TY - JOUR
AU - BSc, Ingrid Smith,
AB - ABSTRACT Strategies to combat auditory overload were studied. Normal-hearing males were tested in a sound isolated room in a mock-up of a military land vehicle. Two tasks were presented concurrently, in quiet and vehicle noise. For Task 1 dichotic phrases were delivered over a communications headset. Participants encoded only those beginning with a preassigned call sign (Baron or Charlie). For Task 2, they agreed or disagreed with simple equations presented either over loudspeakers, as text on the laptop monitor, in both the audio and the visual modalities, or not at all. Accuracy was significantly better by 20% on Task 2 when the equations were presented visually or audiovisually. Scores were at least 78% correct for dichotic phrases presented over the headset, with a right ear advantage of 7%, given the 5 dB speech-to-noise ratio. The left ear disadvantage was particularly apparent in noise, where the interaural difference was 12%. Relatively lower scores in the left ear, in noise, were observed for phrases beginning with Charlie. These findings underscore the benefit of delivering higher priority communications to the dominant ear, the importance of selecting speech sounds that are resilient to noise masking, and the advantage of using text in cases of degraded audio. INTRODUCTION This study is the second in a series to evaluate aids for auditory overload during military operations.1 Auditory overload, the problem of competing messages, has been the subject of numerous studies dating from the early 1950s.2,3,4,5 Although listeners can understand a single talker among many (the “cocktail party effect”), they have difficulty understanding competing talkers. Intelligibility improves if the talkers are differentiated by such features as their gender, vocal intensity, spatial location, and importance.6,7,8,9,10 This scenario is particularly challenging in military operations where radio operators may be tasked with monitoring, responding to and relaying messages arriving concurrently over right and left earphones of one or more communications headsets, from loudspeakers in a crew compartment, and by direct dialogue with team members. The task may be further complicated by the masking effects of background noise from vehicle engines, weapon fire, and nonattended conversation.1,11 Finomore et al11 investigated the benefits of a multimodal communications suite (MMC) incorporating three-dimensional (3D) audio, a repeat function, and text-based messaging (chat), along with the standard radio. MMC was compared with monaural radio communications, 3D audio, and chat for detecting and replying to target messages from six different channels, over a 27-minute session. Listeners responded by pressing a push-to-talk button on the correct channel and repeating back the message. Target message detection and verbal response accuracy were significantly greater with MMC and chat than with 3D audio and radio. Detection response times were faster for 3D audio and radio. Using a different strategy, Abel et al1 assessed the benefit of visual cues to direct attention to communication channels. Concurrent messages were played over right and left earphones of a noise attenuating communication headset and loudspeaker array, in quiet or noise. Listeners encoded keywords contained only in messages beginning with a preassigned call sign. Although the at-ear speech-to-noise ratios (SNR) for messages presented over the headset and loudspeakers were the same, messages from the loudspeakers were more difficult to understand. The explanation was that the headset barrier decreased at-ear levels at the speech frequencies below levels for the noise. Provision of a visual icon on a computer monitor showing the source of the target messages significantly increased the percentage correct for the loudspeaker messages. This result is in line with the finding of Best et al12 that a vision cue signifying the location and time of occurrence of a message can improve its identification. A variable that has not been investigated for military communications is whether one or the other ear is better at disentangling simultaneous messages. The scientific literature shows evidence for right ear dominance.13,14 Kimura13 presented pairs of dichotic (different) digits over headphones to patients with temporal lobe lesions. More digits arriving at the right ear than the left were reported correctly regardless of the side of lesion. The outcome was attributed to a right-ear advantage in accessing the speech processing center in the left hemisphere through the dominant crossed neurologic pathway. Patients with confirmed speech processing by the right hemisphere showed a left-ear advantage. Kinsbourne and coworkers15 argued against prewiring in the brain, in favor of attentional bias. The present experiment evaluated the benefits for auditory overload of replacing audio with text (visual) or audiovisual messages for one of several simultaneously active communication channels. The channels included right and left earphones of a communication headset and a loudspeaker array. Research