Previous psychophysical studies of hypoxia's effects on auditory sensitivity have provided mixed results but the weight of evidence supports the conclusion that sensitivity is unaffected by hypoxia. This conclusion is discrepant with that drawn from physiological studies in which hypoxia has been found to affect auditory-evoked response (AER) latency. One possible explanation of this discrepancy concerns the relatively low maximum frequency (8 kHz) for which hypoxia's effects were assessed in the psychophysical studies. We have extended the range of frequencies over which hypoxia's effects have been examined to include frequencies up to 16 kHz. Thresholds for 1, 8, 10, 12, 14 and 16 kHz tones were measured at levels of hypoxia equivalent to altitudes of 0, 1200, 2400 and 3700 m. Our results indicate that sensitivity for frequencies up to 16 kHz is unaffected by hypoxia. We suggest that AER latency does not always provide a valid measure of auditory sensitivity.

Key Words: hypoxia, auditory sensitivity, auditory threshold, hearing threshold

The effect of hypoxia on auditory sensitivity has been examined in a number of studies. In one of the earliest, Gellhorn and Spiesman observed that hypoxia equivalent to that experienced on exposure to an altitude of 6100 m and induced by breathing an appropriate gas mixture resulted in a significant decrease in auditory sensitivity. Other studies in which hypoxia was found to reduce sensitivity include those by McFarland, Klein and Klein et al.. In each of these studies, hypoxia was observed to reduce sensitivity to a subset of the 5-8 pure tone frequencies that were tested but either enhance or leave unaltered sensitivity to the remainder. Other studies, however, suggest that auditory sensitivity is unaffected by hypoxia. This conclusion is supported by the early studies of Lewis (18) and Bagby, who found that hypoxia induced by exposure to simulated altitudes high enough to result in "general functional impairment" and loss of consciousness had no effect on absolute thresholds for air or bone conducted sounds, and more recent studies by Curry and Boys and Burkett and Perrin. The study by Burkett and Perrin is the most comprehensive carried out to date and compared thresholds for six air conducted pure tones ranging in frequency from 250 to 8000 Hz measured at ground level and simulated altitudes of 4600 and 6100 m. Thresholds were found to be unaffected by hypoxia for all frequencies tested. The weight of evidence from the above studies suggests that auditory sensitivity is unaffected by hypoxia.

Hypoxia’s effects on the auditory system have also been examined using auditory-evoked response (AER) techniques. Deecke et al., Carlile et al., Carlile and Paterson, Fowler and Lindeis, Lucertini et al. and Wesensten et al. found that the latencies of various waves in human AERs increased significantly with increasing hypoxia level. Some of these authors  argued that wave latency provides an objective measure of auditory sensitivity and interpreted the latency increases they observed as evidence of a hypoxia-induced reduction in sensitivity. Using stimulus-level versus wave-latency functions measured for each of their subjects at sea level, Carlile and Paterson calculated that the latency increases observed in their study correspond to an average reduction in sensitivity of about 9 dB. In contrast, Mosko et al., Sohmer et al., Urbani and Lucertini and Bouchet et al. found that AER wave latencies were unaffected by hypoxia. This discrepancy may be accounted for in part by inter-study differences in the duration of exposure to hypoxia prior to AER recording. The studies by Mosko et al., Sohmer et al. and Bouchet et al.  involved particularly brief (1-2 minute) or long (24 and 72 hour) exposures that have been found by others to be outside the range associated with latency increases. Urbani and Lucertini, however, used exposures that were similar in duration to those used by Carlile et al. , Lucertini et al.  and Wesensten et al. and it is unclear why they did not observe the latency increases described by these other authors.

On the assumption that AER wave latency does provide an objective measure of auditory sensitivity, the above studies in which hypoxia was found to induce latency increases seem inconsistent with those in which hypoxia was observed to have no effect on psychophysical thresholds. This discrepancy cannot be explained with reference to the way in which hypoxia was induced, as both groups of studies included instances in which normobaric and hypobaric hypoxia were involved. It can, however, be partially explained with reference to the length of time subjects were exposed to hypoxia prior to the assessment of its effects. In the study by Burkett and Perrin referred to above, subjects were exposed to each simulated altitude for only five minutes before threshold measurement began. Carlile et al. found that latency increases do not develop until subjects have experienced hypoxia for at least 20 minutes and suggested that the shorter pre-measurement exposure period employed by Burkett and Perrin  could explain their failure to observe an effect. The length of time subjects in the studies by Lewis and Bagby were exposed to hypoxia prior to threshold measurement is unclear but as each session in the study by Bagby was completed within 30 minutes, it is unlikely that in that case it was greater than a few minutes. In the study by Curry and Boys, however, thresholds were measured following 30 minute's exposure to hypoxia, so the fact that they did not observe an effect of hypoxia cannot be explained in this way.

