12 normal-hearing subjects and designated as 0 db nhl. The corresponding levels for the peak SPL and peak equivalent SPL were 30 db and 26 db, respectively. Previous authors have noted that the relative level of the click with respect to the noise is critical (Pratt and Bleich, 1982 ; Hecox et al, 1989). We reasoned that setting the broadband noise at an effective level of 73 db nhl (i.e., 8 db above the 65 db nhl click) would ensure adequate masking yet approximate an optimum ABR. The corresponding SPL for the broadband noise was 95 db. To estimate the spread-of-masking, we obtained puretone thresholds in highpass masking and in lowpass masking for one subject at 1000 Hz and 4000 Hz. This testing revealed slopes of approximately 85 db/octave for the highpass filters, but only about 30 db/octave for the lowpass filters. This substantial spread-ofmasking is consistent with previous research (Picton et al, 1979 ; Chung, 1981 ; Beattie and Boyd,1985) and yielded a notch depth for the 0.5 octave condition of only about 10 db. Procedures Initially, each subject was tested with 65 db nhl clicks in quiet, in a 95 db SPL broadband noise, and in notch noises centered around 1000 Hz or 4000 Hz with bandwidths of 0.5, 1.0, 1.5, 2.0, and 2.5 octaves. Because the first three subjects from each group showed no responses to the 0.5 octave bandwidth on three tracings of 5000 repetitions each, we omitted the 0.5 octave condition with the remaining subjects. Preliminary testing included two tracings in response to a 65 db nhl click in quiet to verify a normal ABR. Additionally, to ensure that the noise was sufficiently intense to mask the ABR, tracings to 5000 stimulus repetitions were obtained with the click (65 db nhl) mixed with the 95 db SPL broadband noise. A minimum of two tracings and a maximum of three tracings were obtained for each condition to assess response replicability. Three thousand stimulus repetitions were used for the quiet condition and 5000 repetitions were employed during the notch noise conditions. RESULTS Response Detectability We examined the frequency ofoccurrence of wave V in response to clicks presented at 65 db nhl in quiet and in notch noises centered around 1000 Hz and 4000 Hz. The results revealed that all ten subjects in both the 1000 Hz and 4000 Hz groups yielded identifiable responses for the click in quiet and for the 2.5 octave and 2.0 octave conditions. For the 1.5 octave notch noise, eight subjects responded at 270

3 mti1 vi1' ' 1111', -1 oi it"i o f vi 1c Notch Noise Bandwidth and ABR to Clicks/Beattie et al Table 1 Data for Wave V Latencies Notch Bandwidth (Octave) Frequency Quiet Hz M Md SD Range Hz M Md SD Range Measurements are in milliseconds and were in the presence of notch masking noise. M = mean ; Md = median ; SD = standard deviation Hz and nine subjects responded at 1000 Hz. In contrast, the 1.0 octave notch noise condition yielded responses in four subjects for the 4000 Hz group and in three subjects for the 1000 Hz group. Latencies Table 1 presents means, medians, standard deviations, and ranges for wave V latencies for the quiet and notch noises centered around 1000 Hz and 4000 Hz. Because the data for the 1.0 octave bandwidth are based on three or four subjects, they must be viewed with reservation. The mean data also are illustrated in Figure 1. Both groups exhibited similar mean latencies (-6.15 msec) and standard deviations (-0.22 msec) for the quiet condition. Compared to the quiet condition, mean latencies increased substantially ( msec) when the clicks were presented in notch noises centered around 1000 Hz. Moreover, longer latencies were observed when the notches were 1.0 and 1.5 octaves wide (-8.9 msec) than when the bandwidths were 2.0 and 2.5 octaves (-8.2 msec). For the 4000 Hz group, Figure 1 shows that latency was shortest for the quiet condition (6.08 msec) and systematically increased as notch bandwidth decreased from 2.5 octaves (6.23 msec) to 1.0 octave (6.62 msec). The trends described above can be seen in Figure 2, which presents representative waveforms for the quiet and notch noise conditions. I N M S E C 5.5 r i J 6.0 QUIET BANDWIDTH Figure I Latency in milliseconds (msec) is displayed as a function of bandwidth (quiet, and 2.5, 2.0, 1.5, and 1.0 octave bandwidths) for notches centered around 1000 Hz and 4000 Hz. Figure 2 Representative waveforms obtained for the quiet condition, broadband noise (BBN) condition, and notch noise conditions centered around 1000 Hz and 4000 Hz. 271

