Ambient noise is a ubiquitous feature of modern environments, frequently present in classrooms, offices, and homes. Its influence on cognitive performance, particularly attention, has been studied for several decades. Accumulated evidence suggests that the effects of background sound depend jointly on the demands of the concurrent task, the acoustic characteristics of the sound, and the characteristics of the listener. Tasks that place high demands on working memory or sustained attention, such as proofreading or mental arithmetic, are especially susceptible to interference from sounds carrying semantic information or dynamic pitch variations [1]. In contrast, more rule-based or highly structured tasks, such as grammatical judgments or missing-item detection, tend to be comparatively robust to auditory distraction [2]. Responses to background noise also vary substantially across individuals. While some people report improved focus or learning efficiency in the presence of background music or ambient noise [3,4], others experience increased distraction and performance decrements under similar conditions [1]. This heterogeneity has been attributed to multiple interacting factors, including developmental stage (eg, children vs adults), individual cognitive and affective traits, task characteristics, and the acoustic properties of the noise itself [1,2,5-9]. These findings underscore the importance of considering both acoustic and individual factors when investigating the effects of background sound on attention.

This study builds on 3 areas of prior research, each of which leaves an important question unresolved. First, a substantial body of controlled research has examined the cognitive effects of music and stationary nonspeech noise, particularly white and pink noise, and systematic reviews have characterized this literature in considerable detail [1,10]. Research has also examined ambient noise derived from human voices, which more closely resembles everyday environments such as cafés, libraries, and shared workspaces [2,3,7,9]. Meta-analytic evidence further indicates that task-irrelevant speech can interfere with reading performance [11]. However, studies using speech stimuli have generally focused on speech intelligibility, semantic content, or the presence vs absence of speech rather than isolating paralinguistic configuration itself. Consequently, prosodic, rhythmic, and other dynamic vocal features often remain confounded with linguistic meaning, making it difficult to determine whether nonsemantic properties of human voices independently influence attention. Although recent studies have begun to examine paralinguistic speech features in relation to outcomes such as stress recovery [12], their specific role in attentional performance remains poorly understood.

Second, existing theoretical accounts provide different explanations for how background sounds may influence attention. The Moderate Brain Arousal (MBA) model proposes that moderate levels of noise can facilitate cognitive performance by introducing neural variability that may enhance weak signals through stochastic resonance [13]. Under this account, the effects of noise depend partly on baseline arousal and individual characteristics. Individuals with lower baseline dopamine activity, including some individuals with attention-deficit hyperactivity disorder (ADHD), have been hypothesized to benefit more from moderate noise, whereas those with higher baseline arousal may be more susceptible to distraction [4]. The MBA model is therefore concerned primarily with noise-induced arousal and does not, in its standard formulation, distinguish effects arising from speech structure from those arising from other sound sources.

Interference-based accounts instead emphasize the structural and attentional properties of sound. Changing-state accounts propose that acoustically varying sequences can interfere with task-relevant serial processing, while attentional-capture accounts emphasize disruption caused by salient or deviant auditory events; the duplex-mechanism account distinguishes these 2 routes to auditory distraction [2,6,8]. Paralinguistically rich but semantically inaccessible speech provides a useful case for examining these different explanations because it retains dynamic acoustic properties while minimizing accessible linguistic meaning. From an interference perspective, paralinguistic variation may contribute to disruption by creating changing-state or attention-capturing signals, whereas the MBA model emphasizes noise-induced arousal and individual differences rather than speech structure per se. Examining nonsemantic vocal stimuli may therefore help clarify the conditions under which these mechanisms contribute to attentional performance.

Third, limitations in existing measurement approaches make it difficult to determine the processes underlying observed behavioral effects. Accuracy and response time can establish whether performance differs across auditory conditions but provide limited information about accompanying changes in cognitive effort or physiological arousal. Psychophysiological measures can provide complementary information about these responses. For example, pupil dynamics are associated with arousal and cognitive effort, electrodermal activity (EDA) with sympathetic activation, heart rate variability with autonomic activity, and spontaneous blink rate with aspects of attentional state [14-16]. Studies combining behavioral and physiological measures have demonstrated the value of such multimodal approaches in occupational and applied settings [15,16], but they remain uncommon in the background-speech literature, where physiological investigations have more often focused on affective or stress-related outcomes than on sustained attention [12]. In addition, individual characteristics including noise sensitivity, working memory capacity, sleep quality, and mood have been associated with subjective or cognitive responses to background noise [17-19]; yet, they have rarely been examined together within a common experimental framework.

