ABSTRACT
Introduction Understanding how listeners execute a dual-task paradigm for listening effort would provide a benchmark for future studies and clinical implementations. Therefore, this study aimed to examine the impact of instructions on the prioritization strategy employed by individuals during a dual-task paradigm for listening effort by assessing dual-task interference.
Methods The dual-task paradigm consisted of a primary speech understanding task in different listening conditions and a secondary visual memory task, both performed separately (baseline) and simultaneously (dual-task). Twenty-three normal-hearing participants (mean age: 36.8 years; 14 females) were directed to prioritize the primary speech understanding task in the dual-task condition, whereas another twenty-three (matched for age, gender, and education level) received no specific instructions regarding task priority. Both groups performed the dual-task paradigm twice (mean interval: 14.8 days). Dual-task interference was assessed by plotting the dual-task effect of the primary and secondary task against each other. Participants were classified based on their patterns of interference.
Results The prioritizing group had more participants who achieved stable or better scores on the primary task in the dual-task condition compared to baseline. However, there was considerable variability in the prioritizing strategy employed at the individual level across listening conditions and test moments, regardless the given prioritization instructions.
Conclusion Providing prioritization instructions was insufficient to ensure that an individual will mainly focus on the primary task and will stick to this strategy across listening conditions and test moments. These results raised certain reservations about the current usage of dual-task paradigms for listening effort.
INTRODUCTION
Increased listening effort – defined as the deliberate allocation of mental resources to overcome obstacles in goal pursuit during a listening task (Pichora-Fuller et al., 2016) – is one of the main complaints of individuals with hearing loss and plays a vital role in listeners’ communication and quality of life (Cañete et al., 2023; Pang et al., 2019). One commonly reported approach to measure listening effort is a dual-task paradigm (Gagne et al., 2017). In the context of listening effort, a dual-task paradigm evaluates an individual’s ability to understand speech (primary task) while concurrently performing another task (secondary task). Both tasks are performed separately and concurrently, denoted as baseline and dual-task condition, respectively.
Dual-task paradigms are based on Kahneman’s theory that the available resources to perform a task are limited in capacity (Kahneman, 1973). When the attentional and cognitive resources required for performing the primary and secondary tasks concurrently are lower than the total available resources, both tasks can be handled effectively. However, if the required resources to simultaneously perform the primary and secondary task exceed the listeners’ maximum cognitive capacity, the processing system will prioritize one of the tasks under the dual-task condition. In the context of listening effort, it is traditionally instructed to prioritize the primary speech understanding task (Gagne et al., 2017). One might assume that providing this prioritization instruction will result in stable scores on the primary task in the baseline and dual-task condition. When this assumption is met, the dual-task effect (DTE) of the secondary task (i.e. change in secondary task performance from the baseline condition to the dual-task condition) can serve as a measure of listening effort, calculated through the formula DTE = 100 × [score in dual-task condition - score in baseline condition]/score in baseline condition (Gagne et al., 2017). Negative DTE values imply a decrease in performance in the dual-task relative to the single-task (i.e. dual-task cost), whereas positive DTE values suggest an advancement in performance in the dual-task (i.e. dual-task benefit) (Plummer & Eskes, 2015).
Within the traditional approach, dual-task interference is not taken into account. Dual-task interference encompasses the DTE of both the primary and secondary tasks (Plummer & Eskes, 2015). For example, a dual-task cost in the secondary task (indicating listening effort) may occur with a reciprocal dual-task cost, no change, or a dual-task benefit in the primary task. Due to the prioritization instruction, no change in the primary task is expected. However, a previous study with children has shown that instructions alone were ineffective to ensure that a participant will mainly focus on the primary task (Choi et al., 2008). There is currently a lack of results on adults in this regard. Moreover, the assumption of stable primary speech understanding scores between the baseline and dual-task condition is mostly evaluated on a group level (e.g. Degeest et al., 2022). This approach neglects individual differences as they might be smoothed out through averaging (Ceuleers et al., 2024). Hence, it remains unclear which prioritization strategy individuals employ when their total cognitive capacity is exceeded.