4 II Journal of the American Academy of Audiology/Volume 3, Number 4, July 1992 Table 2 Data for Wave V Amplitudes Frequency Quiet Hz M Md SD Range Hz M Md SD Range Notch Bandwidth (Octave) Measurements are in nanovolts and were in the presence of notch masking noise. M = mean, Md = median ; SD = standard deviation. Each panel displays the average waveform of two tracings, representing 6000 to stimulus repetitions. Amplitudes Descriptive statistics for wave V amplitudes are presented in Table 2 for the quiet and notch noise conditions centered around 1000 Hz and 4000 Hz. Standard deviations were largest for the quiet conditions and tended to decrease as bandwidth decreased from 2.5 octaves to A M P L T UD 300 E I N 200 ~ - N V O L T S HZ HZ QUIET BANDWIDTH Figure 3 Amplitude in nanovolts (nv) is displayed as a function of bandwidth (quiet and 2.5, 2.0, 1.5, and 1.0 octave bandwidths) for notches centered around 1000 Hz and 4000 Hz. octave. The mean data also are illustrated in Figure 3. Note that amplitudes were largest for the quiet condition (-315 nv) and that they decreased systematically as bandwidth decreased from 2.5 octaves (-V220 nv) to 1.5 octaves (-100 nv). The 1.0 octave amplitudes must be viewed with caution because they are based only on the three or four subjects that had observable waves (i.e., the three or four subjects with the largest responses). DISCUSSION Response Detectability Recall that no responses were observed to the 0.5 octave condition, and that only 30 to 40 percent of our subjects responded to the 1.0 octave notch noise. These findings were found despite using a signal-to-noise ratio that approached the maximum (-8 db) and using very steep rejection rates (>200 db/octave). However, the results were not unexpected because the low-frequency spread-of-masking yielded slopes of -30 db/octave which, compared to the frequency response of the filters, reduced the notch depth. Our results are not consistent with those of Pratt and his associates (1982, 1984) who used a notch width of 0.5 octave. Pratt and Bleich (1982) reported notch depths of -15 db (acoustically determined) and stated that the clicks were only 5 to 10 db higher than the masking noise. This information, together with our psychoacoustic data showing substantial spread-of-masking into the notch, suggests that the level of the masking noise selected by Pratt and associates (1982, 1984) may not have been 272!3i1~'~~ I1k!VTq A!?M 1 1 1! JWI1 11! I'` ' , iljli~ill~l l1 liidl

5 inllllp 1111!11' I! I I Iil~l s 1'~. Sr i ; Notch Noise Bandwidth and ABR to Clicks/Beattie et al sufficient. That is, the ABRs may have been elicited by a broadband (unmasked) click rather than the restricted spectrum associated with the 0.5 octave notches. Latencies The latency of wave V was longer for 1000 Hz than 4000 Hz. These differences were about 1.96 msec for the 2.0 octave and 2.5 octave notches, and about 2.36 msec for the 1.0 octave and 1.5 octave notches. The longer latency for 1000 Hz is consistent with Don and Eggermont (1978) and van Zanten and Brocaar (1984), but in contrast to Pratt and his associates (1982, 1984) who found that latency was independent of the center frequency of the notch. As mentioned earlier, these discrepancies may be due to insufficient masking. Amplitudes The results revealed that amplitude was largest for the signal having the widest bandwidth (quiet condition) and that amplitudes decreased as the notch width narrowed from 2.5 octaves to 1.0 octave. These results were expected because the notch noise masks those cochlear regions outside the notch, thereby eliminating some of the neural elements that contributed to the unmasked response (Stapells and Picton, 1981 ; Thummler et al, 1981). The number of neurons excluded increases as notch width decreases, with a consequent attenuation in the amplitude of wave V. Moreover, the basal region of the cochlea may contribute to a more synchronized neural response for the quiet condition, as contrasted to the notch noise conditions, which mask the more basal regions of the cochlea (Eggermont and Odenthal, 1974 ; Picton et al, 1979). Examination of our amplitudes, in conjunction with the identifiability data, indicate that when ABRs are obtained to 65 db nhl clicks in notch noise, the notch width should exceed 1.0 octave. Moreover, because of the relatively low amplitudes elicited with the 1.5 octave bandwidth, it appears preferable to select notches that are 2.0 to 2.5 octaves wide. To assess the tradeoff between frequency specificity and identifiability, additional research is required comparing the 2.0 octave and 2.5 octave notch widths using both normal-hearing and hearingimpaired subjects. Additionally, the effects of notch bandwidth on the ABR should be investigated using low and moderate noise levels, which are associated with less spread of masking (Wegel and Lane, 1924). Perhaps narrower bandwidths (< 2.0 octaves) are