The Point of Pointing: Deictic Gestures Modulate Attentional Shifts and Cognitive Load in Simultaneous Interpreting

The Point of Pointing: Deictic Gestures Modulate Attentional Shifts and Cognitive Load in Simultaneous Interpreting | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article The Point of Pointing: Deictic Gestures Modulate Attentional Shifts and Cognitive Load in Simultaneous Interpreting Dongpeng Pan, Kilian G. Seeber This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9294880/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Spoken language processing requires the integration of acoustic signals with visual information. Deictic gestures, such as pointing, may facilitate this integration by constraining the referential domain before linguistic disambiguation takes place. However, it remains unclear how multimodal cues modulate processing under increased cognitive load. Simultaneous interpreting provides a task environment for the examination of speech processing under substantial cognitive load. In this study, we combined the Visual World Paradigm (VWP) with pupillometry to investigate how congruent, incongruent, and neutral pointing gestures influence attention and cognitive load in 24 professional simultaneous interpreters. Eye-tracking revealed that congruent gestures elicited early anticipatory visual target fixations. Incongruent gestures directed gaze toward visual competitors before rapid correction, while the absence of gestures delayed visual target identification. Counterintuitively, the gesture-absent condition produced greater pupil dilation than the incongruent condition, while congruent and incongruent conditions did not differ significantly in cognitive load. These results challenge the assumption that cue accuracy drives processing cost and instead suggest that the presence of a visual cue, even a misleading one, is the critical variable that reduces cognitive demands. Biological sciences/Neuroscience Biological sciences/Psychology Social science/Psychology speech processing simultaneous interpreting deictic gestures Visual World Paradigm eye-tracking cognitive load Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Language processing involves integrating speech with visual signals [ 1 , 2 ]. Beyond the speech signal, communicators deploy visual cues such as eye gaze and hand movements that interlocutors integrate in real time to construct meaning and generate predictions about upcoming content [ 3 ]. Specifically, referential gaze and pointing gestures guide visual attention toward relevant objects, thereby grounding linguistic expressions in the physical environment [ 4 , 5 ]. Theoretical frameworks suggest that speech and gesture function as a system where the two modalities mutually and intrinsically influence one another during comprehension [ 6 ]. At the same time, behavioural evidence indicates that listeners process congruent multimodal information more efficiently, whereas conflicting inputs increase processing costs [ 6 , 7 ]. Furthermore, meta-analytic reviews confirm that gestures benefit comprehension, particularly when they depict motor actions or convey information that supplements the spoken message [ 8 ]. This multimodal integration appears critical in complex tasks like simultaneous interpreting, which represents an extreme form of bilingual language control requiring concurrent comprehension and production [ 9 , 10 ]. In this cognitively demanding context, visual input such as pointing may act as aid that reduces cognitive load and facilitates the retrieval of meaning [ 11 – 13 ]. Specifically, visual aids facilitate the prediction of message content, whereas the lack of these cues in many distance interpreting settings has been suggested to increase extraneous cognitive load and to contribute to fatigue [ 11 , 14 , 15 ]. Among the various forms of nonverbal communication, deictic gestures, i.e., pointing actions that orient attention toward specific referents, fulfil a unique role by anchoring speech to the environment [ 4 , 16 ]. While iconic gestures depict semantic attributes and beat gestures emphasise prosody, pointing provides referential precision. This facilitates target identification without imposing significant interpretive demands [ 5 ]. There is substantial behavioural evidence identifying pointing as the dominant cue for referent selection, consistently serving as the strongest predictor of attention even when pitted against conflicting eye gaze or verbal demonstratives [ 4 ]. Viewers observing deictic gestures also demonstrate better retention than those watching beat gestures or no gestures, suggesting that pointing supports memory [ 16 , 17 ]. During the cognitively demanding task of simultaneous interpreting, these gestures can function as complementary visual cues that can reduce the processing burden associated with auditory comprehension [ 12 , 18 ]. The VWP provides a solid methodological framework to explore how multimodal cues modulate real-time language processing [ 19 , 20 ]. By tracking eye movements as participants listen to spoken language while viewing visual displays, the VWP enables the real-time measurement of incremental language processing with high temporal resolution [ 20 , 21 ]. Fixation probabilities, for example, align with the predictions of lexical activation models, revealing that words with similar onsets or rhymes compete during spoken word recognition [ 19 , 22 ]. Anticipatory eye movements based on information provided by the subject and/or verb emerge before the relevant noun is heard, revealing that listeners use linguistic and visual input to predict upcoming content [ 23 , 24 ]. In bilingual and multilingual research, the VWP has allowed to uncover cross-linguistic co-activation and the role of visual cues in resolving competition between languages [ 25 ]. Fridman and Meir [ 26 ], for instance, used the VWP to show asymmetrical activation in trilingual speakers, where visual contexts, including deictic elements, accelerate fixation shifts and mitigate interference from dominant languages. Abashidze et al. [ 27 ] illustrated how deictic contexts counteract biases in bilingual prediction, fostering efficient resolution by guiding gaze to intended targets. Today, methodological advances, particularly generalised additive mixed models (GAMMs), enable researchers to capture the non-linear temporal dynamics in these fixation trajectories [ 28 – 30 ]. This analytical approach accounts for the autocorrelation inherent in time-series data, offering a precise estimation of when effects arise during online processing [ 31 , 32 ]. Language processing has also shown to mobilise physiological mechanisms that index the intensity of cognitive work [ 33 ]. As a complement to eye-tracking data, pupillometry can be used to measure task-evoked pupillary responses (TEPRs) and thus quantify the allocation of mental effort on a moment-to-moment basis [ 34 ]. The underlying theoretical rationale is that pupil diameter increases in response to processing demands, reflecting the consumption of limited attentional resources required for task performance [ 34 , 35 ]. This metric has proved highly sensitive to linguistic complexity; for instance, sentences containing syntactic ambiguities or conflicting prosodic cues elicit greater dilation than their simpler counterparts, indicating the immediate cost of resolving structural conflict [ 33 , 36 ]. The modality of input also modulates this response; auditory presentation of tasks typically induces larger dilations compared to visual presentation, potentially because visual formats support dual coding strategies that alleviate the burden on working memory [ 37 ]. In multilingual contexts, pupillometry has been used to capture the cognitive cost of cross-linguistic competition, where bilinguals engaging in code-switching or processing a second language exhibit sustained dilation patterns [ 38 ]. These findings are particularly relevant for simultaneous interpreting, a task that generates significantly higher pupillary responses compared to listening or shadowing due to the concurrent demands of comprehension and production [ 39 ]. Furthermore, specific linguistic factors such as syntactic asymmetry between source and target languages trigger measurable increases in pupil diameter, providing objective evidence of the load associated with restructuring verbal input under time pressure [ 40 ]. The present study combines the VWP with pupillometry to investigate how congruent, incongruent, and neutral deictic gestures influence attention allocation and cognitive load in professional conference interpreters. We hypothesise that congruent pointing guides attention to the target referent [ 4 ], thereby reducing the cognitive load associated with reference resolution. Conversely, we predict that incongruent gestures will direct attention toward competitor objects [ 22 ], inducing a cognitive conflict that may produce distinct pupillary signatures reflecting cross-modal reconciliation. Finally, we anticipate that the neutral condition (without deictic gestures) will impose a greater processing load than congruent pointing, as the interpreter must rely exclusively on the auditory signal to construct meaning [ 3 ]. Results Fixation Dynamics Across Gesture Conditions Descriptive fixation proportions revealed that interpreters predominantly directed gaze toward the speaker's head region throughout the sentence. Condition-specific modulations in object-directed fixations emerged after modifier onset, with divergence patterns indicating differential engagement with the visual display across gesture conditions (Fig. 1 ). Logistic generalised additive mixed models (GAMMs) modelling binary target fixation (target = 1, non-target = 0) quantified these condition effects across the three temporal windows (Table 1 ). Fitted smooth functions for all model terms are presented in Supplementary Material S2.1 (Figures S1 –S3); model diagnostics and autocorrelation function (ACF) plots appear in S2.2 (Figures S4–S6). During the modifier window, where the adjective preceded the disambiguating noun, a strong baseline target preference was evident in