suggests that concurrent presentation of visual and auditory stimuli can result in either facilitation (faster response latencies and greater accuracy) or competition (because of sensory dominance). Sinnett et al16 presented trains of randomly intermixed auditory, visual, and bimodal stimuli separated by brief gaps. Participants responded to either visual or auditory targets. Response latencies to unimodal auditory and visual stimuli were no different, showing that they were equally salient. Responses were significantly slower to the bimodal stimuli on the auditory response trials (competition) than on the visual response trials, and faster to the bimodal than the unimodal visual stimuli on the visual response trials (facilitation). Results reported above led us to hypothesize that participants in the present study would benefit from audiovisual presentation of messages presented over a communication channel (facilitation). Audio messages presented on other channels would be neglected during concurrent visual or audiovisual presentations because of visual dominance (competition).16 We predicted that accuracy would be higher for messages presented over the right headphone than the left because of right ear dominance,13 particularly for right-handed participants.17 METHODS AND MATERIALS Participants Nineteen males, aged 23–59 years (median 33 years) participated. The study was restricted to males to minimize possible confounding of gender differences between the listeners (the participants) and talkers (recorded male voices).18 All were experienced at responding to concurrent audio messages. Fifteen were members of the Canadian Armed Forces (CAF) and had used combat radios. Four were civilians who had participated in similar auditory experiments. Participants were screened for ear disorders, including a history of wax buildup, middle ear disease, and tinnitus. Bilateral air conduction hearing thresholds measured using Békésy audiometry19 were not greater than 25 decibels (dB) hearing level (HL) from 0.25 to 8 kHz, except for one individual whose 2-kHz threshold in 1 ear was 30 dB HL. None had an interaural difference greater than 15 dB, minimizing a bias favoring one or the other ear. Participants were screened for claustrophobia as they would be confined in a sound proof room, the use of medications that might affect sustained attention over a 2-hour period, or the need for eyeglasses to read instructions on a laptop monitor as these might affect the fit of the headset that would be worn.20 Proficiency in speaking English was required to eliminate the effect of nonfluency on the understanding of the speech materials that would be used.21,22 All obtained a score of at least 85% on an adapted timed (20 minutes) paper and pencil version of a test offered by The Skylark School of English.23 All but three reported that they were right-handed. Test Facility The test facility has been described previously.1 Participants were tested individually while seated in front of a laptop computer in a mock-up of a CAF land vehicle, the Bison Command, Control, Communications, and Intelligence mobile command post (C3I MCP), in our center's Noise Simulation Facility.24 The Noise Simulation Facility is a semi-reverberant room, 10.55 m (L) by 6.10 m (W) by 3.05 m (H). An array of speakers comprising four low-frequency drivers (Bass Tech 7; ServoDrive, Glenview, Illinois), eight mid-frequency drivers (Gane G218; Equity Sound Investments, Bloomington, Indiana), and four high-frequency drivers (DMC 1152A; Electro-Voice, Burnsville, Minnesota) occupies the shorter rear wall. These are powered by 14 amplifiers (8 stereo model 4B and 6 mono model 7B; Bryston, Peterborough, Ontario, Canada). This array allows the acoustic simulation of a wide range of CAF environments, in terms of level, energy spectrum and phase, and can produce levels in excess of 130 decibels, sound pressure level (dB SPL). The ambient noise is 28 dB SPL at the speech frequencies (i.e., 0.5, 1, and 2 kHz). Within the mock-up, a set of four loudspeakers (EVID 3.2t; Electro-Voice, Burnsville, Minnesota), suspended from the framework, surrounded the participant at azimuth angles of approximately 45°, on either side of the midline, front and back, at a distance of 1 m, level with the top of the head. Tasks Participants engaged in two equally important concurrent tasks. The first (Task 1) involved listening and responding to dichotic (i.e., pairs of different) phrases presented over right and left earphones of a headset. The second (Task 2) required agreement or disagreement with mathematical equations occurring randomly during the delivery of the phrases. The equations were presented (1) over the four-speaker array surrounding the participant (audio modality), (2) as text on the laptop monitor (visual modality), (3) in both the audio and the visual modalities simultaneously (audiovisual modality), or (4) not at all (control condition). They were performed either in quiet or during playback of a digital recording of the noise heard within a Bison C3I MCP vehicle driving along a highway. In all, there were eight listening conditions (four presentation modes for Task 2 by two levels of the background). The order of the absence or presence of the noise, and within each background, the order of the four Task 2 presentation modes were counterbalanced across participants. Task 1 In each of the eight conditions, participants were presented 60 dichotic pairs of phrases with simultaneous onsets. The phrases differed within and across pairs and were chosen randomly for each participant and listening condition. They were taken from the Coordinate Response Measure, a nonstandardized speech corpus for multitalker communications research, that measures speech intelligibility in military environments.25 Each phrase consists of a recording of a talker speaking a call sign followed by a color-number combination within a carrier phrase, e.g., “Baron go to Blue Five now.” In all, there are 256 phrases, made up of combinations of eight call signs (Charlie, Ringo, Laker, Hopper, Arrow, Tiger, Eagle, and Baron), four colors (blue, red, white, and green), and eight numbers (1, 2, 3, 4, 5, 6, 7, and 8), recorded by four male and four female talkers. In the present study, the phrases spoken by two of the males were used for the right and left earphones, respectively, to maximize the distinction between these communication channels. These were stored digitally on the hard drive of a desktop computer, with the carrier words “go to” and “now” removed. Phrases beginning with two of the call signs (Baron and Charlie) were used. Ten of the participants were assigned Baron and nine were assigned Charlie as their target call sign. The target call sign began a phrase on 16 of the 60 trials (27% probability of occurrence), eight randomly to each of the right and left ears, with the restriction that it could not occur at the same time in both ears, and never twice in succession in the same ear. Across the 16 target trials the choice of color-number pairings was random, with the restriction that each could only occur once. For the remaining 44 nontarget trials the color-number pairings were again random, and none could occur at the same time in both ears. Speech materials were presented at an at-ear level of 70 dB SPL over a noise attenuating communications headset (Racal Slimgard II RA108/1148; Esterline Technologies, Bellevue, Washington, DC) used by military personnel operating the Bison C3I MCP. The noise recording was played over the loudspeaker array in the test room (outside the Bison C3I MCP mock-up) at an at-ear level under the headset of 65 dBA (decibels, A-weighted). The at-ear level measured using an acoustic test fixture without the headset worn was 95 dBA, which is about 5 dB lower than the level inside a light armored vehicle driving along a highway.26 This level was chosen so that with a comfortable listening level for the speech of 70 dB SPL, the at-ear SNR under the headset would be 5 dB. The 5 dB SNR results in speech understanding scores of 60%–80%.27Figure 1 shows the right and left at-ear spectra of the vehicle noise without and with the headset fitted to the acoustic test fixture. FIGURE 1 View largeDownload slide At-ear energy spectra for the vehicle noise (NH-no headset worn and H-headset worn), separately for right and left ears. FIGURE 1 View largeDownload slide At-ear energy spectra for the vehicle noise (NH-no headset worn and H-headset worn), separately for right and left ears. Participants responded whenever a phrase presented to either ear started with their call sign. They pressed three keys on a standard laptop computer keyboard, indicating the perceived ear (one of two labeled keys), the color (one of four labeled keys), and the number (one of eight labeled keys), respectively. The keys for each were located centrally on different rows of the keyboard. Each phrase took 3 seconds to present. Rate of presentation was one pair of phrases every 6 seconds, allowing 3 seconds for the response. The total duration for the block of 60 trials was 10 minutes, including the time to describe the upcoming condition. Task 2 Participants were required to agree or disagree with 16 arithmetic equations. These were presented randomly, half correct and half incorrect, on the 44 trials when the phrases did not begin with the participant's call sign. Half of each of the correct and incorrect equations involved addition (e.g., 4 + 1 = 5) and half subtraction (e.g., 4 − 1 = 3) of single digits whose result was a single digit. Digits were chosen randomly for each participant and listening condition. A response, depressing the “correct” or “incorrect” key on the laptop keyboard, was required within 2 seconds of the termination of the equation. These keys were located above those used for the phrases. Late responses were counted as wrong. For the audio modality, recordings of the equations were made by a male colleague. The at-ear level was 70 dB SPL, the same as that for the headset phrases. For the visual modality, the equations appeared centered on the participant's laptop computer monitor. Audio and/or text began at the same time as the presentation of the phrases in Task 1. The text presentation was approximately