Because of its ecological validity, implementing a dual-task paradigm for measuring listening effort in clinical practice seems valuable, emphasizing the need of reliable individual measurements. Results from cognitive-motor dual-task studies already showed that evaluating dual-task interference can deepen our understanding of individual prioritization strategies as it provides insights into changes in attention allocation as well as dual-task capacity (Plummer & Eskes, 2015; Yogev-Seligmann et al., 2012). Hence, dual-task interference has been suggested as a possible metric to measure listening effort (Degeest et al., 2021; Gagne et al., 2017) though it has received limited attention in audiological research to date.
Therefore, the current study aimed to examine the prioritization strategy employed by individuals during a dual-task paradigm for listening effort by assessing dual-task interference. Specifically, this study sought to determine the potential impact of instructions on the individuals’ chosen prioritization strategy as well as whether participants maintained their preferred strategy consistently across different listening conditions and test moments. The strategy to perform a dual-task is determined by the drive to minimize risk and maximize pleasure (Williams, 2006). Those aspects are individually driven and can vary over time. In the context of listening effort, it has been shown that various aspects related to the listener (e.g. age, hearing sensitivity, personality, motivation) and listening environment (e.g. background noise and reverberation) can impact how a listener will recruit the attentional and cognitive resources to understand speech (Peelle, 2018; Pichora-Fuller et al., 2016; Strand et al., 2018). Likewise, when performing a listening effort dual-task paradigm, a large interindividual variability in chosen prioritization strategy is hypothesized across different listening conditions and test moments, regardless of the given prioritization instructions.
MATERIALS AND METHODS
The current study was approved by the local Ethics Committee and was conducted in accordance with the Helsinki Declaration. Prior to participation, an informed consent was signed. No financial reward was given to the participants.
Participants
Normal-hearing adults between 18 and 69 years of age were recruited through convenience sampling. All participants were native Dutch (Belgium) speakers who had completed a minimum of 12 years of education (equivalent to high school graduation). All participants had either normal or corrected-to-normal vision, as assessed by the Near Vision Snellen Eye Chart (Snellen, 1873). Exclusion criteria encompassed tinnitus, self-reported excessive noise exposure within the past 72 hours, attention deficits, as well as learning, psychiatric, or (a history of) neurological disorders. For participants aged 60 years and older, the Montréal Cognitive Assessment (with a cutoff score of 26) was administered to screen for potential cognitive impairment (Nasreddine et al., 2005).
Bilateral normal middle-ear status was confirmed for all participants through otoscopy and 226 Hz tympanometry (TympStar Pro). Air-conduction pure-tone thresholds were obtained bilaterally through headphones for octave frequencies between 250 and 8000 Hz using the modified Hughson-Westlake method (Equinox 2.0 audiometer with TDH39 headphone, Interacoustics). Pure-tone audiometry, performed in a sound-treated booth, adhered to the International Organization for Standardization (ISO) 8253-1 guideline for ambient sound pressure levels, ensuring accurate testing ((ISO, 2010). Participants were included when air-conduction thresholds at each tested octave frequency were symmetrical and when the better-ear audiometric threshold averaged over 0.5, 1, 2, and 4 kHz was equal or below 20 dB HL (World Health Organization, 2021). Asymmetric hearing was identified as a difference of 15 dB or more between the right and left ears at three consecutive frequencies.
Dual-task paradigm
A dual-task paradigm, based on that reported in Degeest et al. (2015), was used. The primary and secondary task consisted respectively of a speech understanding and visual memory task, and were both performed separately (baseline condition) and simultaneously (dual-task condition).
