Novel sounds as a psychophysiological measure of listening effort in older listeners with and without hearing loss

Status:: #status/reviewed

Metadata

Item Type: Article #journalArticle
Authors: Sibylle Bertoli, Daniel Bodmer
Date: 05/2014
Date Added: 2023-07-02
URL: https://linkinghub.elsevier.com/retrieve/pii/S1388245713011280
DOI: 10.1016/j.clinph.2013.09.045
Cite key: bertoliNovelSoundsPsychophysiological2014
Topics: misc soruces for topics/ideas, ALC lab journal club
Tags: #Auditory
PDF Attachments
- Bertoli_Bodmer_2014_Novel sounds as a psychophysiological measure of listening effort in older.pdf

other Files

Zotero links

Article Review Table: (Bertoli & Bodmer, 2014)


Background/Purpose::	Background- • aging is associated with changes to the auditory system, including hearing loss, lowed processing speed, and reduced selective attending to a target speaker (inhibitory defect). All of these changes may demand more effort to follow conversations • existing audiometric measures (ex: pure-tone diagrams (?)) do not measure the effort/stress involved in subject hearing performance • Auditory event-related potentials (ERPs) have not been explicitly used to estimate listening effort, despite their high temporal resolution. • an objective measure of listing effort (like using auditory ERPs could aid in audiological evaluations and be used to explore differential outcomes of hearing treatments (i.e. hearing aids) between people with similar hearing damage. Purpose- • examine if task-irrelevant novel sounds played during auditory task can provide info on degree of listing effort.
Participants/samples::	• 18 older normal hearing adults (62-78years old) and 26 older adults (60-86 years old) with hearing loss. both groups recruited from local institution “Senior University”. same subjects for both EXPs(?)
Measures::	• ERPs • German version of the SPIN test- a test to measure speech understanding, sentences with low predictability are presented to participants in a background noise of 60 dB SPL. If a participant's pure-tone average for 0.25, 0.5 and 1 kHz is higher than 30 dB HL, the noise level is increased. The participants then repeat the final words of the sentences, which have low predictability from the preceding context. Two Lists of 15 sentences are administered prior to the experiment to determine the individual signal-to-noise ratio (SNR) for which 50% of the final words are correctly identified. Hypothesis- • Novelty P3 and LPP amp would change with task difficulty • Novelty P3 and LPP amplitudes would decrese if there was enough spatial separation to establish two channels • Novelty P3 and LPP amp would be enhanced with task difficulty if separation failed to allow for two perceptual channels
Procedures::	General • ERPs recorded for novel sounds during a frequency discrimination task (EXp1) and a speech-perception-in-noise (SPIN) test (EXP2). EXP1 • individual subjects frequency discrimination threshold for 1000Hz pure tones was established • subjects were run through a novelty oddball design with frequency discrimination task of three difficulty levels. • three types of stim: standards (80% n = 400) , targets ( 14% n = 70) and novels ( 6%, n = 30). target would consist of a 1200Hz (easy) tone or the frequency of individual discrimination threshold ( medium/hard). stim delivered with a fixed ISI of 700ms (easy/medium) or a randomized value between 350-650ms (hard) . [note this diff in ISIS for diff conditions is why RT’s where not looked at] • standards/targets delivered to the right ear, novels to the left ear ( to favor channel separation) • subjects hit a button when they heard the target. hit/false alarm rates were calculated, where hits = responding within 100-1000ms response window. FA = response to standard/novel in the 100-1000ms response window EXP2 • German version of SPIN test given to subjects • listeners had to repeat the last word of a sentence. sentences differed based on how predictable the last word would be based on the contexts of the rest of the sentence , either being low predictable (LP) or high predictable (HP). Only LP’s were used • task difficulty manipulated by altering signal to noise ratio (SNR) to three levels (easy, medium, hard) • behavior recoded by having subjects write down the last word after each sentence following a fixed 5s silent retention interval b/w sentences. • %correct was calculated per condition
Analysis::	• grand mean average waveforms generated for resposnes to novels and grouped by hearing condition (normal vs impaired) per condition in each EXP. • N1, N2 and Novelty P3 peak amplitudes and latencies were measured using a three-factor repeated-measures ANOVA. with hearing as the between-subject factor, difficulty and electrode site as the within-subject factors. Significance was determined with Bonferroni’s post hoc measure (alpha level < 0.05). • Significance was determined using a Greenhouse-Geiser epsilon (e) correction with Bonferroni's post hoc measures and gp2 partial eta-squared coefficient for effect sizes. • Correlational analyses were also performed to analyze the relationship between the ERPs and behavioral data.
Results::	• latency windows for peak measures determined by grand average waveforms : N1 75–150 ms (Experiment 1) and 75–250 ms (Experiment 2), N2 150–250 ms, Novelty P3 200–500 ms General- • Novelty P3 amplitudes and late positive potentials increased with task difficulty • N1 and N2 amplitudes decreased with task difficulty • hearing loss subjects had sig bigger LPP amps in the easy condition of SPIN test compared to normal -hearing subjects • lack of correlations b/w ERP and behavioral data suggests it is not a simple equivalent of behavioral performance
Discussion/Limits::	Conclusion- • suggests that LPP amp is the most sensitive component for assessing listening effort reflecting subject arousal compared to N1 due to the more consistent direction of amp change across conditions • suggests that further research is needed to determine the optimal parameters for measuring listening effort using novel sounds as an objective psychophysiological measure, and that the LPP component of ERP’s shows promise in capturing the amount of stress experienced by a listener during a task Limitations-
Relevance::	• shows potential of recording ERP’s (specifically LPP) to novel sounds during hearing tests as an easy objective physiological measure of listening effort. • specifically points to amplitude of LPP as most useful index of listening effort. it specifically calls it reflective of general arousal and thus emotional processing (rather then cognitive processing). this might be useful for justifying our focus on freq band power?