one-third second longer than the audio to allow sufficient time to be read. The extra time allotted was based on the investigators' trial and error experience. Procedure The protocol was approved by the ethics committee of our research center. Volunteers were asked to sign a consent form that described the experiment before participation. At the start of the session, they were fitted with the headset by a trained technician. They then listened to a sample of the vehicle noise and listened to and/or visualized and responded to the stimulus materials for Tasks 1 and 2. Feedback was given for the practice but not the experimental trials. Before each trial block, participants were told the background and Task 2 mode. In the case of Task 2 visual or audiovisual presentations, they were instructed to focus on the monitor to read the equations and then shift their gaze to the keyboard to respond. Either one or both hands could be used to respond. Short breaks separated the eight conditions. The duration of the session, including the time for instructions, practice, breaks, and debriefing, was 2 hours. RESULTS The dataset for each of the four Task 2 presentation modes (none, audio only, visual only, and audiovisual), in combination with the quiet and noise backgrounds, consisted of (1) the percentage of correct identifications of all of the ear, color, and number for the right and left ears for Task 1; (2) the percentage of correct identifications of the ear, color, number, and color-number combination taken separately; (3) the percentage of false alarms, i.e., responding when neither member of a pair of phrases started with the participant's call sign; (4) the percentage of misses, i.e., failure to respond to a target phrase; and (5) the percentage of correct responses for Task 2. Communication Headset Messaging (Task 1) Table I shows the mean percentage, averaged across participants, of correctly encoded target phrases (i.e., all of correct ear, color, and number) for right and left ears, for combinations of the call sign (Baron or Charlie), background (quiet or vehicle noise), and Task 2 presentation modes (N—none, A—audio, V—visual, or AV—audiovisual). The outcomes ranged from 77.8% to 100.0%, with standard deviations of 0% to 24.8%. A four-way repeated measures analysis of variance (ANOVA) was applied to these data.28 Mauchley's test indicated that the variances of the differences between all possible levels of each of the independent variables were not significantly different. This finding, along with the similar sample sizes, ensured that the assumptions of the ANOVA had not been violated.29 The ANOVA showed that there were significant effects of ear (F1,17 = 14.06; p < 0.002), ear by call sign (F1,17 = 6.18; p < 0.02), and ear by background (F1,17 = 5.67; p < 0.03). The difference because of ear was 7%, favoring the right ear. The ear by call sign interaction indicated that the poorer performance observed for the left ear compared with the right occurred mainly when phrases began with Charlie (85.07% vs. 97.22% for a difference of 12.2%) and to a lesser extent with Baron (93.13% vs. 95.78% for a difference of 2.6%). Figure 2 shows the at-ear energy spectra, as well as the overall dBA and dB SPL levels for phrases beginning with the two call signs, for the right and left ears. These appear similar. The ear by background interaction indicated that the poorer performance for the left ear compared with the right occurred mainly in noise (86.08% vs. 97.57% for a difference of 11.5%) and to a lesser extent in quiet (92.12% vs. 95.43% for a difference of 3.3%). Post hoc pairwise comparisons using Fisher's LSD test28 showed that while the ear by background interaction was statistically significant (p < 0.05, 1-tailed), the ear by call sign interaction was not. TABLE I Percentage of Correctly Encoded Phrases for Left and Right Ears Call Sign  Background  Task 2 Mode  Correct Phrase  Left Ear  Right Ear  Baronb  Quiet  N  93.8 (10.6)a  90.0 (12.9)  A  95.0 (8.7)  97.5 (5.3)  V  95.0 (10.5)  95.0 (8.7)  AV  96.3 (6.0)  96.3 (8.4)  Noise  N  85.0 (16.5)  100.0 (0.0)  A  91.3 (11.9)  97.5 (5.3)  V  95.0 (6.5)  96.3 (11.9)  AV  93.8 (10.6)  93.8 (8.8)  Charliec  Quiet  N  93.1 (12.7)  94.4 (12.7)  A  87.5 (10.8)  98.6 (4.2)  V  88.9 (19.2)  97.2 (5.5)  AV  87.5 (12.5)  94.4 (9.1)  Noise  N  83.3 (17.7)  97.2 (5.5)  A  80.6 (15.5)  98.6 (4.2)  V  81.9 (19.9)  100.0 (0.0)  AV  77.8 (24.8)  97.2 (5.5)  Call Sign  Background  Task 2 Mode  Correct Phrase  Left Ear  Right Ear  Baronb  Quiet  N  93.8 (10.6)a  90.0 (12.9)  A  95.0 (8.7)  97.5 (5.3)  V  95.0 (10.5)  95.0 (8.7)  AV  96.3 (6.0)  96.3 (8.4)  Noise  N  85.0 (16.5)  100.0 (0.0)  A  91.3 (11.9)  97.5 (5.3)  V  95.0 (6.5)  96.3 (11.9)  AV  93.8 (10.6)  93.8 (8.8)  Charliec  Quiet  N  93.1 (12.7)  94.4 (12.7)  A  87.5 (10.8)  98.6 (4.2)  V  88.9 (19.2)  97.2 (5.5)  AV  87.5 (12.5)  94.4 (9.1)  Noise  N  83.3 (17.7)  97.2 (5.5)  A  80.6 (15.5)  98.6 (4.2)  V  81.9 (19.9)  100.0 (0.0)  AV  77.8 (24.8)  97.2 (5.5)  N, none; A, audio; V, visual; AV, audiovisual. a Mean (SD). b Number of participants assigned call sign Baron = 10. c Number of participants assigned call sign Charlie = 9. View Large TABLE I Percentage of Correctly Encoded Phrases for Left and Right Ears Call Sign  Background  Task 2 Mode  Correct Phrase  Left Ear  Right Ear  Baronb  Quiet  N  93.8 (10.6)a  90.0 (12.9)  A  95.0 (8.7)  97.5 (5.3)  V  95.0 (10.5)  95.0 (8.7)  AV  96.3 (6.0)  96.3 (8.4)  Noise  N  85.0 (16.5)  100.0 (0.0)  A  91.3 (11.9)  97.5 (5.3)  V  95.0 (6.5)  96.3 (11.9)  AV  93.8 (10.6)  93.8 (8.8)  Charliec  Quiet  N  93.1 (12.7)  94.4 (12.7)  A  87.5 (10.8)  98.6 (4.2)  V  88.9 (19.2)  97.2 (5.5)  AV  87.5 (12.5)  94.4 (9.1)  Noise  N  83.3 (17.7)  97.2 (5.5)  A  80.6 (15.5)  98.6 (4.2)  V  81.9 (19.9)  100.0 (0.0)  AV  77.8 (24.8)  97.2 (5.5)  Call Sign  Background  Task 2 Mode  Correct Phrase  Left Ear  Right Ear  Baronb  Quiet  N  93.8 (10.6)a  90.0 (12.9)  A  95.0 (8.7)  97.5 (5.3)  V  95.0 (10.5)  95.0 (8.7)  AV  96.3 (6.0)  96.3 (8.4)  Noise  N  85.0 (16.5)  100.0 (0.0)  A  91.3 (11.9)  97.5 (5.3)  V  95.0 (6.5)  96.3 (11.9)  AV  93.8 (10.6)  93.8 (8.8)  Charliec  Quiet  N  93.1 (12.7)  94.4 (12.7)  A  87.5 (10.8)  98.6 (4.2)  V  88.9 (19.2)  97.2 (5.5)  AV  87.5 (12.5)  94.4 (9.1)  Noise  N  83.3 (17.7)  97.2 (5.5)  A  80.6 (15.5)  98.6 (4.2)  V  81.9 (19.9)  100.0 (0.0)  AV  77.8 (24.8)  97.2 (5.5)  N, none; A, audio; V, visual; AV, audiovisual. a Mean (SD). b Number of participants assigned call sign Baron = 10. c Number of participants assigned call sign Charlie = 9. View Large FIGURE 2 View largeDownload slide At-ear energy spectra for phrases beginning with the call signs Baron and Charlie, separately for right (R) and left (L) ears. FIGURE 2 View largeDownload slide At-ear energy spectra for phrases beginning with the call signs Baron and Charlie, separately for right (R) and left (L) ears. Repeated measures ANOVA were also applied to the mean percentages correct, averaged across participants, for each of ear, color, number, and color-number, taken separately, to evaluate the effects of call sign, ear, background, and Task 2 mode. In each case, there were statistically significant background by ear interactions (p < 0.04 or better), the pattern mirroring the effect found for percentage correct phrases. The results are displayed in Table II, averaged across call sign, which was not significant. Outcomes ranged from 87.5% to 99.3% with standard deviations of 1.7% to 13.0%. For color and color-number, there was also a significant main effect of ear of 6.4% and 7.1%, respectively, favoring the right ear (p < 0.004 or better). Post hoc pairwise comparisons for the ear by background interaction using Fisher's LSD test28 indicated that for each of color, number, and color-number, in noise, the percentage correct for the right ear was significantly greater than that for the left ear by 9.9%, 4.9%, and 11.3% (p < 0.05, 1-tailed), respectively. TABLE II Percentage of Correct Responses for Each of Ear, Color, Number and Color—Number Observed in Quiet and Vehicle Noise for Left and Right Ears Ambient  Message Element  Left Ear  Right Ear  Quiet  Ear  98.8 (2.4)a  97.4 (4.1)  Color  94.6 (7.3)  97.4 (4.0)  Number  97.0 (3.9)  96.4 (4.9)  Color-Number  92.8 (8.4)  95.6 (5.4)  Noise  Ear  95.6 (6.6)  98.5 (3.8)  Color  89.3 (10.6)  99.0 (1.8)  Number  94.5 (7.2)  99.3 (1.7)  Color-Number  87.5 (13.0)  98.5 (3.0)  Ambient  Message Element  Left Ear  Right Ear  Quiet  Ear  98.8 (2.4)a  97.4 (4.1)  Color  94.6 (7.3)  97.4 (4.0)  Number  97.0 (3.9)  96.4 (4.9)  Color-Number  92.8 (8.4)  95.6 (5.4)  Noise  Ear  95.6 (6.6)  98.5 (3.8)  Color  89.3 (10.6)  99.0 (1.8)  Number  94.5 (7.2)  99.3 (1.7)  Color-Number  87.5 (13.0)  98.5 (3.0)  a Mean (SD); N = 19. View Large TABLE II Percentage of Correct Responses for Each of Ear, Color, Number and Color—Number Observed in Quiet and Vehicle Noise for Left and Right Ears Ambient  Message Element  Left Ear  Right Ear  Quiet  Ear  98.8 (2.4)a  97.4 (4.1)  Color  94.6 (7.3)  97.4 (4.0)  Number  97.0 (3.9)  96.4 (4.9)  Color-Number  92.8 (8.4)  95.6 (5.4)  Noise  Ear  95.6 (6.6)  98.5 (3.8)  Color  89.3 (10.6)  99.0 (1.8)  Number  94.5 (7.2)  99.3 (1.7)  Color-Number  87.5 (13.0)  98.5 (3.0)  Ambient  Message Element  Left Ear  Right Ear  Quiet  Ear  98.8 (2.4)a  97.4 (4.1)  Color  94.6 (7.3)  97.4 (4.0)  Number  97.0 (3.9)  96.4 (4.9)  Color-Number  92.8 (8.4)  95.6 (5.4)  Noise  Ear  95.6 (6.6)  98.5 (3.8)  Color  89.3 (10.6)  99.0 (1.8)  Number  94.5 (7.2)  99.3 (1.7)  Color-Number  87.5 (13.0)  98.5 (3.0)  a Mean (SD); N = 19. View Large The percentage of missed target phrases, averaged across Task 2 mode, was at most 3.8% (phrases beginning with the call sign Charlie for the left ear in noise). The percentage of false alarms, responding to phrases that did not begin with the participant's call sign, was 0.2% in quiet and 0.8% in noise, averaged across the two call signs and Task 2 mode options. Effect of Visual/Audiovisual Messaging (Task 2) Participants' ability to agree with correct mathematical equations (correct math) and disagree with incorrect equations (incorrect math) is shown in Table III. The percentages of accurate judgments, averaged across the 19 