Baseline condition
The primary task was a speech understanding task presented in different listening conditions, varying in listening demand. Words of the Brugge-Leuven-Utrecht (BLU) list were used as speech stimuli. This BLU list is endorsed by the Belgian National Institute for Health and Disability Insurance in, among other things, the context of hearing aid reimbursement. Consequently, in Belgium, this word list is considered gold standard within daily audiological practice and was, therefore, used within the current study. The BLU list contains two-syllable words. All words were spoken by a native Flemish female speaker without coarticulation. A steady state noise was generated and spectrally shaped to reflect the long-term average speech spectrum of the speech material. Words and noise were digitally mixed to create different signal-to-noise ratios (ranging from +8 dB SNR to −4 dB SNR in steps of 4 dB) whereby a constant noise level (i.e. 65 dB SPL) was utilized to which the intensity level of the speech stimuli was altered. Before stimulus onset, 10 seconds noise followed by a one second attention pure-tone of 1000 Hz were played to adapt to the noise and to focus participants’ attention to stimulus presentation, respectively. In addition to these noise conditions, a quiet listening condition without background noise was used. Within each listening condition, three trials of five words were presented. The words and background noise were presented through two loudspeakers, positioned at 45° and 315° azimuth at a distance of 90 cm from the participant’s ear. After each trial, participants verbally repeated the words. A word scoring was used, leading to a total score of 15 per listening condition.
The secondary task was a visual memory task in which a raster (i.e. 12 separated squares) was shown on a white standard computer screen (see example in Degeest et al. (2015)). Participants were seated on a chair and sat 70 cm in front of the computer, positioned at eye height. Within the raster, series of five identical blue filled circles appeared successively for 1 s each. Participants had to memorize the squares in which the circles appeared and had to indicate the position on a score form. It was obligated to indicate five squares, even if guessing was necessary. For each square correctly designated, one point was assigned. Five trials of five circles were presented. To determine the secondary baseline score, only the score of the fourth and fifth trials were used in order to control for learning effects (Degeest et al., 2015). Eventually, this score on ten was multiplied by 1.5, to reach a score on 15, making it comparable with the primary scores.
Dual-task condition
In the dual-task condition, participants underwent three trials where sequences of five words were presented alongside the appearance of five circles. This occurred under each of the five fixed listening conditions (quiet, SNR +8 dB, SNR +4 dB, SNR 0 dB, SNR −4 dB SNR), mirroring the approach used in the baseline conditions. To mitigate any potential conditioning effect, it was ensured that the auditory presentation of a word and the visual appearance of a circle did not happen simultaneously. Participants had to verbally repeat the words, after which the position of squares was assigned on the score form. For each correctly repeated word and correctly designated square, one point was counted for the primary and secondary task respectively, leading to a total score of 15 per listening condition for both tasks.
Dual-task interference
Each participant’s dual-task interference was measured by calculating the DTE for both the primary and secondary task separately (DTE = 100 × [score in dual-task condition - score in baseline condition]/score in baseline condition) (Gagne et al., 2017; Plummer & Eskes, 2015). By plotting the DTE of the primary and secondary task against each other (Figure 1), nine distinct patterns of dual-task interference can be identified: (1) speech priority trade off (+DTE in speech understanding task, -DTE in visual memory task), (2) mutual facilitation (+DTE in both tasks), (3) visual priority trade off (+DTE in visual memory task, - DTE in speech understanding task), (4) mutual interference (-DTE in both tasks), (5) visual memory facilitation (zero DTE in speech understanding task, +DTE in visual memory task), (6) visual memory interference (zero DTE in speech understanding task, -DTE in visual memory task), (7) speech facilitation (zero DTE in visual memory task, +DTE in speech understanding task), (8) speech interference (zero DTE in visual memory task, -DTE in speech understanding task), and (9) no interference (zero DTE in both tasks) (Plummer & Eskes, 2015).
Framework of the dual-task interference patterns by plotting the dual-task effect (DTE) of the primary and secondary task against each other.
Test procedure
The dual-task paradigm was carried out in a quiet, non-reverberant room illuminated with daylight and standard room lighting. At the beginning of each test, a practice trial was performed to verify whether the task was understood. If not, additional instructions were given until the participants fully understood the task.