Notes

zettle notes

my questions

how is "predictability" measured in the context of the SPIN test? specifically because the end words were supposed be low or high predictability based on the rest of the sentence.? pdf
- i looked up the OG paper. for the English version, predictability was set by audiologists based on number of “pointer” words.
- in the German version (the one used here) determined predictability by first having a sample of people (n28) complete a set of phrases by adding monosyllabic nouns to the end. testing 630 phrases this way, predictability was determined by the percentage of correct identifications

my comments

highlights

This text discusses the use of novel sounds as a measure of listening effort in older individuals with and without hearing loss.
The study found that amplitudes of the Novelty P3 and late positive potential (LPP) increased with increasing task difficulty, and subjects with hearing loss had larger LPP amplitudes compared to those with normal hearing.
The text suggests that ERPs could be a valuable tool for studying the complex processes underlying hearing and speech comprehension.
The text highlights the need for objective measures of listening effort in addition to traditional audiometric assessments.
LPP is more relevant for emotion then for cog? it has been suggested that the amplitude of LPP is reflective of general arousal of the subject, being more relevant to affective processing then it is to cognitive processing as it generally is larger for emotional stimuli compared to neutral stim (highlight).

authors spit balling why there was no sig difference in p3 between the medium and hard condition-

- they suggest that it might mean that listing effort was the same for both conditions despite the differences in behavior. this idea has seen support with other measures of orienting response (skin conductance, pupillometer)

listing effort might be the same for tasks with clearly diff behavioral performance, specifically for hard tasks. some studies suggest a non-liner relationship b/w listening effort and task difficulty
the parms selected for the medium and hard conditions might not have been optimal

misc. notes

i found out later that Matt and Ali have already read this, as they cited it in a 2021 paper “Wisniewski, M. G., Zakrzewski, A. C., Bell, D. R., & Wheeler, M. (2021). Eeg power spectral dynamics associated with listening in adverse conditions. Psychophysiology, 58(9). https://doi.org/10.1111/psyp.13877”