participants, were calculated for combinations of the background (quiet or noise) and mode of presentation of the equations (A—audio, V—visual, and AV—audiovisual). A three-way ANOVA28 showed that there were statistically significant effects of the background (F1,18 = 8.95; p < 0.008), mode (F2,36 = 67.09; p < 0.001), type of equation (F1,18 = 126.15; p < 0.0001), and mode by type of equation (F2,36 = 60.28; p < 0.001). Although the homogeneity of variance assumption for the ANOVA was not met for mode and the interaction of mode by type of equation, the sample size was sufficiently large to validate the assumptions of the test.29 The results showed a decrement in accuracy because of the background noise of 4.7%. Visual and audiovisual presentations resulted in an increment in the percentage correct relative to the audio presentation of 21.9% (97.37% compared with 75.49%). The mode by type of equation interaction is shown in Figure 3. In the audio modality, participants had relatively greater difficulty agreeing with correct equations (58.55%) than disagreeing with incorrect equations (92.43%). Post hoc pairwise comparisons indicated that this difference of 33.9% was statistically significant (p < 0.01, 2-tailed). Results for the visual and audiovisual modes were not different. TABLE III Percentage Correct Responses for Mathematical Equations Background  Mode  Math  Average  Correct  Incorrect  Quiet  A  66.4 (19.1)a  94.1 (8.7)  80.3 (11.3)  V  97.4 (5.2)  99.3 (2.9)  98.4 (2.8)  AV  98.0 (4.7)  99.3 (2.9)  98.7 (3.3)  Noise  A  50.7 (24.5)  90.8 (12.4)  70.7 (14.7)  V  94.1 (7.6)  98.0 (8.6)  96.1 (7.3)  AV  97.4 (5.2)  95.4 (8.5)  96.4 (4.3)  Background  Mode  Math  Average  Correct  Incorrect  Quiet  A  66.4 (19.1)a  94.1 (8.7)  80.3 (11.3)  V  97.4 (5.2)  99.3 (2.9)  98.4 (2.8)  AV  98.0 (4.7)  99.3 (2.9)  98.7 (3.3)  Noise  A  50.7 (24.5)  90.8 (12.4)  70.7 (14.7)  V  94.1 (7.6)  98.0 (8.6)  96.1 (7.3)  AV  97.4 (5.2)  95.4 (8.5)  96.4 (4.3)  A, audio; V, visual; AV, audiovisual. a Mean (SD); N = 19. View Large TABLE III Percentage Correct Responses for Mathematical Equations Background  Mode  Math  Average  Correct  Incorrect  Quiet  A  66.4 (19.1)a  94.1 (8.7)  80.3 (11.3)  V  97.4 (5.2)  99.3 (2.9)  98.4 (2.8)  AV  98.0 (4.7)  99.3 (2.9)  98.7 (3.3)  Noise  A  50.7 (24.5)  90.8 (12.4)  70.7 (14.7)  V  94.1 (7.6)  98.0 (8.6)  96.1 (7.3)  AV  97.4 (5.2)  95.4 (8.5)  96.4 (4.3)  Background  Mode  Math  Average  Correct  Incorrect  Quiet  A  66.4 (19.1)a  94.1 (8.7)  80.3 (11.3)  V  97.4 (5.2)  99.3 (2.9)  98.4 (2.8)  AV  98.0 (4.7)  99.3 (2.9)  98.7 (3.3)  Noise  A  50.7 (24.5)  90.8 (12.4)  70.7 (14.7)  V  94.1 (7.6)  98.0 (8.6)  96.1 (7.3)  AV  97.4 (5.2)  95.4 (8.5)  96.4 (4.3)  A, audio; V, visual; AV, audiovisual. a Mean (SD); N = 19. View Large FIGURE 3 View largeDownload slide Effect of presentation mode (A—audio, V—video, and AV—audiovisual) on the percentage of correct judgments, separately for correct and incorrect math equations. FIGURE 3 View largeDownload slide Effect of presentation mode (A—audio, V—video, and AV—audiovisual) on the percentage of correct judgments, separately for correct and incorrect math equations. DISCUSSION Participants had relatively little difficulty discriminating between phrases presented simultaneously to the two ears. They achieved mean scores of at least 78% across the various listening conditions. However, accuracy was significantly higher by 7% for the right ear (averaged across backgrounds, call signs, and Task 2 mode). The left ear disadvantage was particularly apparent in noise, averaged across call signs and Task 2 modes, where the interaural difference was 12%. This outcome was statistically significant for correct identifications of each of color, number, and color-number taken separately, and corroborates previous findings by Kimura.13,14 The effect has been attributed to the processing of speech by the dominant left hemisphere of the brain in right-handed individuals. Left-handers with speech processed by the left hemisphere also show this effect, indicating that handedness alone does not predict the outcome.14 In the present study all but three of the participants were right-handed. They, and 11 of the 16 right-handers, showed a right-ear advantage. This finding underscores the possible benefit of delivering higher priority communications to the dominant right ear in situations of auditory overload. Relatively higher scores were achieved in the left ear, in noise, if the call sign was Baron rather than Charlie. The percentage of missed targets was also relatively highest at 3.8% for phrases beginning with Charlie, presented to the left ear in noise, compared with the other combinations of call sign, background, and ear. This outcome was not expected, given that the at-ear energy spectra for phrases beginning with the two call signs were similar for the two ears. One explanation is that voiceless consonants (“ch”) are more susceptible to noise masking than voiced consonants (“b”).30 Voiced consonants involve vibration of the vocal chords. This finding points to the importance of considering the composition