mlds in forward and backward masking

LETTERS TO THE EDITOR

Navy Submarine Med. Res. Lab., Groton, Connecticut, Rep. No. 591.

Kmrx•R, K., Woox), W., Mx•J.v.R, J., and Exa)amov., D. (1%6). "Hazardous Exposure to Intermittent and Steady-State Noise," J. Acoust. Soc. Amer. 39, 451-464. Sv.•O•.ANT, R. L., and DuvrY, J. (1970). "Acoustic Analysis of

Speech in Air Compressed from 14.7 and 147 PSIA," J. Acoust. Soc. Amer. 47, 128.

Su•amT•, J. (1970). (personal communication, Washington, D.C.)

WALSH--HEALEY Pu•.zc CONTRACTS ACT (1969). Federal Register 34, No. 96, 7946-7949.

Received 26 March 1971 4.3, 4.7, 4.8, 4.15

MLDs in Forward and Backward Masking

A.M. SMALL, J. BootESS, R. KLxca, D. KUEHN, J. THELIN, AND T. WXLEY

University of Iowa, Iowa City, Iowa 52240

Masking-level differences (MLDs) of substantial magnitudes were found in forward and backward masking situations. Their magnitude diminishes exponentially as the silent interval between signal and masker increases. Implications of the data for current models relating to the MLD are suggested.

INTRODUCTION I. PROCEDURE

The release from masking brought about by a shift in the interaural phase of either signal or masker in a binaural listening situation [masking-level difference (MLD)• is an intriguing phenomenon (Hirsh, 1948a). Models developed to account for this phenomenon, for example, Jeffress et al. (1956) and Durlach (1963), have for the most part been concerned with steady-state effects. They have to a lesser degree addressed themselves to questions relating to time-dependent proper- ties of the auditory system and neural representation. However, there are certain experimental conditions which force us to attend to these latter considerations

to a greater extent than is traditional. For example, if MLDs obtain when signal and masker are not simultaneously present, it may be necessary to modify or extend those models which postulate that arithmetic operations are performed on binaural waveforms. The research that we describe deals specifically with whether MLDs, conventionally defined, can be observed in backward and forward masking conditions.

The masker was a 500-msec wide-band noise whose

pressure spectrum level was 46 dB re 0.0002 ubar and whose rise/fall time was about 0.1 msec. The signal con- sisted of five complete cycles of a 250-Hz sinusoid gated without envelope shaping at zero axis crossings. Data were gathered for NOS0, NOS•r and SO, S•r with no noise. A two-alternative temporal forced-choice (2ATFC) method was used, and for each condition the signal level required for d'= 1.0 was estimated. Figure 1 illustrates the stimulus paradigm for each section of the experi- ment. The detection criterion value was estimated by randomly presenting either of two signal levels on any given trial. Based on preliminary work, these signal levels were chosen so that they bracketed the level corresponding to d'= 1.0 (approximately 68o/0 and 84% correct). The exact level corresponding to d'= 1.0 was then determined by interpolation. Five normal-hearing listeners, while experienced generally in listening tasks including those involving signals of short duration, received no specific practice in this experimental task.

Fro. 1. Stimulus paradigm for each of the three kinds of temporal masking. A 2ATFC method was used in which the duration of the observation intervals (O•, 02) was 1 sec. The masker appeared at the midposition of both observation intervals, while the signal occurred in either O• or O2 with a probability of 0.5. The interstimulus interval, ISI, was defined as shown and took values' of 5, 10, 20, 30, and 50 msec. For the "simultaneous" condition, the signal was tem- porally centered in the noise. A warning light was presented 1 sec prior to O• and a vote light im- mediately followed 02. The listener's vote caused the correct response to be displayed and initiated a new stimulus Sequence. The duration of the re- cycle interval was at least 1 sec.

