(PP1313) Self-Reported Task Load as a Function of Signal-to-Noise Ratio and Perceived Loudness

Abstract: Speech comprehension is an intricate process that expands far beyond an audiometric assessment, and is further compounded by competing background noise. It is well established that speech-in-noise performance declines as a function of declining signal-to-noise ratios (SNR). However, little is known if performance is also influenced as a function of perceived loudness. This study investigated self-reported task load for a speech-in-noise task that was delivered at four different SNRs at three normalized loudness levels to normal hearing listeners. Results suggest that task load such as frustration, effort, and mental demand are dependent on both SNR and perceived loudness.

Summary: Rationale/Purpose: The allocation of cognitive resources such as attention and/or working memory (i.e., listening effort) plays a role in understanding speech in the presence of background noise (Pichora-Fuller et al., 2016). As listening conditions become more challenging, greater cognitive resources are required to maintain the same performance (Peelle, 2018). This scenario is further compounded for individuals with cochlear hearing loss, and the increased cognitive load is suggested by some to contribute to the onset of dementia (Lin et al., 2013). Patients with cochlear hearing loss also experience an abnormal growth of loudness due to a reduced dynamic range. It is currently unknown if cognitive load is dependent on perceived stimulus loudness, which could have large implications for individuals who have reduced dynamic ranges due to cochlear hearing loss or hyperacusis. We hypothesized that greater cognitive load would be reported when a speech in noise task was presented in soft or loud conditions compared to a comfortably loud (i.e., medium) condition.

Design: As a preliminary investigation, 24 adults with normal audiometric thresholds were recruited to participate. Three trials of categorical loudness scaling (CLS) was performed using an 11-point scale and multi-talker babble presented binaurally from 30-90 dB HL in 5 dB steps. The average intensity for each individual’s ranking of soft, medium and loud was used to normalize three speech-in-noise loudness conditions. The connected sentences test (CST) was presented in quiet, +5-, 0- and -5 dB SNR for each of the three-loudness conditions. After each condition, participants completed the NASA Task Load Index to
report the amount of effort, frustration, perceived performance, and mental demand they applied to the preceding set of sentences. Each condition consisted of 10 sentences and were scored on the percent correct response to 25 target words.

Results: Actual performance and perceived performance on the CST significantly declined as a function of decreasing SNRs for all loudness conditions. Self-reported effort, frustration and mental demand all significantly increased as a function of declining SNRs for all loudness conditions. In the +5 dB SNR soft loudness condition, effort, frustration, and mental demand was significantly greater and CST score was significantly lower than the medium or loud stimulus conditions of the same SNR. There was no significant difference between soft, medium, or loud conditions for any of the task load variables and CST scores were not significantly different when sentences were presented at 0 dB SNR. In the -5 dB SNR loud stimulus condition, listeners reported significantly more effort and CST scores were significantly lower than the medium or soft condition presented at the same SNR.

Conclusions: In addition to SNR, perceived stimulus loudness may also influence the amount of cognitive load endured while listening to speech in noise. Further investigation is needed to determine if individuals with a reduced dynamic range due to cochlear hearing loss or hyperacusis show greater cognitive load as a function of perceived loudness.