the model intercept (β = 9.41, SE = 0.39, z = 24.20, p < .001). Both neutral and incongruent gestures produced significant reductions in target fixation probability relative to the congruent baseline. The neutral condition reduced target fixation by approximately 1.10 log-odds units (β = −1.10, SE = 0.38, z = − 2.89, p = .004), while incongruent gestures produced a somewhat larger reduction of 1.36 log-odds units (β = −1.36, SE = 0.36, z = − 3.81, p < .001). Examination of the condition-specific smooth terms revealed that only the congruent condition exhibited significant time-varying dynamics (edf = 2.98, χ² = 2.90, p = .047). Neither neutral (edf = 2.00, χ² = 0.07, p = .965) nor incongruent conditions (edf = 2.00, χ² = 0.79, p = .674) showed significant temporal dynamics. Pairwise difference curves confirmed a congruent advantage over incongruent cues in early and late modifier intervals (normalised time 0.004–0.487 and 0.688–1.000) and over neutral cues similarly (0.004–0.386 and 0.879–1.000). No significant difference emerged between neutral and incongruent conditions at any point in the modifier window (Fig. 2 , Modifier column). In the object window, the overall target preference persisted (β = 9.14, SE = 0.31, z = 29.25, p < .001) with attenuated but still significant condition effects. The neutral condition reduced target fixation by 0.67 log-odds units relative to congruent (β = −0.67, SE = 0.21, z = − 3.26, p = .001), while incongruent gestures produced a similar reduction of 0.74 log-odds units (β = −0.74, SE = 0.20, z = − 3.67, p < .001). Spatial lag terms showed increased complexity during this window (Time × Lagged Distance to Target: edf = 4.31, χ² = 706.46, p < .001), suggesting intensified competition dynamics. Condition-specific time-varying effects were non-significant across all three conditions. Pairwise comparisons revealed that congruent cues maintained an advantage over incongruent cues during the first half of the window (normalised time 0.043–0.486) and over neutral cues during the second half (0.668–0.940) (Fig. 2 , Object columns). Table 1 GAMM results for fixation analysis: condition effects across temporal windows. Reference level: congruent. Parametric coefficients are on the log-odds scale. Modifier Object Spillover Term Est/edf SE/Ref.df z/χ² p Est/edf SE/Ref.df z/χ² p Est/edf SE/Ref.df z/χ² p Parametric Intercept 9.41 0.39 24.20 < .001 9.14 0.31 29.25 < .001 11.03 0.45 24.39 < .001 Cond: Neutral −1.10 0.38 −2.89 .004 −0.67 0.21 −3.26 .001 0.08 0.16 0.48 .631 Cond: Incongruent −1.36 0.36 −3.81 < .001 −0.74 0.20 −3.67 < .001 0.49 0.16 3.12 .002 Lagged Target Fix 3.61 0.14 26.44 < .001 3.49 0.11 33.25 < .001 3.26 0.08 42.16 < .001 Smooth Time × Prior Target Fix 1.40 1.67 9.94 .011 7.09 8.12 18.96 .014 4.49 5.51 19.47 .003 Time × Lag Dist Target 3.49 4.08 289.72 < .001 4.31 5.29 706.46 < .001 3.69 4.59 1096.25 < .001 Time × Lag Dist Comp 2.16 2.25 68.30 < .001 4.58 5.61 56.83 < .001 2.21 2.75 38.49 < .001 Time (Overall) 1.01 1.02 3.80 .052 1.79 2.23 0.20 .949 1.94 2.41 3.13 .277 Trial Order 4.35 5.32 12.50 < .001 5.23 6.41 11.50 .083 7.35 8.46 36.72 < .001 Participant (RE) 12.65 23.00 15.20 < .001 6.75 23.00 44.43 < .001 10.20 23.00 193.16 < .001 Item (RE) 21.65 29.00 18.40 < .001 25.41 29.00 323.28 < .001 27.07 29.00 557.08 < .001 Time × Cond: Cong 2.98 3.71 2.90 .047 3.79 4.52 1.23 .926 3.21 3.95 6.48 .161 Time × Cond: Neut 0.02 0.05 0.10 .952 4.63 5.54 3.61 .706 1.43 1.95 1.99 .368 Time × Cond: Inc 1.02 1.04 0.50 .462 4.57 5.48 1.43 .950 2.59 3.15 5.07 .184 During the spillover window capturing the speaker's sentence offset to capture gaze patterns following the end of auditory input, the target preference reached its highest magnitude (β = 11.03, SE = 0.45, z = 24.39, p < .001). Critically, the pattern of condition effects reversed compared to earlier windows. The neutral condition no longer differed significantly from congruent (β = 0.08, SE = 0.17, z = 0.48, p = .631). In contrast, incongruent gestures produced a significant enhancement of target fixation relative to the congruent baseline (β = 0.49, SE = 0.16, z = 3.12, p = .002). Pairwise comparisons confirmed this incongruent advantage over both congruent (normalised time 0.053–0.184 and 0.648–0.960) and neutral conditions (0.496–0.980, Fig. 2 , Spillover columns). Target–Competitor Competition Dynamics To characterise within-condition competition independent of between-condition differences, we modelled binary area-of-interest (AOI) outcomes restricted to target and competitor fixations (target = 1, competitor = 0), following the analytical approach of Stone et al. [ 41 ]. This analysis focused on the temporal divergence term s(Time):RegionBin , which captures the time-course of target versus competitor differentiation within each condition and window. Pairwise difference curves generated via the plot_diff() function ( itsadug package [ 42 ]) identified intervals where fixations to target and competitor significantly diverged from equality. All intervals are expressed in normalised time units (0–1 within each window). Full model specifications are described in Supplementary Material S3.1; complete parametric and smooth term results for all nine models appear in S3.2 (Tables S1–S9); model diagnostics and ACF plots are presented in S3.3 (Figures S7–S15). When gestures were congruent, target dominance emerged rapidly and persisted across all three windows. In the modifier window, the temporal divergence term was significant (edf = 2.98, χ² = 27.85, p < .001), with target fixations exceeding competitor fixations across a sustained interval spanning normalised time 0.034–0.849, encompassing the vast majority of the pre-disambiguation window. In the object window, target dominance was again significant (edf = 3.10, χ² = 38.72, p < .001), with target exceeding competitor from normalised time 0.164 through window end (1.000). In the spillover window, target dominance was significant (edf = 2.00, χ² = 24.19, p < .001) across normalised time 0.003–0.930(Fig. 3 , upper row). When gestures were incongruent, a biphasic trajectory emerged. In the modifier window, the temporal divergence term was significant (edf = 2.00, χ² = 19.85, p < .001), but critically, the direction was reversed compared with the congruent condition. Here, competitor fixations exceeded target fixations across nearly the entire window (normalised time 0.004–0.950). In the object window, the divergence term remained significant (edf = 2.46, χ² = 28.50, p < .001) but revealed a crossover pattern. The competitor initially exceeded target during normalised time 0.003–0.335, followed by target exceeding competitor from 0.627–1.000 as disambiguating noun input accumulated. By the spillover window, recovery was complete: the divergence term was highly significant (edf = 2.00, χ² = 56.68, p < .001), with target exceeding competitor across the full window (0.003–1.000, Fig. 3 , middle row). Table 2 GAMM results for target-competitor divergence across conditions and temporal windows. Condition Modifier Object Spillover Term edf Ref.df χ² p edf Ref.df χ² p edf Ref.df χ² p Congruent s(Time):ValueLag1 2.40 3.01 5.11 .168 2.00 2.00 2423.18 < .001 1.00 1.00 12.98 < .001 s(Time):RegionBin 2.98 3.48 27.85 < .001 3.10 3.59 38.72 < .001 2.00 2.00 24.19 < .001 s(Time) 1.00 1.00 0.92 .336 1.07 1.12 8.57 .005 1.00 1.00 4.28 .039 Incongruent s(Time):ValueLag1 4.05 4.85 1052.23 < .001 2.00 2.00 2902.54 < .001 1.00 1.00 3.29 .070 s(Time):RegionBin 2.00 2.00 19.85 < .001 2.46 2.77 28.50 < .001 2.00 2.00 56.68 < .001 s(Time) 1.00 1.00 1.07 .302 1.00 1.00 0.07 .794 1.00 1.00 0.93 .334 Neutral s(Time):ValueLag1 2.18 2.72 10.55 .015 2.00 2.00 1806.83 < .001 2.87 3.60 12.43 .013 s(Time):RegionBin 2.00 2.00 4.47 .107 2.00 2.00 7.12 .028 3.75 4.43 23.32 < .001 s(Time) 1.00 1.00 2.75 .097 1.00 1.00 2.64 .104 1.00 1.00 0.78 .377 When gestures were absent, processing was delayed. In the modifier window, the temporal divergence term failed to reach significance (edf = 2.00, χ² = 4.47, p = .107). A late-emerging tendency toward competitor dominance (normalised time 0.708–0.970) was observed but did not reach significance. In the object window, significant divergence emerged (edf = 2.00, χ² = 7.12, p = .028) but was restricted to the latter half of the window (normalised time 0.486–1.000). In the spillover window, the divergence term was significant (edf = 3.75, χ² = 23.32, p < .001). (Fig. 3 lower row). Cognitive Load: Pupillometric Evidence Baseline-corrected pupil diameter served as a continuous index of cognitive load dynamics across gesture conditions. Pupillometric data were modelled using GAMMs with scaled-t family distributions to accommodate the heavy-tailed characteristics typical of pupil size data [ 43 ]. A two-stage modelling approach addressed temporal autocorrelation: initial models estimated autoregressive coefficients ( rho ), which were subsequently incorporated in final models. All models included Time × Condition smooth interaction terms (k = 20 basis functions) to capture condition-specific temporal dynamics, along with gaze-position smooths to control for the relationship between pupil size measurements and eye position on the display. Table 3 presents the key model outputs; full model specifications appear in Supplementary Material S4.1; model diagnostics and ACF plots are presented in S4.2 (Figures S16–S18). Table 3 GAMM results for pupillometry across temporal windows. Section Modifier Object Spillover Term Est/edf SE/Ref.df t/F p Est/edf SE/Ref.df t/F p Est/edf SE/Ref.df t/F p Parametric Intercept 322.29 15.68 20.55 < .001 306.41 15.36 19.95 < .001 290.10 14.46 20.06 < .001 Cond: Incongruent −14.21 21.93 −0.65 .517 −18.80 21.49 −0.88 .382 −18.77 20.23 −0.93 .353 Cond: Neutral 8.93 21.82 0.41 .682 23.26 21.37 1.09 .277 24.25 20.12 1.21 .228 Smooth s(Time):Cond Cong 7.21 9.51 0.43 .881 4.66 6.32 0.45 .869 1.63 2.08 5.21 .004 s(Time):Cond Inc 1.01 1.01 0.21 .649 3.09 4.16 0.62 .601 10.03 12.21 3.65 < .001 s(Time):Cond Neut 12.86 14.58 2.15 .004 11.23 13.34 1.11 .356 12.28 13.96 2.21 .006 s(Xgaze, Ygaze) 28.95 29.00 807.93 < .001 28.92 29.00 576.79 < .001 28.90 29.00 1057.05 < .001 s(Time, Subject_Item) 4803.27 4899.00 2597.63 < .001 4809.06 4903.00 2495.71 < .001 4828.46 4902.00 2180.87 < .001 Parametric condition effects on mean pupil diameter were non-significant across all three temporal windows. In the modifier window, neither neutral (t = 1.21, p = .228) nor incongruent (t = 0.89, p = .372) conditions differed from the congruent baseline in overall dilation level. Similarly, object and spillover windows showed no significant parametric differences (all |t| ≤ 1.15, all p ≥ .251). The temporal dynamics of pupillary responses, captured by the Time × Condition smooth interaction terms, revealed condition-specific processing patterns that emerged progressively across the three windows. In the modifier window, only the neutral condition exhibited a significant time-varying effect (edf = 12.86, F = 2.15, p = .004). Congruent (edf = 7.21, F = 0.43, p = .881) and incongruent (edf = 1.01, F = 0.21, p = .649) conditions showed no significant temporal dynamics, with the low edf for incongruent (1.01, approaching linearity) suggesting minimal temporal structure. Pairwise comparisons confirmed no significant intervals of differentiation between any condition pair during the modifier window. Condition-contingent differences emerged during the object window, where linguistic disambiguation made it possible for the cognitive system to assess cue accuracy. The Time × Condition smooth terms remained non-significant for all three conditions individually (congruent: edf = 4.66, F = 0.45, p = .869; incongruent: edf = 3.09, F = 0.62, p = .601; neutral: edf = 11.23, F = 1.11, p = .356), but pairwise difference curves identified a sustained interval during which the neutral condition produced significantly greater pupil dilation than the incongruent condition (normalised time 0.333–0.899). No significant differences emerged for the congruent versus incongruent or congruent versus neutral comparisons during the object window (Fig. 4 , middle row). This pattern of cognitive load in the neutral condition exceeding that in the incongruent condition intensified during the spillover window. There, all three conditions exhibited significant Time × Condition smooth effects, indicating that cognitive load dynamics continued to evolve post-articulation. The congruent condition showed a significant but relatively simple smooth (edf = 1.63, F = 5.21, p = .004), the incongruent condition showed a more complex pattern (edf = 10.03, F = 3.65, p < .001), and the neutral condition showed the most complex temporal structure (edf = 12.28, F = 2.21, p = .006). Pairwise comparisons again showed significantly greater pupil dilation for neutral than incongruent (normalised time 0.141–0.697), replicating and extending the object-window finding into post-articulatory processing. No other pairwise differences reached significance in the spillover window (Fig. 4 , lower row). Discussion In this study we combined the VWP with pupillometry to investigate how congruent, incongruent, and neutral deictic gestures modulate attentional allocation and cognitive load in professional simultaneous interpreters. The fixation and pupillometric measures capture different facets of this process. While fixation patterns track the spatiotemporal direction of the referential search, pupil dilation reflects the temporal dynamics of the processing effort required to sustain that search [ 34 , 35 ]. The three gesture conditions produced distinct patterns in both fixation and pupil dilation. Congruent pointing guided anticipatory attention toward target referents. Incongruent gestures initially directed fixations toward competitor objects and produced transient increases in pupil dilation, but this conflict resolved as linguistic input accumulated. The neutral condition, in which no gestural cue was available, imposed the greatest cognitive load, exceeding even the cost associated with gestural incongruity. Congruent gestures produced target dominance already in the modifier window. Target fixations surpassed competitor fixations from the modifier onset and maintained this advantage through both the object and spillover intervals. This early bias reflects a process in which visual cues constrain the referential domain before linguistic disambiguation becomes available [ 24 , 44 ]. Concurrent verbal working memory load delays predictive eye movements in both L1 and L2 speakers without altogether suppressing them [ 45 ]. The nature of load, therefore, is not inconsequential. Whereas verbal working memory load disrupts phonological-level prediction, visual working memory load preserves semantic-level anticipation [ 46 ]. In our congruent condition, interpreters maintained anticipatory fixations throughout the task, indicating that anticipatory processing persisted under the concurrent demands of simultaneous interpreting. This observation is consistent with the pattern reported by Ito et al. [ 45 ], showing how prediction survived concurrent load, and with that of Liu et al. [ 46 ], suggesting that semantic-level anticipation is more resilient than phonological-level anticipation under resource constraints. By reducing referential ambiguity early in the unfolding sentence, congruent cues shortened the interval over which competitors remained active, consistent with models of situated comprehension in which visual scene information is incrementally recruited to constrain ongoing linguistic interpretation [ 47 ]. The facilitatory effect of congruent cues aligns with findings in adverse listening conditions, where gestures compensate for degraded auditory input [ 48 ]. Our results extend this compensatory function to simultaneous interpreting, where resource limitations arise from concurrent task demands. Interestingly, incongruent gestures produced a bifurcated fixation pattern. Initial fixations favoured the gestured competitor throughout the modifier window and into the early object interval, indicating that misleading gestural cues were automatically integrated into the unfolding referential interpretation. This obligatory influence of gesture on speech processing is consistent with evidence that the two modalities form a coupled system during comprehension in which integration occurs regardless of cue validity [ 6 ]. The time course of this erroneous bias mirrors established competitor dynamics in the VWP, where cohort activations rise from approximately 200 ms after target onset and decline to baseline by 500 ms as disambiguating input accumulates [ 22 ]. In our data, target recovery was complete by the spillover window. Incongruent target fixations then surpassed congruent fixations during this interval, a post-conflict rebound in which the resolution of cross-modal discrepancy produced higher target fixation proportions than conditions where no conflict arose. Electrophysiological evidence from multimodal discourse studies indicates that speech-gesture integration recruits visuospatial working memory resources, indexed by heightened alpha and beta power suppression as the cognitive system reconciles cross-modal discrepancies [ 49 ]. The visual interference observed here highlights the dominance of the visual modality during simultaneous interpreting, where interpreters frequently failed to override misleading written cues despite professional standards, a pattern analogous to the Colavita visual dominance effect, in which visual stimuli can take precedence over auditory input in detection tasks [ 50 ]. In the neutral condition, target dominance emerged only during noun articulation and showed limited consolidation in the spillover interval. Without gestural cues, disambiguation depended entirely on accumulating phonological evidence from the unfolding speech signal. This delayed resolution converges with several lines of evidence on the costs of unimodal processing. Sekicki and Staudte [ 3 ] found that the absence of referential gaze precluded early anticipatory eye movements, resulting in higher cognitive load as listeners relied on verbal constraints alone. Hostetter and Bahl [ 51 ] showed that prohibiting gesture elevated verbal cognitive load in descriptive tasks, and Gieshoff [ 52 ] documented that audio-only conditions prolonged processing in simultaneous interpreting relative to multimodal input containing visible lip movements. The contrast between neutral and incongruent conditions is interesting. Both conditions lack a veridical visual cue, yet only the incongruent condition provides a spatial reference that the unfolding speech can eventually override. The pupillometric findings provide converging evidence for the distinction between cue-present and cue-absent processing. During the modifier window, no reliable differences in pupil dilation emerged across conditions, and the condition-specific time courses were largely flat. This uniformity indicates that gestural cues do not modulate cognitive load until the disambiguating noun allows the cue–speech relationship to be evaluated, consistent with ERP evidence that gesture-speech integration effects are locked to the point of semantic disambiguation [ 53 ]. Across all three regions, no significant mean-level differences in pupil size emerged between conditions. The critical distinction lay instead in the temporal profile of the pupillary response. In the spillover window, the incongruent condition shifted from a near-flat trajectory during earlier windows to a markedly more dynamic pattern, possibly suggesting a delayed processing cost that surfaces only after the critical noun has passed. This temporal pattern suggests that cross-modal conflict incurs a processing cost that emerges not during the mismatch itself but during post-disambiguation reconciliation, as linguistic input overrides the misleading cue (cf. [ 54 ] for converging pupillemetric evidence that gesture-speech incongruence increases cognitive load). The neutral condition, by contrast, exhibited the most complex temporal dynamics across all windows together with the numerically largest mean dilation, consistent with evidence that auditory-only processing demands greater sustained working memory resources than multimodal encoding [ 37 ]. The congruent condition showed the simplest spillover trajectory, suggesting that