preferable when testing at these lower intensities. Some authors recommend using clicks in noise to obtain a two-point audiogram in order to distinguish among rising, flat, and falling audiometric configurations (Eggermont, 1982 ; Fjermedal and Laukli, 1989). The present investigation suggests that presenting clicks in notch noise (2.0 to 2.5 octaves wide) centered around 1000 Hz and 4000 Hz may allow assessment of audiometric configuration. As noted by Beattie and Spence (1991), however, the upper range of testable hearing loss is limited by (1) a high ABR threshold and (2) the high noise levels necessary to mask the click (-95 db SPL of BBN is required to mask a 65 db nhl click). Several other frequency-specific procedures have been suggested for estimating sensitivity, including presenting tonepips in quiet, highpass noise, and notch noise (Davis and Hirsh, 1979 ; Picton et al, 1979 ; Kileny, 1981 ; Suzuki et al, 1981 ; Laukli, 1983 ; Davis et al, 1984 ; Beattie and Boyd, 1985 ; Jerger et al, 1985 ; Gorga et al, 1988). Considering the wide bandwidth required for high-level click in notch noise testing, and the several other problems discussed in this paper, it appears that tonepips are more promising stimuli for assessing the ABR. Research comparing these procedures is necessary to establish their relative advantages. REFERENCES Beattie RC, Beguwala FE, Mills DM, Boyd RL. (1986) Latency and amplitude effects of electrode placement on the early auditory evoked response. JSpeech Hear Disord 51 : Beattie RC, Boyd RL. (1985) Early/middle evoked potentials to tone bursts in quiet, white noise and notched noise. Audiology 24 : Beattie RC, Spence J. (1991) Auditory brainstem response to clicks in quiet, notch noise, and highpass noise. J Am Acad Audiol 2: Chung DY. (1981) Tone-on-tone masking in subjects with normal hearing and with sensorineural hearing loss. J Speech Hear Res 24 : Davis H, Hirsh SK. (1979) A slow brain stem response for low-frequency audiometry. Audiology 18 : Davis H, Hirsh SK, Popelka GR, Formby C. (1984) Frequency selectivity and thresholds of brief stimuli suitable for electric response audiometry. Audiology 23 : Don M, Eggermont JJ. (1978) Analysis of the click-evoked brainstem potentials in man using high-pass noise masking. JAcoust SocAm 63 :

6 Journal of the American Academy of Audiology/Volume 3, Number 4, July 1992 Eggermont JJ, Odenthal DW. (1974) Frequency selective masking in electrocochleography. Rev Laryngol Otol Rhinol 95 : Eggermont JJ. (1982) The inadequacy of click-evoked auditory brainstem responses in audiological applications. Ann N YAcad Sci 388: Fjermedal O, Laukli E. (1989) Low-level 0.5 and 1 khz auditory brainstem responses. Scand Audiol 18 : Gorga MP, Kaminski JR, Beauchaine KA, Jesteadt W. (1988) Auditory brainstem responses to tone bursts in normally hearing subjects. J Speech Hear Res 31 : Hecox KE, Patterson J, Birman M. (1989) Effect of broadband noise on the human brain stem auditory evoked response. Ear Hear 10 : Jerger J, Oliver T, Stach B. (1985) Auditory brainstem response testing strategies. In : Jacobson JT, ed. The Auditory Brainstem Response. San Diego : College-Hill Press, Kileny P. (1981) The frequency specificity of tone-pip evoked auditory brain stem responses. Ear Hear 2: Laukli E. (1983) High-pass and notch noise masking in suprathreshold brainstem response audiometry. Scand Audiol 12 : Picton TW, Ouellette J, Hamel G, Smith AD. (1979) Brainstem evoked potentials to tonepips in notched noise. J Otolaryngol 8: Pratt H, Bleich N. (1982) Auditory brain stem potentials evoked by clicks in notch-filtered masking noise. Electroencephalogr Clin Neurophysiol 53 : Pratt H, Ben-Yitzhak E, Attias J. (1984) Auditory brainstem potentials evoked by clicks in notch-filtered masking noise : audiological relevance. Audiology 23 : Stapells DR, Picton TW. (1981) Technical aspects of brainstem evoked potential audiometry using tones. Ear Hear 2: Stapells DR, Picton TW, Perez-Abalo M, Read D, Smith A. (1985) Frequency specificity in evoked potential audiometry. In : Jacobson JT, ed. The Auditory Brainstem Response. San Diego : College-Hill Press, Suzuki T, Hirai Y, Horiuchi K. (1981) Simultaneous recording of early and middle components of auditory electric response. Ear Hear 2: Thummler I, Tietze G, Matkei P. (1981) Brain-stem responses when masking with wide-band and high-pass filtered noise. Scand Audiol 10 : van Zanten GA, Brocaar MP. (1984) Frequency-specific auditory brainstem responses to clicks masked by notched noise. Audiology 23 : Wegel RL, Lane CE. (1924) The auditory masking of one pure tone by another and its probable relation to the dynamics of the inner ear. Phys Rev 23 : ~ie-i rlf~, IVil :!.I R