These gaps motivate our experimental design, which isolates differences in paralinguistic configuration while minimizing semantic accessibility and examines behavioral and physiological responses under controlled auditory conditions. Our protocol also measures individual differences that may contribute to variability in responses to background speech, allowing their potential moderating roles to be explored.

Before finalizing the present protocol, we conducted a formative pilot with 9 participants using an abbreviated procedure comprising the N-back task and 2 of the 3 auditory-condition blocks. The pilot used a separate sample and was conducted solely to assess procedural and technical feasibility, calibrate the newly developed reading comprehension task, determine whether the auditory stimuli were perceptually distinguishable, and refine procedural parameters. It was neither designed nor powered to test the study hypotheses, no inferential analyses were performed, and the pilot data will not contribute to the full study described in this protocol.

The pilot demonstrated that the behavioral tasks and concurrent eye-tracking and wrist-worn physiological recording could be completed within a single laboratory session. Participant feedback and task responses were used to refine the reading comprehension materials and indicated that the focused and diffused auditory stimuli were perceived as qualitatively distinct. Observations regarding playback-volume selection and task recovery also informed retention of the available volume range and adoption of a 5-minute rest interval between experimental blocks.

The primary objective of this study is to investigate whether paralinguistic characteristics of background human speech influence attentional performance. Specifically, the study will compare focused and diffused vocal stimuli that differ in paralinguistic configuration while minimizing semantic accessibility, with silence included as a reference condition.

The primary analysis will examine whether attentional performance, measured by the d2 concentration performance (CP) score, differs between the focused and diffused vocal conditions. Reading comprehension will be examined as a secondary behavioral outcome. Secondary analyses will examine physiological responses across auditory conditions using pupillometry, EDA, and heart rate variability. The study will also explore whether individual differences, including noise sensitivity, ADHD symptomatology, working memory capacity, sleep quality, and mood, moderate behavioral and physiological responses to the vocal conditions.

The study is guided by the following hypotheses:

This study is a nonrandomized, within-subject laboratory experiment involving 3 counterbalanced auditory conditions: focused vocal stimuli, diffused vocal stimuli, and silence. Applicable items from the SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) 2025 reporting guideline were used to support comprehensive reporting of the protocol (Checklist 1) [20]. Figure 1 illustrates the overall experimental procedure and timeline.

Members of the public were not formally involved in the design, conduct, or reporting of this study. Participant experience nonetheless informed the final design. Feedback from the 9 formative pilot participants on session length, task difficulty, stimulus volume, and the perceived character of the auditory conditions directly shaped the session structure, the retained 20‐65 dB volume range, and the 5-minute rest intervals, as described in the Introduction. Public involvement in interpreting and disseminating the findings will not be planned beyond the participant summary described under dissemination policy.

Ethics approval for the study was obtained from the Ethics Review Board of Kyoto University of Advanced Science (approval number 26E01; June 22, 2026). The separate formative pilot described in the Introduction was conducted under approval number 22E01. Participation will be entirely voluntary. Prospective participants will receive a written information sheet describing the study purpose and procedures, physiological measurements, expected session duration, data handling, compensation, and their right to withdraw. Written informed consent will be obtained from all participants before participation in the study. Participants will be informed that they may withdraw at any time without providing a reason and without penalty. They will also be informed of the procedures governing data collected before withdrawal, including whether and under what circumstances previously collected identifiable data can be deleted.

Each participant will be assigned a study identifier at enrollment. Questionnaire, behavioral, and physiological data will be stored using this identifier, while the linkage file connecting study identifiers with personally identifying information will be stored separately and accessible only to the principal investigator and authorized study personnel. Analytic datasets will be deidentified, although they will not be considered fully anonymized while the linkage key is retained. Data-management computers will be maintained offline with access controls, and installation of unauthorized software will be restricted. Study personnel with data access will complete institutional information-security training. Raw eye-tracking recordings, including scene-camera footage, and raw physiological recordings will be treated as potentially identifiable or sensitive data and subject to additional safeguards. These data will be stored in encrypted offline media, will not be publicly shared, and will be deleted 5 years after publication in accordance with the institutional retention policy.