a valid gestural cue reduces not only the magnitude but also the temporal complexity of the cognitive load associated with reference resolution [ 3 ]. The distinction, then, is not between high load and low load but between transient and sustained load. Cross-modal conflict generates a cost that terminates upon resolution, whereas unimodal processing without visual scaffolding imposes a maintenance cost that persists. The observed difference in temporal profiles between cue-present and cue-absent conditions is consistent with a cognitive offloading mechanism that connects several established theoretical frameworks. The Integrated-Systems Hypothesis posits that speech and gesture form a unified communicative system in which integration is obligatory, a view supported by evidence that gesture and speech recruit overlapping neural resources during comprehension [ 6 , 49 ]. Our temporal patterns are compatible with this account: both gesture conditions, whether congruent or incongruent, produced relatively simple pupillary trajectories during earlier processing windows, diverging from the more complex dynamics observed in the neutral condition. Dual-coding theory offers a complementary account: verbal and nonverbal systems operate as distinct but additive representational channels, distributing information across modalities and potentially reducing the load on any single channel [ 55 , 56 ]. In the neutral condition, the system operates within a single channel, where processing demands may accumulate as the candidate set is narrowed on phonological evidence alone [ 57 ]. At a mechanistic level, deictic gestures function as external pointers that bind objects to cognitive programs, allowing the processor to reference the environment directly rather than relying solely on internal representations [ 58 ]. When such external pointers are unavailable, referential maintenance falls more heavily on internal working memory [ 56 ]. These findings carry substantive implications for applied human language processing. While Effort Models of simultaneous interpreting have historically prioritised the management of auditory and memory constraints [ 59 ], our results suggest that deictic gestures are not merely secondary; rather, they actively modulate the attentional dynamics of referent resolution. The absence of these cues measurably increases sustained cognitive demands, a conclusion that converges with evidence that audio-only interpreting conditions precipitate longer silent pauses due to increased load [ 13 ]. Furthermore, the influence of the visual modality is evident in the "Colavita effect" observed during interpreting with text, where visual stimuli can override auditory inputs during conflict, leading interpreters to prioritise visual information even when it contradicts the audio source [ 50 ]. Consequently, these data validate professional standards advocating for direct visual access and underscore the cognitive risks of remote interpreting platforms: by restricting the visual channel, such platforms may inadvertently escalate fatigue by forcing the interpreter to rely on unimodal maintenance without the benefit of "virtual presence" [ 11 , 14 ]. Ultimately, this supports the broader principle that multimodal integration is the cognitive default, effectively reducing processing load in high-stakes communicative environments [ 48 ]. Several limitations warrant consideration. First, the sample comprised 24 professional conference interpreters, predominantly female, with French or Spanish as A languages and English as B or C language. This limits generalisability to other interpreter profiles and to broader populations who process multimodal speech under cognitive load [ 60 ]. Replication across populations in a broader community would clarify whether the observed offloading patterns are expertise-dependent or general. Second, the study focused exclusively on deictic pointing gestures, yet speakers produce iconic, beat, and non-manual cues that may engage different integration mechanisms [ 61 ]. Extending the paradigm to these cue types would inform models of multimodal comprehension beyond the referential function of pointing. Third, pupillometry captures only the temporal envelope of processing effort, not its neural substrates. Complementary methods, such as EEG time-frequency analysis [ 49 ] could isolate the stages of conflict detection and referential updating. Fourth, the laboratory setting used scripted video stimuli with a fixed spatial layout that may have increased gestural salience relative to naturalistic conditions [ 12 , 50 ], and future designs employing live interlocutors, varied spatial configurations, and competing visual distractors would test whether the cue-presence advantage generalises beyond controlled settings to the multimodal environments in which language comprehension ordinarily occurs. Methods Participants. Twenty-four professional conference interpreters working in Geneva were recruited from among staff interpreters at the United Nations Office in Geneva and members of the International Association of Conference Interpreters (AIIC) (22 women, two men; mean age = 41.7 yrs, SD = 10.8). Participants' A language was either French (n = 12) or Spanish (n = 12), with English as a working language. Professional experience averaged 11.6 years (SD = 9.0). All participants had normal or corrected-to-normal vision. The gender distribution (92% female) reflects the demographic composition of conference interpreting, where women constitute approximately 70–80% of the profession. Ethical approval was obtained from the Ethics Committee of the Faculty of Translation and Interpreting at the University of Geneva (Approval Number: FTI-N./Réf. 33), in accordance with the Declaration of Helsinki and institutional human research guidelines. Informed consent was obtained prior to participation, following a briefing about the study's objectives, procedures, potential discomforts (e.g., mild ocular fatigue from eye-tracking), and withdrawal rights without repercussion. Participants received a 50 Swiss Franc supermarket voucher as compensation. Stimuli. Linguistic Stimuli. The experiment comprised 30 experimental and two practice items. Each item consisted of a critical sentence with a “Prepositional phrase + Subject + Verb + Qualifier + Object” structure (e.g.: “In the desert, the explorer sees a large camel”) paired with a filler sentence (to allow ear-voice span management) following the structure, “There are + Number + Noun + in the + Qualifier + Noun Phrase” (e.g.: “There are two deserts on the vast continent”) to allow for additional interpreting time. Verb-modifier and modifier-object collocations were verified using the British National Corpus. English-speaking participants recruited via Amazon Mechanical Turk ( N = 160; 40 per list) assessed four stimulus lists (32 sentences each) for plausibility. ANOVAs with post-hoc t -tests identified optimal item pairings; the two lowest-plausibility pairs served as practice trials. The resulting 64 sentences (sentence length M = 9.27 words, SD = 0.7) were distributed across two matched lists, with matched word frequency and length confirmed with t-tests. A second and confirmatory norming round was performed by 34 native English speakers (17 per list) to rate the plausibility on a 7-point Likert scale. Mean ratings (List A: M = 5.53, SD = 0.73; List B: M = 5.99, SD = 0.75) showed no significant difference between lists (Mann-Whitney U, p = .072). Visual Stimuli. Licensed images of the two objects named in each sentence (N = 64) were obtained from Adobe Stock. They were selected based on resolution levels and background colours. Picture naming agreement was assessed via LimeSurvey, with participants providing the first descriptor elicited by each image. Responses with minor orthographic variations and semantically equivalent alternatives sharing lemmatic or phonological onset characteristics were retained (e.g., bike/bicycle, stairway/staircase). As the initial Amazon Mechanical Turk sample (N = 50) yielded anomalous responses suggestive of automated image recognition tool usage, a second norming iteration explicitly prohibiting AI-assisted identification yielded 38 valid respondents. Final naming agreement was 93.22% (SD = 8.5%, range: 63.16%–100%)." Each trial presented a speaker on the left of the screen (covering two-thirds of the image) with two potential referents arranged vertically on the right, covering the remaining third (Fig. 5 ). A right-handed male American English speaker recorded all sentences in a soundproof studio under controlled lighting conditions. Pointing gestures were performed with the non-dominant (left) hand, and initiated at the modifier onset. The speaker kept looking forward throughout the trials. Sentence duration averaged 6.98 s ( SD = 0.66), with the critical part lasting 3.69 s ( SD = 0.40). Modifier onset occurred at M = 2.58 s (SD = 0.36) post-video onset. Silent frames extended the total duration to 18 s. Videos were recorded against a green screen, replaced with a 25% grey background at 1920 × 1080 resolution, 25 fps, and 44.1 kHz stereo audio. Areas of interest (AOIs) were defined around each object, the speaker's head, and hand movement regions (See Fig. 5 ) using SR Research EyeLink Experiment Builder (Version 2.5.90). Conditions and Counterbalancing. Three gestural conditions were implemented with 10 items each: (a) congruent, i.e., pointing to the target after the modifier onset; (b) incongruent, i.e., pointing to the competitor; and (c) neutral, i.e., forward gaze without gestures. Six versions of each item (3 conditions × 2 vertical positions) controlled for positional preferences. A Latin square design distributed versions across six lists, rotated separately within French and Spanish interpreter subgroups. Apparatus. Eye-tracking and pupillometric data were collected with an SR Research EyeLink Portable Duo desktop-mounted eye-tracker operating in Head Stabilised Tracking mode at a sampling rate of 1000Hz. Monocular recording tracked participants' left eye (right eye in one case due to superior calibration stability). Participants were seated approximately 60 cm