SIGNAL

I

Ol

MASKER

!

i

•.,•

500 MSEC ' i

FOR;WARD i

i i

5 O0 MSEC

SIMULTANEOUS

ß ' TIME

•-• 0 2

' ,", MASKER

, ,,,,, BACK:WARD SIGNAL

20 MSEC

The Journal of the Acoustical Society of America 1365

Redistribution subject to ASA license or copyright; see http://acousticalsociety.org/content/terms. Download to IP: 140.254.87.149 On: Sat, 20 Dec 2014

07:00:49


z IO

z

II

6

z

• o

Z -•

0

0

TIME

I I ,

40 20 BACKWARD

' I ' ' 5 5 I0 SIMULTANEOUS

IN MILLISECONDS

i

2o 4'0 FORWARD

Fro. 2. Average MLDs for backward, forward, and simultaneous masking. This MLD is the difference in signal level required for d'-1.0 for the NOS0 and NOS•r interaural phase conditions. The straight-line functions are visually fitted to the data points, each of which is based on 400 judgments/ listener.

II. RESULTS

Figure 2 shows the average MLD (NOS0-NOS•r) for each of the experimental conditions pooled across listeners. Note that the abscissa is logarithmic. The functions are simply visual best-fit straight lines and indicate to a first approximation that the MLD is an exponential function of the silent interval between signal and masker. Although not presented here, our data also suggest that the MLD is nearly proportional to the amount of masking produced. This latter finding is the same as is seen in simultaneous masking data of Hirsh (1948b)and Canahl (1965). Thus, the differ- ential slope of the MLD function for forward and backward masking reflects the amount of masking produced for each condition. Similar slope differentials have been reported for simple monaural masking by Pickett (1959), Elliott (1962), and others. The magnitude of the MLD produced in the simultaneous masking condition, 10.5 dB, is somewhat less than would be ex- pected based on earlier studies under similar although not identical conditions, 12-15 dB (Green, 1966; McFadden, 1966).

III. DISCUSSION

In those instances in which our experimental conditions were similar to those of earlier studies, similar results were obtained. However, more importantly, our data show clearly tha• MLDs may be produced by a signal and masker that are not simultaneously present at the ears. The data of Deatherage and Evans (1969) suggest a similar conclusion, although their stimulus paradigms do not allow a conventional measurement of the magnitude of the MLD. Also Doland and Trahiotis (1970) presented a paper in which they report the existence of an MLD for backward masking.

Our results have several implications for mechanisms underlying the MLD phenomenon as well as those re- sponsible for backward and forward masking. It has been suggested that backward and forward masking are reflections of fundamentally different processes. Forward masking is often thought of as originating in the peripheral portion of the auditory system (Ltischer and Zwislocki, 1949) while backward masking is generally conceded to be central in origin (Elliott, 1962). Other data, for example, those of Deatherage and Evans (1969), suggest that forward masking is mediated, in part at least, centrally. Our results allow us to go a step further. Not only do they support the idea that both backward and forward masking are largely central in origin, but they indicate that these two forms of masking may be the result of similar processes. In other words, however the various aspects of the stimulus may be represented neurally, their binaural (central) interaction produces the same result, an MLD, whether the signal precedes or follows the masker.

Our data also suggest a reexamination of current theories dealing with the MLD to the extent that they may require the simultaneous presence of signal and masker. For example, the simple subtraction of the signal-masker waveforms at the two ears thus eliminat- ing the in-phase components seems unlikely as does the production of an interaural time difference between the waveforms at the two ears by the addition of an out- of-phase signal (or masker). Although it is clear that the interactions between signal and masker and between left and right ear must be in terms of the neural representation of the stimuli, it is altogether too easy to fall into the habit of thinking of the stimuli themselves interacting. It is important to note that although the stimuli may be clearly nonsimultaneous, it is presumed that the neural representations of the signal and masker overlap in time. One obvious way in which our data

1366 Volume 51 Number 4 (Part 2) 1972


07:00:49


could be reconciled with current theories is to assume

that the neural representation contains information about the phase as well as the magnitude of the stimuli

regardless of whether the stimuli are present at that moment. Whether phase information can be preserved neurally in this way is an as yet unanswered question.