Participants will receive a ¥3000 Amazon (ACI Gift Cards, Inc) gift card (approximately US $20) for completing the study session. Participants who withdraw before completing the session will not receive compensation.

Participants will be recruited through convenience sampling, primarily through campus advertisements.

Participants must meet the following criteria: (a) be 18 years of age or older; (b) self-report English proficiency of intermediate level or higher (≥3 on a 5-point Likert scale); (c) be able to attend the in-person experimental session; and (d) be able to complete eye-tracking tasks without wearing glasses (contact lenses permitted).

Participants will be excluded if they self-report (a) a diagnosed neurological disorder, or (b) severe medical or psychiatric disorders that may interfere with task performance or physiological measurements.

No additional restrictions based on demographic characteristics will be imposed. To account for individual differences that could influence cognitive outcomes, participants will complete a battery of psychometric instruments assessing sleep quality, mood, chronotype, ADHD symptoms, dyslexia risk, and noise sensitivity, as well as the N-back test. These measures will be used to characterize individual differences that may influence cognitive performance, with prespecified measures examined as exploratory moderators as described in the analysis plan.

The sample size was determined for the primary confirmatory outcome, the CP score of the d2 Test of Attention, and the prespecified focused-vs-diffused contrast. This contrast directly addresses the primary hypothesis (H1) that attentional performance differs according to paralinguistic configuration when both conditions contain speech-derived background sound.

No previous study has examined the focused-vs-diffused paralinguistic manipulation used in the present protocol with d2 CP as the outcome, and a directly applicable effect-size estimate is therefore unavailable. The sample-size calculation was therefore based on the minimum standardized within-participant effect that the study should be able to detect. For the prespecified focused-vs-diffused contrast, sample size was calculated for a two-sided paired comparison with α=.05, power of 0.80, and a standardized within-participant effect of dz=0.40. Using the noncentral t distribution, the minimum required sample was 52 participants with complete primary-outcome data. To provide a modest allowance for incomplete or unusable primary behavioral data, we plan to recruit 60 participants. The study may have limited power to detect effects smaller than dz=0.40; effect estimates and 95% CI will therefore be reported alongside hypothesis tests.

The sample-size calculation applies specifically to the primary focused-vs-diffused comparison under H1. The physiological outcomes under H2 were not independently used to determine the sample size because directly applicable effect-size estimates are unavailable for the present stimulus manipulation. The moderation analyses under H3 involve interaction effects and were also not used to determine the target sample size. These analyses are therefore designated exploratory and will be interpreted primarily using interaction estimates and 95% CIs.

This study adopts a within-subject, single-session experimental design. Each participant will complete 1 laboratory session lasting approximately 120 minutes, during which attention-related tasks are performed under 3 auditory conditions: (1) background noise with focused vocal stimuli, (2) background noise with diffused vocal stimuli, and (3) silence (control). All participants will be exposed to all 3 conditions within the same session. The order of auditory conditions will be counterbalanced across participants using a predetermined schedule to control for order effects. Figure 1 illustrates the experimental timeline, which consists of the following phases.

The primary objective of the analysis will be to determine whether attentional performance differs between the focused and diffused vocal conditions. Secondary analyses will examine the effects of auditory condition on reading comprehension and physiological responses. Exploratory analyses will examine whether individual differences, including noise sensitivity, ADHD symptoms, sleep quality, mood, and working memory performance, may moderate these effects.

The study poses minimal risk. The principal foreseeable burdens were fatigue and possible discomfort from the auditory stimuli, both of which were mitigated by design. Playback volume will not exceed 65 dB, participants will set the volume themselves within the permitted range, and rest intervals will separate the experimental blocks. Participants will be reminded at the start of the session that they may pause or discontinue participation at any time, and the researcher (C. D. K.) will monitor for signs of fatigue or discomfort throughout the session. Any unexpected or serious adverse event related to study participation will be documented and reported to the Ethics Review Board in accordance with institutional procedures.