of target words when creating cryptic messages for degraded audio conditions. The findings reported above for dichotic messages are limited to an at-ear under the headset SNR of 5 dB and noise with a low-frequency bias. Studies have shown that speech understanding will become progressively worse as the SNR decreases.27 Speech understanding also depends on the genders of the talkers31 and the hearing status of the listeners,21 as well as the spectrum (e.g., low frequency, continuous, or speech) and the time course (continuous or interrupted) of the masker.32 Female voices tend to mask each other more than male voices.31 With respect to Task 2, participants performed significantly better by 20% when the mathematical equations were presented either visually or audiovisually, compared with audio alone. This outcome is in line with the finding of Finomore et al11 that listeners are more accurate when messages are delivered by chat compared with a standard radio in situations of auditory overload. Much of our understanding of the benefit of supplementary visual information derives from studies of either multisensory integration (e.g., Sinnett et al16) or studies of audiovisual speech, in which audio messages are accompanied by simultaneous videos of the talker's utterances.33 In the present study, the results for the visual and audiovisual presentations were no different, suggesting that participants were primarily paying attention to the visual display. There was no evidence of multisensory facilitation. None of the three presentation modes for Task 2, audio, visual, or audiovisual resulted in a decrement in the accurate reporting of the target phrases, compared with the condition in which Task 2 was absent. This outcome could be due to the fact that the equations were only presented on those trials when there were no target phrases or because the two tasks were independent. The latter would support the conclusion that, in situations of auditory overload, presenting information visually for one or more channels over which messages are difficult to understand is a viable option for communication. CONCLUSIONS Male participants had no difficulty understanding and responding to dichotic headset phrases presented at an SNR of 5 dB. A right ear advantage was evident for the headset task, underscoring the benefit of delivering higher priority communications to the dominant right ear. Although the at-ear levels and spectra for phrases beginning with the two call signs were similar, phrases beginning with Charlie were relatively more difficult to understand than Baron, particularly in the left ear in the vehicle noise. This outcome may be due to differences in resilience of the initial consonant to noise masking. The finding points to the importance of speech sound selection for conditions of degraded listening. Participants benefitted from text presentation of the loudspeaker materials in noise in Task 2. Performance of Task 2, whatever the modality of presentation, did not interfere with Task 1, either because the tasks could be managed independently or because the equations for Task 2 only occurred on nontarget trials in Task 1. ACKNOWLEDGMENTS We are indebted to Mr. John Bowen for his help with the management of the equipment and to Mr. Garry Dunn for software development. We are also grateful to the individuals who graciously gave their time to act as participants. This research was supported by the Agility Fund, Defence Research and Development Canada, Toronto. REFERENCES 1. Abel SM, Nakashima A, Smith I Divided listening in noise in a mock-up of a military command post. Mil Med  2012; 177: 436– 43. Google Scholar CrossRef Search ADS PubMed  2. Broadbent DE Listening to one of two synchronous messages. J Exp Psychol  1952; 44: 51– 5. Google Scholar CrossRef Search ADS PubMed  3. Cherry EC Some experiments on the recognition of speech, with one and with two ears. J Acoust Soc Am  1953; 25: 975– 9. Google Scholar CrossRef Search ADS   4. Webster JC, Thompson PO Responding to both of two overlapping messages. J Acoust Soc Am  1954; 26: 396– 402. Google Scholar CrossRef Search ADS   5. Webster JC, Sharpe L Improvements in message reception resulting from “sequencing” competing messages. J Acoust Soc Am  1955; 27: 1194– 8. Google Scholar CrossRef Search ADS   6. Hawley ML, Litovsky RY, Colburn HS Speech intelligibility and localization in a multi-source environment. J Acoust Soc Am  1999; 105: 3426– 48. Google Scholar CrossRef Search ADS   7. Drullman R, Bronkhorst AW Multichannel speech intelligibility and talker recognition using monaural, binaural, and three-dimensional auditory presentation. J Acoust Soc Am  2000; 107: 2224– 35. Google Scholar CrossRef Search ADS PubMed  8. Brungart DS Informational and energetic masking effects in the perception of two simultaneous talkers. J Acoust Soc Am  2001; 109: 1101– 9. Google Scholar CrossRef Search ADS PubMed  9. Arbogast TL, Mason CR, Kidd G Jr The effect of spatial separation on informational and energetic masking of speech. J Acoust Soc Am  2002; 112: 2086– 98. Google Scholar CrossRef Search ADS PubMed  10. Drullman R, Bronkhorst