Another possible explanation of this discrepancy concerns the range of frequencies tested in the above psychophysical studies. In keeping with most audiometric studies, these psychophysical studies only considered frequencies up to 8 kHz. Human hearing, however, extends to considerably higher frequencies and it is possible that hypoxia has effects on thresholds for frequencies above 8 kHz. The cochlea’s sensitivity to high frequency sound in particular is dependent on the integrity of active cochlear processes that are likely to be especially vulnerable to hypoxia. As the majority of the above studies in which hypoxia-induced wave latency increases were observed involved stimuli with sudden onsets and AERs to such stimuli are thought to be dominated by responses to high frequency components a hypoxia-induced reduction in sensitivity to high frequencies may have underlied the latency increases observed.

In this study, therefore, we extended the frequency range over which hypoxia’s effect on auditory thresholds has been examined to cover frequencies up to 16 kHz. Thresholds for six frequencies ranging from 1 to 16 kHz were measured following at least 15 minute's exposure to simulated altitudes of 0, 1200, 2400 and 3700 m. As an additional 40-50 minutes were required to complete all measurements at each simulated altitude, the duration of exposure to hypoxia at the time of threshold measurement for subjects in this study must have fallen within the range associated with AER wave latency increases. Ideally, the range of simulated altitudes would have extended above 3700 m but that possibility was precluded on consideration of ethical issues.

Methods

Thresholds were measured for six frequencies (1, 8, 10, 12, 14 and 16 kHz) at each of four simulated altitudes (0, 1200, 2400 and 3700 m). Each subject participated in four experimental sessions during each of which a different altitude was tested. The order in which altitudes were tested was counterbalanced across the four subjects using a Latin square.

Hypoxia was induced by having subjects breathe an appropriate gas mixture through an aviator’s oxygen mask. Mixtures having oxygen concentrations of 21.0, 18.1, 15.6 and 13.3 % were used to simulate altitudes of 0, 1200, 2400 and 3700 m, respectively. All were blended from bottled air, oxygen and nitrogen no more than 30 minutes before the session in which they were used and stored in 100 l Douglas bags until needed. Testing in each session did not begin until the subject had breathed the relevant gas mixture for at least 15 minutes and their blood oxygen saturation level, which was monitored throughout via non-invasive pulse oximetry, had stabilised at an altitude-appropriate level. The saturation levels at which individual subjects stabilised as a function of altitude are shown in Table 1. An interval of at least one hour separated all pairs of consecutive sessions. No subject experienced hypoxia for more than two sessions on any day.

Thresholds were measured using a two-interval forced-choice task combined with the two-down one-up adaptive procedure described by Levitt. Each measurement involved the presentation of about 40 trials during each of which the subject's task was to determine which of two intervals coincided with the presentation of a brief auditory signal. The occurrence of each interval was signalled by the illumination of a light-emitting diode placed directly in front of the subject. Each interval was 200 ms in duration and the two were separated by 600 ms. On the initial trial the signal level was set well above the subject’s expected threshold and on subsequent trials it was adjusted following a two-down one-up rule in steps of 5 dB until the third reversal, then 2 dB until the eleventh. Threshold was defined as the average of the signal levels associated with the last eight reversals.

All thresholds were measured with the subject seated at the centre of a 3 x 3 m, sound-attenuated chamber, the background noise level within which was less than 10 dB SPL in all 1/3-octave bands with centre frequencies from 0.5 to 16.0 kHz. Two-hundred millisecond pure-tone pulses, incorporating 10 ms rise and fall times, were generated by a PC-controlled digital signal processing system (Tucker-Davis Technologies System II) and presented to the subject via headphones (Sennheiser ). The at-eardrum sound levels produced by this system were measured by a sound level meter (Brüel and Kjær) with a 1-inch microphone (Brüel and Kjær) in an artificial ear (Brüel and Kjær ). A button box interfaced with the controlling PC provided the subject with a means of initiating each trial and indicating the interval during which he/she thought the signal had been presented.

When screening subjects prior to testing, their thresholds were measured first for the left ear and then for the right. For each ear, thresholds were measured twice for each frequency, following an ascending then a descending order, and the averages were compared with the relevant norms. During testing, however, thresholds were measured for one ear only. This ear was the left for two of the subjects and the right for the others. Thresholds were measured three times for each frequency and the median was taken as a representative value. In our experience the median has proven to be a more reliable measure than the mean with respect to threshold measurements of this type. Each experimental session contained three blocks of threshold measurements, in each of which one measurement was made at each frequency in a random order. About 45 minutes were required to make all 18 measurements.

All procedures followed in this study were approved by the Australian Defence Force Medical Ethics Committee.