Ancillary and posttrial care requirements are expected to be limited given the minimal-risk nature of the study. Skin irritation or minor discomfort associated with wearing the wristband or eye-tracking headset will be addressed immediately by the study staff, and participants will be referred to the university nursing office if further attention is needed. Any study-related harm will be managed in accordance with the institutional policies of Kyoto University of Advanced Science. No compensation will be provided for health events unrelated to study participation.

Any substantive amendment to the protocol, including changes to eligibility criteria, study procedures, outcomes, or the analysis plan, will be submitted to the Ethics Review Board of Kyoto University of Advanced Science (KUAS) for approval before implementation. Approved amendments will be documented with the date, nature, and rationale for the change and communicated to study staff before implementation. Where an amendment affects information provided to participants or study procedures requiring consent, the participant information sheet and consent form will be revised and submitted for approval accordingly. All substantive amendments will be reported in the Registered Report Stage 2 manuscript. Any deviation from the prespecified confirmatory analysis plan will be clearly identified and justified, and the affected analyses will be reported as exploratory. Administrative changes that do not affect study conduct, participant safety, or the analysis plan will be documented internally and handled according to institutional requirements.

Ethics approval for the study was granted by the Ethics Review Board of KUAS (approval number 26E01) on June 22, 2026. The study received no external funding. The protocol was finalized following the formative pilot described in the Introduction, which was conducted with 9 participants between January and July 2025. No participants have been recruited for the full study described in this protocol. Recruitment is scheduled to begin in September 2027, with data collection projected to conclude in September 2029. Data analysis is expected to be completed by September 2030. Dissemination of the results is anticipated by September 2031.

Results of the full study will be reported in a subsequent peer-reviewed publication. Authorship will follow the International Committee of Medical Journal Editors (ICMJE) criteria, and no professional writers will be used. Deidentified participant-level behavioral and questionnaire data and derived physiological features will be deposited in PhysioNet (MIT Laboratory for Computational Physiology) [54], subject to the study’s ethics approval and data-sharing requirements. Analysis code will be made available through the project GitHub (GitHub, Inc) repository to support reproducibility. Findings will also be presented at relevant conferences in human-computer interaction and psychophysiology. Participants who wish to receive the study results will be provided with a plain-language summary by email.

This paper presents an experimental protocol designed to investigate whether background human speech with distinct paralinguistic characteristics, operationalized as focused vs diffused vocal stimuli, influences attentional performance. The protocol combines cognitive tasks, psychometric assessments, and physiological measurements in a counterbalanced within-subject design, which allows behavioral and physiological responses to be compared across auditory conditions while accounting for individual differences. The primary contribution of the protocol is the controlled manipulation of paralinguistic characteristics while minimizing semantic accessibility, providing an approach to investigating the role of nonsemantic vocal features in auditory distraction. Concurrent physiological measurements will complement the behavioral outcomes by examining physiological responses associated with auditory condition and attentional performance.

Previous research on noise and cognitive performance has primarily examined music, stationary nonspeech noise such as white or pink noise, or semantically intelligible speech, typically using behavioral outcomes such as accuracy and response time [1,2,5-9]. Theoretical frameworks offer different explanations for the effects of background sound. The MBA model proposes that moderate levels of noise may enhance or impair attention depending partly on baseline arousal and individual characteristics [13], whereas interference accounts emphasize disruption arising from changing-state acoustic structure and attentional capture [2,6,8]. However, comparatively little research has isolated the paralinguistic properties of human speech, such as prosodic and rhythmic variation, from its semantic content.

Research on background speech and cognitive performance has also relied primarily on behavioral measures, which provide limited information about physiological responses accompanying attentional performance. The present protocol combines established psychophysiological measures, including pupillometry, EDA, and BVP, with behavioral outcomes [14-16] to examine whether paralinguistic characteristics of semantically inaccessible background speech are associated with differences in attentional performance and physiological responses. The assessment of individual differences will further allow exploratory investigation of factors that may contribute to variation in susceptibility to auditory distraction.