from a monitor (1920 × 1080 pixels) on a height-adjustable desk; a headrest maintained a consistent eye-to-tracker distance, while a chinrest was removed to permit unconstrained articulation during interpretation. Lighting conditions were kept constant. Audio from video stimuli and participants' interpreted output were recorded simultaneously on separate tracks of a stereo wave file. Procedure. Participants provided informed consent and completed an anonymised background questionnaire assessing demographic information and interpreting experience. The eye-tracker was calibrated using a nine-point grid followed by validation. Instructions were delivered in English, supplemented by experimenter clarification. Two practice trials preceded experimental items, after which participants could ask questions. Each trial began with a drift correction, and audio began approximately 1000 ms after video onset, allowing visual preview of each scene. Participants simultaneously interpreted video narratives from English into their A language (French or Spanish) while viewing the screen. Trials concluded at the video offset. No performance feedback was provided. The experimenter monitored eye-tracking quality throughout, reminding participants to avoid closing their eyes between trials. Total experimental duration was approximately 25 minutes. Data preprocessing. Interpreting validation. Following data acquisition, audio files were split to isolate the interpretation track, which was subsequently batch-transcribed using a Python script interfacing with the Amazon Web Services (AWS) Transcribe API for French and Spanish audio. All automated transcriptions were manually verified against the source audio by the authors, who corrected recognition errors and annotated disfluencies to ensure alignment accuracy. The validation protocol was designed to exclude trials in which eye-movement data might be confounded by misinterpretation or temporally misaligned deliveries. Trials were excluded in the case of: substantive lexical substitutions altering the proposition (e.g., "robots have fictitious markets" for "robots predict future markets"); omissions or mistranslations of the introductory locative, which shift the eye-tracking timeline by eliminating the temporal anchor; loss of the grammatical Subject, precluding determination of whether picture or utterance guided fixations; omissions or mistranslations of the critical Object; confusion of target with competitor referent (e.g., "monkey" for "otter"); severely incomplete, inaudible, or truncated sentences indicating temporal misalignment with the stimulus; or structural paraphrases conflating clauses or introducing extraneous information. Modifier adjectives (e.g., "expensive," "mischievous") functioned as fillers; therefore, their omission alone did not trigger exclusion as they were not central to referential resolution. Minor syntactic reordering or synonymous wording preserving source semantics was accepted, accommodating natural variation in interpreting styles. This validation stage yielded 674 of 720 trials (93.6%) for subsequent preprocessing. Fixation Data Preprocessing. Gaze data were inspected in SR Research Data Viewer (Version 4.4.1) and preprocessed in R (v 4.3.2 [ 62 ]) using VWPre (v 1.2.4 [ 29 ]). Onset and offset timestamps for linguistically defined segments (Context, Subject, Verb, Modifier, Object, and numeral-containing follow-up sentence) were extracted from AWS-generated transcripts, converted to message markers, and written to raw EDF files for millisecond-accurate temporal segmentation. Samples were mapped to four rectangular Areas of Interest (AOIs): Picture-up, Picture-down, Hand, and Head (see Fig. 5 ). The two picture AOIs were subsequently reclassified as Target or Competitor based on condition-specific position arrangements. Centroid coordinates for target and competitor pictures were extracted for distance calculations. Critical sentences were segmented into three analytical windows: Modifier (qualifier region), Object (noun region), and Spillover (noun offset to extra sentence onset), with + 200 ms adjustment for oculomotor delays. Valid gaze data percentages were computed for each time window by determining the proportion of samples with valid coordinates not marked as track loss (blinks, saccades, off-screen gaze). Trials with < 50% valid gaze data within critical windows were excluded. This threshold was determined through systematic evaluation of retention rates across 5% increments from 50% to 80%. Crucially, the multi-window design imposed a conjunctive constraint: each trial comprised three critical windows, and retention required all of them to simultaneously satisfy the threshold. At 80%, only 43% of trials survived; at 70%, 64%; whereas 50% preserved 84.2% of trials. We therefore selected 50% as the optimal balance between data quality and retention. Following validation and sparse-sample elimination, 606 of 720 trials (84.2%) were retained. For each 1-ms sampling interval, we computed: (i) binary indicators for gaze within Target or Competitor AOIs; (ii) Euclidean distance from current gaze position to Target and Competitor centroids; and (iii) lag-one versions of all binary and distance measures to capture temporal dependencies essential for GAMM autoregressive modelling [ 32 , 63 ]. As window durations varied across trials, time resampling was performed separately within each window using Python's decimate function, a downsampling method preserving signal integrity. For each window, the shortest series was identified, and all others resampled to match, producing uniform window-specific grids across Subject and Item. Samples were assigned normalised time indices (0 = window onset; 1 = window offset), enabling comparison on standardised proportional scales. Pupillometry Preprocessing. Raw pupil traces were processed in R (v 4.3.2 [ 62 ]) following the PupilPre pipeline (v 0.6.2 [ 64 ]) and protocols outlined by Mathôt and Vilotijević [ 65 ]. EyeLink sample reports were filtered to retain only trials that passed the interpreting validation. Timestamps were adjusted relative to "Context Onset" markers to establish baseline correction anchors. Off-screen samples (beyond monitor boundaries) were flagged. Blinks were replaced with missing values, including 50 ms padding on either side to account for partial blinks. Unmarked blinks and high-velocity artefacts were identified via a robust median-absolute-deviation procedure (100 ms padding, 2 × MAD threshold) and treated likewise; automatic corrections underwent manual review. Following van Rij [ 43 ], missing values were neither interpolated nor smoothed. Baseline correction subtracted the mean pupil diameter from the 500 ms interval preceding context onset from all subsequent measurements. Data were segmented into three linguistically defined epochs (Modifier, Object, Spillover) using embedded time markers with + 200 ms latency adjustment. Applying the same overall 50% validity threshold described above, requiring all four sub-windows (baseline plus three epochs) to simultaneously satisfy the criterion, yielded 545 of 720 trials (75.7%) for pupillometric analysis. Cleaned epoch-specific data were resampled using Python's decimate function following identical procedures to fixation preprocessing, producing uniform window-specific grids with normalised time indices (0–1) for cross-trial comparison. Statistical Analysis. All analyses employed generalised additive mixed models (GAMMs) via the mgcv package in R, selected for their capacity to accommodate non-linear, spatio-temporal dependencies and pronounced temporal autocorrelations inherent in high-resolution eye-tracking and pupillometric data [ 32 , 63 ]. Results are reported with effective degrees of freedom ( edf ), reference degrees of freedom ( Ref.df ), chi-square or F-statistics, and p -values for smooth terms. Target fixation across conditions. Binary target fixation outcomes were modelled with logistic GAMMs across three temporal windows (Modifier, Object, Spillover). Following Brown-Schmidt et al.[ 63 ] and Cho et al. [ 32 ], autocorrelation was addressed by incorporating the lag-one value of the response variable, accounting for the substantial influence of prior fixations on subsequent gaze positions. Lag-one distances between gaze position and Target/Competitor centroids were additionally integrated, capturing the dynamic role of spatial proximities in modulating gaze behaviour [ 63 ]. Model formulae included: Condition as parametric factor; Time-by-Condition tensor product smooths; and Subject × Item factor smooths for random effects. Models were fitted using bam() with discrete optimisation. Model diagnostics confirmed robust estimation: residual plots exhibited random scatter with no heteroscedasticity; ACF analyses revealed minimal persistence beyond lag 1; smooth term approximations confirmed appropriate basis dimensions ( k -index > 0.9). A tensor product interaction ti(Trial, Time) was additionally included to control for trial-by-time trends. Target Versus Competitor Fixation. Drawing upon the GAMM framework for deriving divergence points in VWP research [ 41 ], we modelled binary AOI outcomes (1 = target fixation, 0 = competitor fixation) to assess temporal competition dynamics. Models incorporated normalised time, a lagged fixation predictor ( ValueLag1 ) accounting for gaze inertia, smooth terms for temporal non-linearities, and random effects for Subject × Item clustering. Primary focus rested on the smooth term s(Time):RegionBin , capturing temporal divergence between target and competitor—a key metric for competition dynamics. Pairwise difference curves generated via plot_diff() identified temporal intervals of significant AOI divergence. All models demonstrated excellent fit (adjusted R ² ≥ 0.994; deviance explained ≥ 98.7%) with robust diagnostics ( k -index ≈ 0.97–0.99, p > 0.05; negligible residual autocorrelation). Pupillometric Analysis. Baseline-corrected pupil diameter was modelled with scaled-t ( scat ) family GAMMs to accommodate heavy-tailed distributions typical of pupillometric data [ 43 ]. Two-stage modelling addressed autocorrelation: initial models