Results

Thresholds averaged across all four subjects are plotted as a function of frequency and simulated altitude in Figure 1. At an altitude of 0 m (solid circles in the figure) thresholds increased with increasing frequency in a pattern generally consistent with that for published norms. Threshold values, however, were somewhat lower than the relevant norms for all frequencies other than 1 kHz. This discrepancy may reflect the use of different psychophysical and/or calibration procedures in our and the norm-defining studies. At other altitudes thresholds varied with frequency in a similar manner and for any given frequency, data points for all four altitudes tended to overlap. There was no apparent effect of altitude on sensitivity at any frequency. This was confirmed by an analysis of variance, incorporating a Greenhouse-Geisser correction, which revealed a significant effect of frequency (F(1.4,4.19)=14.51, p=.015) but no effect of altitude (F(1.56,4.67)=.56, p=.563) or frequency-by-altitude interaction (F(2.29,6.87)=1.97, p=.21). Estimates of omega-squared for the frequency and altitude treatments were calculated to be .287 and -.014, respectively. The negative value for altitude suggests that this treatment has a negligible effect (14). A power analysis following the procedures outlined by Keppel  revealed the presence of sufficient power (³ .8) to reliably detect an altitude effect of 4-5 dB.

Discussion

The results of this study indicate that auditory sensitivity for frequencies up to 16 kHz is unaffected by hypoxia induced by exposure to simulated altitudes as high as 3700 m.

As discussed above, studies that have used AER techniques to examine hypoxia's effects on the auditory system have found that the latencies of various waves in human AERs increase significantly with increasing hypoxia level. These latency increases have been interpreted as evidence of decreased auditory sensitivity, as at constant levels of hypoxia they are associated with reduced stimulus levels. Carlile et al. have suggested that the discrepancy between these studies and those in which hypoxia was observed to have no effect on auditory thresholds  resulted from differences in pre-measurement exposure to hypoxia. Whereas subjects in Burkett and Perrin's study were exposed to each simulated altitude for only 5 minutes before threshold measurement began, Carlile et al. found that latency increases do not develop until subjects have experienced hypoxia for at least 20 minutes. In the study described here, however, threshold measurement did not begin until subjects had been exposed to each simulated altitude for at least 15 minutes and required an additional 40-50 minutes to complete. The duration of exposure to hypoxia prior to threshold measurement, therefore, coincided with the range associated with AER latency increases and differences with respect to that variable cannot account for differences between the findings of our and the AER studies.

An alternative explanation of the discrepancy between previous psychophysical and AER studies is that latency increases observed in the AER studies resulted from a hypoxia-induced reduction in sensitivity to frequencies greater than the 8 kHz maximum tested in the psychophysical studies. That possibility prompted us to extend the range of frequencies over which hypoxia's effects on sensitivity have been examined to cover frequencies up to 16 kHz. The fact that thresholds for all frequencies tested in this study, which included five in the range from 8 to 16 kHz, were found to be unaffected by hypoxia, indicates that this explanation also is untenable.

A more likely explanation of the discrepancy between these studies is that AER wave latency does not always provide a valid measure of auditory sensitivity. As noted above, authors who have interpreted wave latency increases as evidence of decreased auditory sensitivity have done so on the basis of the well-established association between wave latency and stimulus level. The existence of this association, however, does not justify the use of wave latency as a sensitivity measure. A variable such as level of hypoxia could affect wave latency via a mechanism that has no effect on auditory sensitivity. A prime candidate for such a mechanism, considered in two of the above reports, involves the time required to transmit information from the auditory receptor to the site of wave generation. A disassociation between AER wave latency and auditory sensitivity with respect to hypoxia 's effects is suggested by the fact that hypoxia exposures capable of inducing latency increases have been found in several studies to have no effect on wave amplitude.

The absence of a hypoxia effect as demonstrated in this study is encouraging with respect to the continued development and use of auditory displays in environments such as aircraft cockpits where operators are in many cases routinely exposed to low levels of hypoxia and at risk of exposure to much higher levels. As hypoxia levels normally experienced by operators of modern aircraft are lower than the maximum in this study, our results indicate that no hypoxia-induced loss of auditory sensitivity will normally be experienced in that environment and the auditory modality should provide an uncompromised channel for information transfer. It is possible, however, that this will not remain the case where operators are exposed to higher levels. Severe levels of hypoxia (blood oxygen saturation levels of less than 65%) in non-human subjects have been found to induce significant reductions in cochlear sensitivity, as measured by indices such as evoked-otoacoustic-emission amplitude (24) and AER threshold, and it seems reasonable to expect that audition will eventually be compromised as hypoxia is increased. Arguing against this, however, is the study by Bagby, in which exposure to hypoxia levels high enough to result in loss of consciousness was found to have no effect on auditory thresholds.