The study may have implications for 3 related areas, although these implications will depend on the observed findings and should therefore be considered prospective. First, for research on auditory distraction, experimentally manipulating paralinguistic configuration while minimizing semantic accessibility would provide a means of examining whether nonsemantic properties of background vocal stimuli contribute to attentional disruption [2,6,8]. A difference in attentional performance between the focused and diffused conditions would provide initial evidence that attentional performance differs between the paralinguistic configurations examined and would motivate further investigation of the acoustic and attentional mechanisms underlying this difference. Conversely, little or no difference between the conditions would suggest that the specific paralinguistic manipulation examined here has limited influence on attentional performance under the tested conditions. In either case, the findings may help refine hypotheses concerning how the structural properties of background speech interact with attention.

Second, the findings may have implications for research on the design of occupational and educational environments. Background speech is common in shared offices, classrooms, libraries, and study spaces, and responses to such environments may vary according to individual characteristics, including noise sensitivity and ADHD symptomatology [4,5,7]. Evidence that particular paralinguistic configurations are associated with greater attentional disruption could inform future research on acoustic environments, including acoustic treatment, quiet-space provision, and masking or ambient sound [9,55]. The exploratory moderation analyses may highlight individual characteristics for further investigation. The present study is not designed to establish individualized accommodation recommendations but may help identify questions for subsequent, adequately powered research.

Third, the combination of behavioral and wearable physiological measurements may inform future research on adaptive sound environments and digital interventions. If physiological changes are associated with auditory condition or attentional performance, they could provide candidate markers for subsequent investigation in adaptive soundscape systems, focus-support applications, or context-aware interventions [15,16,55]. The present protocol provides a controlled setting for evaluating these associations. Determining whether such signals can support reliable individual-level prediction or real-time adaptation would require further validation in larger samples and real-world settings.

A key strength of the protocol is its controlled manipulation of paralinguistic characteristics while minimizing semantic accessibility. Comparing focused and diffused vocal stimuli within the same participants will provide a means of examining whether differences in nonsemantic vocal characteristics are associated with attentional performance while reducing between-participant variability. The inclusion of a silence condition will additionally provide a reference against which the effects of both vocal conditions can be evaluated.

The protocol is also designed to support replication and adaptation. The central auditory manipulation can be applied to different participant populations and cognitive tasks, while psychometric measures can be adapted to characteristics relevant to the population under investigation. Making the reading materials, auditory stimulus samples, and volume-control interface openly available will further support replication and extension of the study.

The use of behavioral and physiological measures will provide complementary information about responses to the auditory conditions. Behavioral outcomes will assess whether performance differs across conditions, whereas pupillometry, EDA, and BVP-derived measures will allow secondary and exploratory examination of accompanying physiological responses. The protocol therefore will provide a basis for investigating whether behavioral differences, if observed, are accompanied by measurable physiological changes.

Several limitations should be acknowledged. First, participants will select the playback volume within a predefined range of 20‐65 dB. Although this approach balances experimental control with participant comfort, it may introduce variability in perceived sound intensity across participants. Playback level will therefore be recorded to characterize between-participant variation in stimulus intensity and may be examined in exploratory sensitivity analyses where appropriate. Second, some procedures will require manual execution, including timing of the d2 Test of Attention and scoring of responses. Standardized instructions and scoring procedures will be used to minimize experimenter-related variability, but greater automation could improve precision in future implementations. Third, the reading comprehension task was developed specifically for this study and has undergone only formative pilot testing. The pilot was used to assess task feasibility and identify potential floor or ceiling effects rather than to establish psychometric validity or reliability. Its measurement properties will therefore require further evaluation using data from the full study. Finally, the study is powered for the prespecified focused-vs-diffused contrast on the primary d2 CP outcome. The sample size was not determined specifically for the secondary physiological outcomes or exploratory moderator interactions, and the study may therefore have limited power to detect effects in these analyses. Findings from these analyses will require confirmation in appropriately powered future studies.

This protocol provides a controlled approach to investigating whether paralinguistic characteristics of semantically inaccessible background speech influence attentional performance. By comparing focused and diffused vocal stimuli within a counterbalanced within-subject design, the study will test their effects on the primary behavioral outcome while examining secondary behavioral and physiological responses and exploring potential individual differences. The findings may contribute to understanding auditory distraction and inform future research on the design of occupational, educational, and adaptive sound environments.