AW Speech perception and talker segregation: effects of level, pitch, and tactile support with multiple simultaneous talkers. J Acoust Soc Am  2004; 116: 3090– 8. Google Scholar CrossRef Search ADS PubMed  11. Finomore V Jr, Popik D, Castle C, Dallman R Effects of a network-centric multi-modal communication tool on a communication monitoring task. In: Proceedings of the Human Factors and Ergonomics Society 54th Annual Meeting , September 2010; 54: 2125– 9. Available at http://pro.sagepub.com; accessed November 17, 2011. Google Scholar CrossRef Search ADS   12. Best V, Ozmeral EJ, Shinn-Cunningham BG Visually-guided attention enhances target identification in a complex auditory scene. J Assoc Res Otolaryngol  2007; 8: 294– 304. Google Scholar CrossRef Search ADS PubMed  13. Kimura D Cerebral dominance and the perception of verbal stimuli. Canad J Psychol  1961; 15: 166– 70. Google Scholar CrossRef Search ADS   14. Kimura D From ear to brain. Brain Cogn  2011; 76: 214– 7. Google Scholar CrossRef Search ADS PubMed  15. Hiscock M, Kinsbourne M Attention and the right-ear advantage: what is the connection? Brain Cogn  2011; 76: 263– 75. Google Scholar CrossRef Search ADS PubMed  16. Sinnett S, Soto-Faraco S, Spence C The co-occurrence of multisensory competition and facilitation. Acta Psychol  2008; 128: 153– 61. Google Scholar CrossRef Search ADS   17. Foundas AL, Corey DM, Hurley MM, Heilman KM Verbal dichotic listening in right and left-handed adults: laterality effects of directed attention. Cortex  2006; 42: 79– 86. Google Scholar CrossRef Search ADS PubMed  18. Markham D, Hazan V The effect of talker- and listener-related factors on intelligibility for a real-world, open-set perception test. J Speech Lang Hear Res  2004; 47: 725– 37. Google Scholar CrossRef Search ADS PubMed  19. Yantis PA Puretone air-conduction testing. In: Handbook of Clinical Audiology , Ed 3, pp 153– 69. Edited by Katz J Baltimore, MD, Williams & Wilkins Press, 1985. 20. Abel SM, Sass-Kortsak A, Kielar A The effect on earmuff attenuation of other safety gear worn in combination. Noise Health  2002; 5: 1– 13. 21. Abel SM, Alberti PW, Haythornthwaite C, Riko K Speech intelligibility in noise: effects of fluency and hearing protector type. J Acoust Soc Am  1982; 71: 708– 15. Google Scholar CrossRef Search ADS PubMed  22. van Wijngaarden SJ, Steeneken HJM, Houtgast T Quantifying the intelligibility of speech in noise for non-native talkers. J Acoust Soc Am  2002; 112: 3004– 13. Google Scholar CrossRef Search ADS PubMed  23. The Skylark School of English Online test of English fluency . Skylark School of English, Msida, Malta. Available at www.skylarkmalta.com; accessed August 8, 2012. 24. Nakashima A, Borland M The Noise Simulation Facility at DRDC Toronto . Technical Report, DRDC Toronto TR 2005-095, Defence R&D Canada—Toronto, October 2005. Available at http://cradpdf.drdc-rddc.gc.ca/PDFS/unc48/p524518.pdf; accessed December 6, 2013. 25. Bolia RS, Nelson WT, Ericson MA, Simpson BD A speech corpus for multitalker communications research. J Acoust Soc Am  2000; 107: 1065– 6. Google Scholar CrossRef Search ADS PubMed  26. Nakashima A, Borland M, Abel SM Measurement of noise and vibration in Canadian Forces armoured vehicles. Ind Health  2007; 45: 318– 27. Google Scholar CrossRef Search ADS PubMed  27. Abel SM, Krever EM, Alberti PW Auditory detection discrimination and speech processing in ageing, noise-sensitive and hearing-impaired listeners. Scand Audiol  1990; 19: 43– 54. Google Scholar CrossRef Search ADS PubMed  28. Daniel WW Biostatistics: A Foundation for Analysis in the Health Sciences . Ed 3, pp 206– 64. New York, Wiley Press, 1983. 29. Howell DC Statistical Methods for Psychology . Ed 4, pp 321– 2. Belmont, CA, Wadsworth, 1997. 30. Dubno JR, Levitt H Predicting consonant confusions from acoustic analysis. J Acoust Soc Am  1981; 69: 249– 61. Google Scholar CrossRef Search ADS PubMed  31. Ericson MA, McKinley RL The intelligibility of multiple talkers separated spatially in noise. In: Binaural and Spatial Hearing in Real and Virtual Environments , pp 701– 24. Edited by Gilkey RH, Anderson TR New Jersey, Erlbaum, 1997. 32. Iyer N, Brungart DS, Simpson BD Effects of periodic masker interruption on the intelligibility of interrupted speech. J Acoust Soc Am  2007; 122: 1693– 701. Google Scholar CrossRef Search ADS PubMed  33. Bernstein JG, Grant KW Auditory and auditory-visual intelligibility of speech in fluctuating maskers for normal-hearing and hearing-impaired listeners. J Acoust Soc Am  2009; 125: 3358– 72. Google Scholar CrossRef Search ADS PubMed  Footnotes 1 This article was presented in part at the 21st International Congress on Acoustics in Montreal, Quebec, June 2–7, 2013. Reprint & Copyright © Association of Military Surgeons of the U.S.
TI - Strategies to Combat Auditory Overload During Vehicular Command and Control
JF - Military Medicine
DO - 10.7205/MILMED-D-13-00556
DA - 2014-09-01
UR - https://www.deepdyve.com/lp/oxford-university-press/strategies-to-combat-auditory-overload-during-vehicular-command-and-XOt0gGPuM7
SP - 1036
EP - 1042
VL - 179
IS - 9
DP - DeepDyve
ER -