REFERENCES

CA, Am., J. A. (1965). "Binaural Masking Level Differences for 167, 250, 500 cps and Noise Signals at Various Levels of Masking Noise," MA thesis, Univ. Iowa (unpublished). D•.ATa•.RAG•., B. H., and EVANS, T. R. (1969). "Binaural Masking: Backward, Forward, and Simultaneous Effects," J. Acoust. Soc. Amer. 46, 362-371. Do•.•m>, T. R., and T•tAmoTis, C. (1970). "Binaural Interaction in Backward Masking," J. Acoust. Soc. Amer. 47, 131 (A). Dv•.•ca, N. I. (1963). "Equalization and Cancellation Theory of Binaural Masking-Level Differences," J. Acoust. Soc. Amer. 35, 1206-1213. E•J.•o•, L. L. (1962). "Backward and Forward Masking of Probe Tones of Different Frequency," J. Acoust. Soc. Amer. 34, 1116-1117.

GR•.•, D. M. (1966). "Interaural Phase Effects in the Masking of Signals of Different Durations," J. Acoust. Soc. Amer. 39, 720-724.

I-h•tsa, I. J. (1948a). "The Influence of Interaural Phase on Inter, aural Summation and Inhibition," J. Acoust. Soc. Amer. 20, 536-544.

H•tsa, I. J. (1948b). "Binaural Summation and Interaural Inhi- bition as a Function of the Level of Masking Noise," Amer. J. Psychol. 56, 205-213.

J•.rr•.ss, L. A., B•.o•)G•.•, H. C., S•m>m., T. T., and Woo•), C. L. (1956). "Masking of Tonal Signals," J. Acoust. Soc. Amer. 28, 416-426.

Lvsca•.R, E., and Zw•s•.oc•cI, J. (1949). "Adaptation of the Ear to Sound Stimuli," J. Acoust. Soc. Amer. 21, 135-139. MCFA•D•.N, D. (1966). "Masking-Level Differences with Con- tinuous and with Burst Masking Noise," J. Acoust. Soc. Amer. 40, 1414-1419. PIcx•T•, J. M. (1959). "Backward Masking," J. Acoust. Soc. Amer. 31, 1613-1615.

Received 9 June 1970

Cochlear-Microphonic and Middle-Ear Pressure Changes during Nitrous Oxide Anesthesia in Cats

A. STARR* A•t) P. SCaWARTZ•C•O•

Stanford University Medical Center, Department of Neurology, Stanford, California 94305

The amplitude of cochlear-microphonic responses in cats was attenuated up to 18 dB during both induction and recovery from nitrous oxide anesthesia, with smaller changes (up to 3 dB) occurring during the periods of maintained anesthesia. This response variability was clearly related to alterations of middle-ear pressure and could not be attributed to changes in middle-ear muscle or efferent olivocochlear bundle activities. Middle-ear pressure rose during inhalation of nitrous oxide with periodic abrupt returns to baseline levels during the period of maintained anesthesia; pressure became negative after withdrawal of nitrous oxide.

4.2.4

In the course of studying auditory responses in cats with chronically implanted round-window elec- trodes, we encountered a source of response amplitude variability attributable to the use of nitrous oxide inhalation anesthesia. Cochlear microphonic (CM) responses to tone signals (1-30 kHz) varied as much as 18 dB during both induction and recovery from nitrous oxide anesthesia and up to 3 dB during maintained anesthesia. Response variability was not the result of middle-ear muscle activity, since it persisted in cats paralyzed with gallamine triethiodide or d-tubo-

curarine, as well as in cats in which the tendons of the middle-ear muscles had been sectioned. Crushing the eighth nerve had no effect on the microphonic changes, indicating that the efferent olivocochlear bundle was also not involved in this phenomenon. Two lines of evidence suggested that nitrous oxide administration affected CM responses by influencing middle-ear pressure. First, nitrous oxide enters body cavities (such as the middle ear) before the normal major component of these cavities (nitrogen) has been resorbed (Matz et al., 1967; Rasmussen, 1967). If the cavity has rigid

Fro. 1. Middle-ear pressure (top trace) and integrated CM response to a steady 3-kHz tone (bottom trace) during induction of nitrous oxide anesthesia.

The Journal of the Acoustical Society of America 1361


07:00:49

mlds in forward and backward masking

Documents