estimated autoregressive coefficients ( rho ), subsequently incorporated via the rho parameter with trial-start event markers for pre-whitening residuals (van Rij et al., 2019). Model formulae included: Condition as parametric factor; Time-by-Condition smooths ( k = 20); gaze position smooths s (Xgaze, Ygaze) controlling for pupil size changes associated with eye position; and Subject × Item factor smooths for random effects. Models were fitted using bam() with discrete optimisation and multi-threading for computational efficiency. Pairwise difference plots via plot_diff() identified temporal intervals of condition divergence. All models demonstrated excellent fit (adjusted R ² ≥ 0.989; deviance explained ≥ 90.7%) with adequate diagnostics ( k -index ≈ 0.78–1.00, p ≥ 0.065; minimal residual autocorrelation beyond lag 1). NotebookLM assisted with reference consolidation, AWS Transcribe enabled batch audio transcription and timestamp extraction, and Claude supported refactoring and optimisation of analytical scripts. Declarations Data and Code Availability The datasets and analysis code for this study are publicly available on the Open Science Framework (OSF) at https://osf.io/jcpv8. Acknowledgements The authors thank Edward A. Gibson for his advice on collecting data via Amazon Mechanical Turk. Author Contributions D.P. and K.G.S. designed the study. K.G.S. verified verbal and visual stimuli and contacted participants for online norming and lab studies. D.P. collected and analysed the data. D.P. wrote the first draft. K.G.S. revised the manuscript. Both authors approved the final version. Competing Interests The authors declare no competing interests. Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. References Özyürek, A., Willems, R. M., Kita, S. & Hagoort, P. On-line integration of semantic information from speech and gesture: insights from event-related brain potentials. J. Cogn. Neurosci. 19 , 605–616 (2007). McNeill, D. Speech and gesture integration. New Directions for Child and Adolescent Development 11–27 (1998). (1998). Sekicki, M. & Staudte, M. Eye'll Help You Out! How the Gaze Cue Reduces the Cognitive Load Required for Reference Processing. Cogn. Sci. 42 , 2418–2458 (2018). Arikan, G., Boddy, P. & Coventry, K. R. The relative importance of language, gaze, and gesture in deictic reference. J. Experimental Psychology: Learn. Memory Cognition . 51 , 1662–1681 (2025). Bangerter, A. Using pointing and describing to achieve joint focus of attention in dialogue. Psychol. Sci. 15 , 415–419 (2004). Kelly, S. D., Özyürek, A. & Maris, E. Two sides of the same coin: speech and gesture mutually interact to enhance comprehension. Psychol. Sci. 21 , 260–267 (2010). Wu, Y. C. & Coulson, S. How iconic gestures enhance communication: an ERP study. Brain Lang. 101 , 234–245 (2007). Hostetter, A. B. When do gestures communicate? A meta-analysis. Psychol. Bull. 137 , 297–315 (2011). Hervais-Adelman, A. G. & Babcock, L. The neurobiology of simultaneous interpreting: Where extreme language control and cognitive control intersect. Bilingualism 23 , 740–751 (2020). Christoffels, I. K., de Groot, A. M. B. & Kroll, J. F. Memory and language skills in simultaneous interpreters: The role of expertise and language proficiency☆. J. Mem. Lang. 54 , 324–345 (2006). Rennert, S. Visual Input in Simultaneous Interpreting. Meta 53 , 204–217 (2008). Seeber, K. G. Multimodal Processing in Simultaneous Interpreting in The Handbook of Translation and Cognition (eds Schwieter, J. W. & Ferreira, A.) 461–475 (Wiley, (2017). Gieshoff, A. C. The impact of visible lip movements on silent pauses in simultaneous interpreting. Interpreting 23 , 168–191 (2021). Moser-Mercer, B. Remote Interpreting: Issues of Multi-Sensory Integration in a Multilingual Task. Meta 50 , 727–738 (2005). AIIC. AIIC Guidelines on Distance Interpreting. (2025). Dargue, N., Sweller, N. & Jones, M. P. When our hands help us understand: A meta-analysis into the effects of gesture on comprehension. Psychol. Bull. 145 , 765–784 (2019). Macoun, A. & Sweller, N. Listening and watching: The effects of observing gesture on preschoolers’ narrative comprehension. Cogn. Dev. 40 , 68–81 (2016). Seeber, K. G. Multimodal Input in Simultaneous Interpreting: An eye-tracking experiment in Proceedings of the 1st International Conference TRANSLATA, Translation & Interpreting Research: yesterday - today - tomorrow. (ed. Zybatow, L., Petrova, A. & Ustaszewski, M.) 343–347Peter Lang, (2012). Huettig, F., Rommers, J. & Meyer, A. S. Using the visual world paradigm to study language processing: A review and critical evaluation. Acta. Psychol. 137 , 151–171 (2011). Tanenhaus, M. K., Spivey, M. J., Eberhard, K. M. & Sedivy, J. C. Integration of visual and linguistic information in spoken language comprehension. Science 268 , 1632–1634 (1995). Cooper, R. M. The control of eye fixation by the meaning of spoken language. Cogn. Psychol. 6 , 84–107 (1974). Allopenna, P. D., Magnuson, J. S. & Tanenhaus, M. K. Tracking the Time Course of Spoken Word Recognition Using Eye Movements: Evidence for Continuous Mapping Models. J. Mem. Lang. 38 , 419–439 (1998). Kamide, Y., Altmann, G. T. M. & Haywood, S. L. The time-course of prediction in incremental sentence processing: Evidence from anticipatory eye movements. J. Mem. Lang. 49 , 133–156 (2003). Altmann, G. T. M. & Kamide, Y. Incremental interpretation at verbs: restricting the domain of subsequent reference. Cognition 73 , 247–264 (1999). Prasad, S., Puri, S., Kapiley, K., Rafeekh, R. & Mishra, R. Looking Without Knowing: Evidence for Language-Mediated Eye Movements to Masked Words in Hindi-English Bilinguals. Languages 10 , 32 (2025). Fridman, C. & Meir, N. Dynamics of competition and co-activation in trilingual lexical processing: An eye-tracking study. Bilingualism , 1–16 (2025). Abashidze, D., Schmidt, A., Trofimovich, P. & Mercier, J. Integration of visual context in early and late bilingual language processing: evidence from eye-tracking. Front. Psychol. 14 , 1113688 (2023). Wieling, M. Analyzing dynamic phonetic data using generalized additive mixed modeling: A tutorial focusing on articulatory differences between L1 and L2 speakers of English. J. PHONETICS . 70 , 86–116 (2018). Porretta, V., Kyröläinen, A. J., van Rij, J. & Järvikivi, J. Visual World Paradigm Data: From Preprocessing to Nonlinear Time-Course Analysis in Intelligent Decision Technologies 2017 (ed. Czarnowski, I., Howlett, R. J. & Jain, L. C.) 268–277Springer International Publishing, (2018). Ito, A. & Knoeferle, P. Analysing data from the psycholinguistic visual-world paradigm: Comparison of different analysis methods. Behav. Res. Methods . 55 , 3461–3493 (2023). Baayen, H., Vasishth, S., Kliegl, R. & Bates, D. The cave of shadows: Addressing the human factor with generalized additive mixed models. J. Mem. Lang. 94 , 206–234 (2017). Cho, S. J., Brown-Schmidt, S., de Boeck, P. & Naveiras, M. Space-time modeling of intensive binary time series eye-tracking data using a generalized additive logistic regression model. Psychol. Methods . 27 , 307–346 (2022). Just, M. A. & Carpenter, P. A. The intensity dimension of thought: pupillometric indices of sentence processing. Can. J. experimental Psychol. = Revue canadienne de psychologie experimentale . 47 , 310–339 (1993). Beatty, J. Task-evoked pupillary responses, processing load, and the structure of processing resources. Psychol. Bull. 91 , 276–292 (1982). Kahneman, D. & Beatty, J. Pupil diameter and load on memory. Science 154 , 1583–1585 (1966). Engelhardt, P. E., Ferreira, F. & Patsenko, E. G. Pupillometry reveals processing load during spoken language comprehension. Q. J. Experimental Psychol. 63 , 639–645 (2010). Klingner, J., Tversky, B. & Hanrahan, P. Effects of visual and verbal presentation on cognitive load in vigilance, memory, and arithmetic tasks. Psychophysiology 48 , 323–332 (2011). Beatty-Martínez, A. L., Tamargo, G., Dussias, P. E. & R. E. & Phasic pupillary responses reveal differential engagement of attentional control in bilingual spoken language processing. Sci. Rep. 11 , 23474 (2021). Hyönä, J., Tommola, J. & Alaja, A. M. Pupil dilation as a measure of processing load in simultaneous interpretation and other language tasks. Q. J. Experimental Psychol. Sect. A-Human Experimental Psychol. 48 , 598–612 (1995). Seeber, K. G. & Kerzel, D. Cognitive load in simultaneous interpreting: Model meets data. Int. J. Biling. 16 , 228–242 (2012). Stone, K., Lago, S. & Schad, D. J. Divergence point analyses of visual world data: applications to bilingual research. Bilingualism 24 , 833–841 (2021). van Jacolien, M., Wieling, R. & Harald Baayen & Hedderik van Rijn. itsadug: Interpreting Time Series and Autocorrelated Data Using GAMMs (2025). van Rij, J., Hendriks, P., van Rijn, H., Baayen, R. H. & Wood, S. N. Analyzing the Time Course of Pupillometric Data. Trends Hear. 23 , 2331216519832483 (2019). Cheung, R. W., Hartley, C. & Monaghan, P. Better early than late: the temporal dynamics of pointing cues during cross-situational word learning. Lang. Cognition . 16 , 1960–1986 (2024). Ito, A., Corley, M. & Pickering, M. J. A cognitive load delays predictive eye movements similarly during L1 and L2 comprehension. Bilingualism 21 , 251–264 (2018). Liu, S., Hu, W. & Liu, X. Different effects of verbal and visual working memory loads on Language prediction. Sci. Rep. 15 , 22432 (2025). Crocker, M. W., Knoeferle, P. & Mayberry Situated sentence processing: The coordinated interplay account and a neurobehavioral model. Brain Lang. 112 , 189–201 (2010). Holler, J. & Levinson, S. C. Multimodal Language Processing in Human Communication. Trends Cogn. Sci. 23 , 639–652 (2019). Momsen, J., Gordon, J., Wu, Y. C. & Coulson, S. Event related spectral perturbations of gesture congruity: Visuospatial resources are recruited for multimodal discourse comprehension. Brain Lang. 216 , 104916 (2021). Chmiel, A., Janikowski, P. & Lijewska, A. Multimodal processing in simultaneous interpreting with text. Target 32 , 37–58 (2020). Hostetter, A. B. & Bahl, S. Comparing the cognitive load of gesture and action production: a dual-task study. Lang. Cognition . 