To assess the effect of instructions on the prioritization strategy, participants were divided into two groups: the prioritizing and non-prioritizing group. The instruction provided during dual-task testing varied between the groups. In the prioritizing group, participants were directed to prioritize the primary speech understanding task (Gagne et al., 2017). Conversely, the non-prioritizing group received no specific instructions regarding task priority. The initial participants were randomly assigned to either the prioritizing or non-prioritizing group, after which matching participants were selected for the other group. To account for influencing factors, one-on-one matching was conducted based on age (±2 years), gender, and educational level (≤ 12 years of education or > 12 years of education). The test sequence of the dual-task paradigm (primary baseline, secondary baseline, and dual-task testing) as well as the sequence of the listening conditions within the primary baseline and dual-task testing were randomized. An identical test sequence was used for matched participants. After two weeks (mean interval: 14.8 days, range: 9 – 23 days), all participants re-executed the dual-task paradigm using an identical test sequence.
Statistical Analysis
The prioritization strategy employed by individuals during a dual-task paradigm for listening effort was assessed through dual-task interference. Descriptive analysis (SPSS version 29) was performed to assess which prioritization strategies were preferred across listening conditions and test moments as well as whether those strategies differed between the prioritization and non-prioritization groups.
RESULTS
Participants
Out of the 48 participants involved, one exhibited hearing loss. Consequently, this participant and his/her/them matching person were excluded, resulting in a total of 46 included participants. Table 1 shows the demographic data of the prioritizing and non-prioritizing group which were matched for gender, age, and educational level.
Dual-task interference
The direction of the dual-task interference was assessed by plotting the DTE of the primary and secondary task against each other for both groups at the first (Figure 2) and second (Figure 3) test moment. Note that markers (circle or rhombus) may overlap if multiple participants exhibited the same outcome. Table 2 displays the number of participants corresponding to each dual-task interference pattern across all listening conditions and test moments.
Framework of the dual-task interference patterns for the non-prioritizing and prioritizing groups at each listening conditions at test moment 1.
Framework of dual-task interference patterns for the non-prioritizing and prioritizing groups at each listening conditions at test moment 2.
The number (n, %) of participants (NP = non-prioritizing group, P = prioritizing group) corresponding to each dual-task interference pattern across all listening conditions and test moment 1 (T1) and 2 (T2). Note that red-marked numbers indicate a decrease, while green-marked numbers indicate an increase compared to test moment 1.
When assessing the first test moment, descriptive analysis revealed the application of diverse strategies for executing the dual-task. Overall, the prioritizing group exhibited a higher occurrence of visual memory interference (pattern 6) and speech priority trade-off (pattern 1) compared to the non-prioritizing group. Conversely, the non-prioritizing group employed mutual interference (pattern 4) and speech interference (pattern 8) more frequently than the prioritizing group. Notably, almost all participants of both the prioritizing and non-prioritizing groups used different strategies across the listening conditions within the first test moment (Table 3).
When descriptively analysing the dual-task interference across the two test moments, a shift in the pattern distribution was apparent (Table 2). In both the prioritizing and non-prioritizing group, a notable shift was observed, characterized by a decreased number of participants grappling with mutual interference (pattern 4) across nearly all listening conditions. Hence, participants showcased improved scores in at least one aspect of the dual-task condition during the second test moment compared to the first test moment. Furthermore, the prioritizing group tented to employ visual memory interference (pattern 6), speech interference (pattern 8), and no interference (pattern 9) more often during the second test moment, while the non-prioritizing group leaned more towards speech priority trade-off (pattern 1) than during the first test moment across nearly all listening conditions. As within the first test moment, the majority of participants in both the prioritizing and non-prioritizing group employed varied strategies across the listening conditions during the second test moment (Table 3).
