The effects of hearing protection devices on spatial awareness in complex listening environments

Abstract

Hearing protection devices (HPDs) remain the first line of defense against hazardous noise exposure and noise-induced hearing loss (NIHL). Despite the increased awareness of NIHL as a major occupational health hazard, implementation of effective hearing protection interventions remains challenging in at-risk occupational groups including those in public safety that provide fire, emergency medical, or law enforcement services. A reduction of situational awareness has been reported as a primary barrier to including HPDs as routine personal protective equipment. This study examined the effects of hearing protection and simulated NIHL on spatial awareness in ten normal hearing subjects. In a sound-attenuating booth and using a head-orientation tracker, speech intelligibility and localization accuracy were collected from these subjects under multiple listening conditions. Results demonstrate that the use of HPDs disrupts spatial hearing as expected, specifically localization performance and monitoring of speech signals. There was a significant interaction between hemifield and signal-to-noise ratio (SNR), with speech intelligibility significantly affected when signals were presented from behind at reduced SNR. Results also suggest greater spatial hearing disruption using over-the-ear HPDs when compared to the removal of high frequency cues typically associated with NIHL through low-pass filtering. These results are consistent with reduced situational awareness as a self-reported barrier to routine HPD use, and was evidenced in our study by decreased ability to make accurate decisions about source location in a controlled dual-task localization experiment.

Fig 1. Illustration of the experimental setup and listening conditions.

A) Participants were seated in the center of a 24-speaker array. The 12 speakers used to present the stimuli are indicated by their location in azimuth at a speaker-to-speaker distance of 30°. A single talker was presented from each quadrant for every trial, indicated by the grey dotted lines. B) The 6 listening conditions (normal hearing, with low-pass filtering, and with over-the-ear (OTE) or in-the-ear (ITE) HPDs, see section E) and the low-pass filtering applied to simulate NIHL are visualized.

Fig 2. Analysis of localization response using head tracking.

A sample target stimulus from the CRM corpus is presented with its associated waveform. The target call sign used was “hopper” and the associated color/number to be repeated are indicated in green. The upper panel shows the head position in azimuth (degrees) and lower panel shows the velocity of the head movement (degrees/second). RT is calculated as the time between the onset of the stimulus and initiation of the head movement. The endpoint of the head movement is set once its velocity profile falls below 20°/s (grey shaded bar).

Fig 3. Listeners’ speech intelligibility performance is shown as overall percent correct identification of color and number spoken by the target talker, and localization performance is presented in MAE for all listening conditions.

Here, lower MAE values indicate better localization performance. Performance for targets presented to the front is shown in blue and to the back in red. Shaded areas show standard error of the mean.

Fig 4. Mean response promptness (inverse of RT) across subjects for each listening condition, SNRs and font/back hemifields (blue/red, respectively).

Shaded areas show standard error of the mean. Participants with measurable RTs are indicated by circles. Those participants whose average RTs across listening conditions exceeded a cutoff of 1.5 s are indicated by squares.

Fig 5. Representative examples of three listeners’ strategies highlight a tradeoff between speech intelligibility, localization, and response promptness.

(A) accuracy for color and number identification, (B) MAE (localization performance), and (C) promptness (inverse of RT) are shown for participant 1 (O), 2 (□), and 3 (Δ). Behavior in +10 dB SNR is shown in the left column and +5 dB SNR in the right column of the figure.