15 , 601–621 (2023). Gieshoff, A. C. Does it help to see the speaker’s lip movements? An investigation of cognitive load and mental effort in simultaneous interpreting. TCB 4 , 1–25 (2021). Holle, H. & Gunter, T. C. The role of iconic gestures in speech disambiguation: ERP evidence. J. Cogn. Neurosci. 19 , 1175–1192 (2007). Morett, L. M., Roche, J. M., Fraundorf, S. H. & McPartland, J. C. Contrast Is in the Eye of the Beholder: Infelicitous Beat Gesture Increases Cognitive Load During Online Spoken Discourse Comprehension. Cognitive Science 44 , (2020). Clark, J. M. & Paivio, A. Dual coding theory and education. Educational Psychol. Rev. 3 , 149–210 (1991). Ping, R. & Goldin-Meadow, S. Gesturing saves cognitive resources when talking about nonpresent objects. Cogn. Sci. 34 , 602–619 (2010). Clough, S., Brown-Schmidt, S., Cho, S. J. & Duff, M. C. Reduced on-line speech gesture integration during multimodal language processing in adults with moderate-severe traumatic brain injury: Evidence from eye-tracking. Cortex 181 , 26–46 (2024). Ballard, D. H., Hayhoe, M. M., Pook, P. K. & Rao, R. P. Deictic codes for the embodiment of cognition. Behav. Brain Sci. 20, 723 – 42; discussion 743 – 67 (1997). Gile, D. Basic concepts and models for interpreter and translator training (John Benjamins, 2009). Dong, Y. & Li, P. Attentional control in interpreting: A model of language control and processing control. Bilingualism 23 , 716–728 (2020). Özyürek, A. Hearing and seeing meaning in speech and gesture: insights from brain and behaviour. Philos. Trans. R. Soc. Lond. B Biol. Sci. 369 , 20130296 (2014). R Core Team. R: A Language and Environment for Statistical Computing. (2025). Brown-Schmidt, S., Cho, S. J., Fenn, K. M. & Trude, A. M. Modeling spatio-temporal patterns in intensive binary time series eye-tracking data using Generalized Additive Mixed Models. Brain Research 1854 , (2025). Kyröläinen, A. J., Porretta, V., van Rij, J. & Järvikivi, J. PupilPre: Tools for Preprocessing Pupil Size Data (2019). Mathôt, S. & Vilotijević, A. Methods in cognitive pupillometry: Design, preprocessing, and statistical analysis. Behav. Res. Methods . 55 , 3055–3077 (2023). Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterials.pdf Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 19 May, 2026 Reviewers agreed at journal 04 May, 2026 Reviewers invited by journal 04 May, 2026 Editor invited by journal 08 Apr, 2026 Editor assigned by journal 02 Apr, 2026 Submission checks completed at journal 02 Apr, 2026 First submitted to journal 01 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9294880","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":625023722,"identity":"2b1ddc50-9860-49c7-b9d4-4a391728193c","order_by":0,"name":"Dongpeng Pan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+klEQVRIiWNgGAWjYBACezhL/gADwwcGBh4QmxmfFsMGGEsigYFxBjFaDA4gaWHmgbLxaznee/h1QcUduwaJ5GePbdvuyMi7NzB+LsCn5cy5NOsZZ54lN8gfMzfObXvGY3jmALP0DHxabuSYGfO2HU5mkGAwk85tO8xjOCOBDe5C3Fr+gbSwf5O2BGmZ/4CgFuPHvA2H7RgkeMykGYFa5CUY8Gsx7Dljxjzj2OEENgmeMsmec4d5DHgSm6XxabFn7zH+XFBz2J5fgn2bxI+yw/by7YcPfsanBQjYpIFEYhvcqQcYG/BrAMbbZwaUhENQwygYBaNgFIw0AADtbEmYfw7EaAAAAABJRU5ErkJggg==","orcid":"","institution":"University of Geneva","correspondingAuthor":true,"prefix":"","firstName":"Dongpeng","middleName":"","lastName":"Pan","suffix":""},{"id":625023726,"identity":"5bd0792d-095c-4c11-9f61-fdf98e3ee750","order_by":1,"name":"Kilian G. Seeber","email":"","orcid":"","institution":"University of Geneva","correspondingAuthor":false,"prefix":"","firstName":"Kilian","middleName":"G.","lastName":"Seeber","suffix":""}],"badges":[],"createdAt":"2026-04-01 17:38:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9294880/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9294880/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107455249,"identity":"ad4dad89-b1c5-4816-8df9-ef43bd7b4c4e","added_by":"auto","created_at":"2026-04-21 15:50:10","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3231261,"visible":true,"origin":"","legend":"\u003cp\u003eTime course of mean fixation proportions to four AOIs across the full sentence, plotted separately for congruent (right), incongruent (middle), and neutral (left) visual cue conditions. Solid lines depict mean fixation proportions to the target (orange), competitor (green), head (purple), and hand (blue); shaded ribbons indicate ±1 SE. Vertical dashed lines mark onsets of key events.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-9294880/v1/7093d6a320b7e758a20cb8c3.png"},{"id":107455254,"identity":"d9c27508-98ee-418c-be37-a2caaf6a14aa","added_by":"auto","created_at":"2026-04-21 15:50:10","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":10666610,"visible":true,"origin":"","legend":"\u003cp\u003eCondition effects on target fixation across sentence regions. GAM-estimated log-odds of target fixation for each condition pair across Modifier (columns 1–2), Object (columns 3–4), and Spillover (columns 5–6). Odd panels show overlaid smooth functions; even panels show the pairwise difference with 95% CI; red markers indicate significant intervals. Row 1: congruent vs. incongruent; Row 2: congruent vs. neutral; Row 3: incongruent vs. neutral.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-9294880/v1/1227cf6eb40c0d128de20c62.png"},{"id":107704455,"identity":"0fd370ef-10c2-4543-ab5a-3d44f448922a","added_by":"auto","created_at":"2026-04-24 08:45:28","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":12655016,"visible":true,"origin":"","legend":"\u003cp\u003eTarget–competitor fixation dynamics across sentence regions and visual cue conditions. Odd panels display GAMM-estimated log-odds of fixating the target (orange) and competitor (green) AOIs across Modifier, Object, and Spillover windows for congruent (Row 1), incongruent (Row 2), and neutral (Row 3) conditions. Even panels show the target–competitor difference with 95% CI; red markers indicate significant intervals.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-9294880/v1/aeaee4087ebabfc1a4e6ce4d.png"},{"id":107455253,"identity":"f9d9dc83-662d-4c63-82a3-bdbed9d6dbb4","added_by":"auto","created_at":"2026-04-21 15:50:10","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":7072215,"visible":true,"origin":"","legend":"\u003cp\u003ePupil size as a function of visual cue condition across sentence regions. GAMM-estimated smooth functions of pupil diameter over normalised time for Modifier (columns 1–2), Object (columns 3–4), and Spillover (columns 5–6). Row 1: congruent vs. incongruent; Row 2: congruent vs. neutral; Row 3: incongruent vs. neutral. Even panels show pairwise differences with 95% CI; red markers indicate significant intervals. No reliable condition differences appear in the Modifier window.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-9294880/v1/f96c970bdeafb65bc099f8cc.png"},{"id":107488992,"identity":"42401b64-16fb-4c44-bc61-c28002f12a4f","added_by":"auto","created_at":"2026-04-22 02:46:23","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":5859654,"visible":true,"origin":"","legend":"\u003cp\u003eSample frames from the three cueing conditions for the critical sentence \"In the desert, the explorer sees a large camel.\" In the congruent condition (a), the speaker begins pointing to the camel at the onset of the qualifier \"large\"; in the incongruent condition (b), the speaker instead points to the cactus; and in the neutral condition (c), the speaker speaks without gestures. The two pictures on the right display the potential referents. Throughout the three conditions, the speaker looks forward. The speaker's eyes are covered in the figure to protect identity; participants could see the speaker's eyes during the experiment in the video. Dashed blue rectangles indicate the pre-defined areas of interest (AOIs) configured in SR Research Experiment Builder for post-hoc eye-tracking analysis; these boundaries were not visible to participants during the experiment.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-9294880/v1/dda5d0ecfdbd35c6034f7767.png"},{"id":107708467,"identity":"0e0b95a3-6433-4abb-afc4-d6938d591a7f","added_by":"auto","created_at":"2026-04-24 09:27:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":33678359,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9294880/v1/40ae1ca1-98bf-464b-bc32-572dd45b2b38.pdf"},{"id":107455251,"identity":"8a384500-5f10-4035-82de-d46ceddd54f8","added_by":"auto","created_at":"2026-04-21 15:50:10","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":4781303,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9294880/v1/a310d74adf0ec2378179c8f8.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"The Point of Pointing: Deictic Gestures Modulate Attentional Shifts and Cognitive Load in Simultaneous Interpreting","fulltext":[{"header":"Introduction","content":"\u003cp\u003eLanguage processing involves integrating speech with visual signals [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Beyond the speech signal, communicators deploy visual cues such as eye gaze and hand movements that interlocutors integrate in real time to construct meaning and generate predictions about upcoming content [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Specifically, referential gaze and pointing gestures guide visual attention toward relevant objects, thereby grounding linguistic expressions in the physical environment [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Theoretical frameworks suggest that speech and gesture function as a system where the two modalities mutually and intrinsically influence one another during comprehension [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. At the same time, behavioural evidence indicates that listeners process congruent multimodal information more efficiently, whereas conflicting inputs increase processing costs [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Furthermore, meta-analytic reviews confirm that gestures benefit comprehension, particularly when they depict motor actions or convey information that supplements the spoken message [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. This multimodal integration appears critical in complex tasks like simultaneous interpreting, which represents an extreme form