DISCUSSION
Understanding how listeners execute a dual-task paradigm for listening effort would provide a benchmark for future studies and clinical implementations. Therefore, this study aimed to examine the prioritization strategy employed by individuals during a dual-task paradigm for listening effort by assessing dual-task interference.
Within the current study, the prioritizing group had more participants who achieved stable or better scores on the primary task in the dual-task condition compared to the non-prioritizing group. This observation might indicate an influence of instruction that does cause attentional and cognitive abilities to be more allocated to the primary task in the prioritizing group. However, there was considerable variability in the prioritizing strategy employed at the individual level across listening conditions and test moments, regardless the given prioritization instructions. Hence, in line with previous research on children (Choi et al., 2008), providing prioritization instructions was insufficient to ensure that an individual will mainly focus on the primary task and will stick to this strategy across listening conditions and test moments. This finding has important implications for the commonly used formula when evaluating listening effort through a dual-task paradigm. More specifically, the DTE of the secondary task is traditionally used as a measure of listening effort, under the assumption of stable primary scores in the baseline and dual-task condition (Gagne et al., 2017). Based on the current results, however, stable primary scores can not be guaranteed through instructions. If only the DTE of the secondary task is taken into consideration, disregarded the DTE on the primary task, incomplete conclusion will be made. For example, when evaluating the impact of hearing rehabilitation on listening effort through the standard formula, no change in the dual-task cost of the secondary task would be interpreted as an indication that the hearing rehabilitation had no effect on listening effort. However, no change in the dual-task cost of the secondary task following rehabilitation may occur with a decrease in the dual-task cost of the primary task. This scenario implies that, due to rehabilitation, more attentional and cognitive resources were accessible which were directed towards the primary speech understanding task, suggesting a decline in listening effort (Gagne et al., 2017). This example emphasizes how an incomplete evaluation of the dual-task performance can offer misleading information.
While a dual-task paradigm for listening effort is frequently used for its ecological validity, the present study raised certain reservations about its clinical applicability. First, the standard formula provides an insufficient or partial representation as it does not take dual-task interference into account. Second, evaluating dual-task interference provides insight into the used strategy to perform a dual-task paradigm for listening effort but cannot serve as a real metric for listening effort. Third, there seems to be no singular strategy when performing a dual-task paradigm for listening effort, not even within the same individual. Fourth, a learning effect might be present since a reduced number of participants handling mutual interference was observed when comparing the performances on the second and first test moment. Hence, some participants scored better on at least one aspect of the dual-task condition during the second test moment. Last, some participants exhibited no interference. This suggests that for both the primary and secondary tasks, there were no differences in performance between the baseline and dual-task conditions. For these participants, the dual-task may have been insufficiently challenging, implying that their cognitive capacity was not exceeded, thereby eliminating the necessity to prioritize between tasks.
Strengths, limitations and future perspectives
While the concept of integrating dual-task interference into research on listening effort was put forth years ago (Degeest et al., 2021; Gagne et al., 2017), this study is, to the best of our knowledge, one of the first to actually implement it. The results provided essential information on how normal-hearing listeners execute a dual-task paradigm for listening effort, questioning its clinical applicability as a metric for listening effort. The current study population does not represent the entire audiological target population. Hence, the current results can be considered a first exploration and should, therefore, be further investigated within diverse populations (such as individuals with hearing loss and/or hearing devices) as well as with diverse dual-task paradigms. Apart from assessing the dual-task paradigm’s reliability as a metric of listening effort, dual-task interference can be used as a counseling tool to enhance listeners’ comprehension of the limited capacity theory and the subsequent experienced listening effort.
Conclusion
Providing prioritization instructions when performing a dual-task paradigm for listening effort was insufficient to ensure that an individual will mainly focus on the primary task and will stick to this strategy across listening conditions and test moments. These results raised certain reservations about the current usage of dual-task paradigms for listening effort.
Data Availability
All data produced in the present study are available upon reasonable request to the authors
Footnotes
Financial disclosures/conflicts of interest: none to report