of bilingual language control requiring concurrent comprehension and production [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. In this cognitively demanding context, visual input such as pointing may act as aid that reduces cognitive load and facilitates the retrieval of meaning [\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Specifically, visual aids facilitate the prediction of message content, whereas the lack of these cues in many distance interpreting settings has been suggested to increase extraneous cognitive load and to contribute to fatigue [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAmong the various forms of nonverbal communication, deictic gestures, i.e., pointing actions that orient attention toward specific referents, fulfil a unique role by anchoring speech to the environment [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. While iconic gestures depict semantic attributes and beat gestures emphasise prosody, pointing provides referential precision. This facilitates target identification without imposing significant interpretive demands [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. There is substantial behavioural evidence identifying pointing as the dominant cue for referent selection, consistently serving as the strongest predictor of attention even when pitted against conflicting eye gaze or verbal demonstratives [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Viewers observing deictic gestures also demonstrate better retention than those watching beat gestures or no gestures, suggesting that pointing supports memory [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. During the cognitively demanding task of simultaneous interpreting, these gestures can function as complementary visual cues that can reduce the processing burden associated with auditory comprehension [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe VWP provides a solid methodological framework to explore how multimodal cues modulate real-time language processing [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. By tracking eye movements as participants listen to spoken language while viewing visual displays, the VWP enables the real-time measurement of incremental language processing with high temporal resolution [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Fixation probabilities, for example, align with the predictions of lexical activation models, revealing that words with similar onsets or rhymes compete during spoken word recognition [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Anticipatory eye movements based on information provided by the subject and/or verb emerge before the relevant noun is heard, revealing that listeners use linguistic and visual input to predict upcoming content [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. In bilingual and multilingual research, the VWP has allowed to uncover cross-linguistic co-activation and the role of visual cues in resolving competition between languages [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Fridman and Meir [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], for instance, used the VWP to show asymmetrical activation in trilingual speakers, where visual contexts, including deictic elements, accelerate fixation shifts and mitigate interference from dominant languages. Abashidze et al. [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] illustrated how deictic contexts counteract biases in bilingual prediction, fostering efficient resolution by guiding gaze to intended targets. Today, methodological advances, particularly generalised additive mixed models (GAMMs), enable researchers to capture the non-linear temporal dynamics in these fixation trajectories [\u003cspan additionalcitationids=\"CR29\" citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. This analytical approach accounts for the autocorrelation inherent in time-series data, offering a precise estimation of when effects arise during online processing [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eLanguage processing has also shown to mobilise physiological mechanisms that index the intensity of cognitive work [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. As a complement to eye-tracking data, pupillometry can be used to measure task-evoked pupillary responses (TEPRs) and thus quantify the allocation of mental effort on a moment-to-moment basis [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. The underlying theoretical rationale is that pupil diameter increases in response to processing demands, reflecting the consumption of limited attentional resources required for task performance [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. This metric has proved highly sensitive to linguistic complexity; for instance, sentences containing syntactic ambiguities or conflicting prosodic cues elicit greater dilation than their simpler counterparts, indicating the immediate cost of resolving structural conflict [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. The modality of input also modulates this response; auditory presentation of tasks typically induces larger dilations compared to visual presentation, potentially because visual formats support dual coding strategies that alleviate the burden on working memory [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. In multilingual contexts, pupillometry has been used to capture the cognitive cost of cross-linguistic competition, where bilinguals engaging in code-switching or processing a second language exhibit sustained dilation patterns [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. These findings are particularly relevant for simultaneous interpreting, a task that generates significantly higher pupillary responses compared to listening or shadowing due to the concurrent demands of comprehension and production [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. Furthermore, specific linguistic factors such as syntactic asymmetry between source and target languages trigger measurable increases in pupil diameter, providing objective evidence of the load associated with restructuring verbal input under time pressure [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe present study combines the VWP with pupillometry to investigate how congruent, incongruent, and neutral deictic gestures influence attention allocation and cognitive load in professional conference interpreters. We hypothesise that congruent pointing guides attention to the target referent [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], thereby reducing the cognitive load associated with reference resolution. Conversely, we predict that incongruent gestures will direct attention toward competitor objects [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], inducing a cognitive conflict that may produce distinct pupillary signatures reflecting cross-modal reconciliation. Finally, we anticipate that the neutral condition (without deictic gestures) will impose a greater processing load than congruent pointing, as the interpreter must rely exclusively on the auditory signal to construct meaning [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eFixation Dynamics Across Gesture Conditions\u003c/p\u003e \u003cp\u003eDescriptive fixation proportions revealed that interpreters predominantly directed gaze toward the speaker's head region throughout the sentence. Condition-specific modulations in object-directed fixations emerged after modifier onset, with divergence patterns indicating differential engagement with the visual display across gesture conditions (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Logistic generalised additive mixed models (GAMMs) modelling binary target fixation (target\u0026thinsp;=\u0026thinsp;1, non-target\u0026thinsp;=\u0026thinsp;0) quantified these condition effects across the three temporal windows (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Fitted smooth functions for all model terms are presented in Supplementary Material S2.1 (Figures \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u0026ndash;S3); model diagnostics and autocorrelation function (ACF) plots appear in S2.2 (Figures S4\u0026ndash;S6).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDuring the modifier window, where the adjective preceded the disambiguating noun, a strong baseline target preference was evident in the model intercept (β\u0026thinsp;=\u0026thinsp;9.41, SE\u0026thinsp;=\u0026thinsp;0.39, z\u0026thinsp;=\u0026thinsp;24.20, p \u0026lt; .001). Both neutral and incongruent gestures produced significant reductions in target fixation probability relative to the congruent baseline. The neutral condition reduced target fixation by approximately 1.10 log-odds units (β = \u0026minus;1.10, SE\u0026thinsp;=\u0026thinsp;0.38, z\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;2.89, p = .004), while incongruent gestures produced a somewhat larger reduction of 1.36 log-odds units (β = \u0026minus;1.36, SE\u0026thinsp;=\u0026thinsp;0.36, z\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;3.81, p \u0026lt; .001). Examination of the condition-specific smooth terms revealed that only the congruent condition exhibited significant time-varying dynamics (edf\u0026thinsp;=\u0026thinsp;2.98, χ\u0026sup2; = 2.90, p = .047). Neither neutral (edf\u0026thinsp;=\u0026thinsp;2.00, χ\u0026sup2; = 0.07, p = .965) nor incongruent conditions (edf\u0026thinsp;=\u0026thinsp;2.00, χ\u0026sup2; = 0.79, p = .674) showed significant temporal dynamics. Pairwise difference curves confirmed a congruent advantage over incongruent cues in early and late modifier intervals (normalised time 0.004\u0026ndash;0.487 and 0.688\u0026ndash;1.000) and over neutral cues similarly (0.004\u0026ndash;0.386 and 0.879\u0026ndash;1.000). No significant difference emerged between neutral and incongruent conditions at any point in the modifier window (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, Modifier column).\u003c/p\u003e \u003cp\u003eIn the object window, the overall target preference persisted (β\u0026thinsp;=\u0026thinsp;9.14, SE\u0026thinsp;=\u0026thinsp;0.31, z\u0026thinsp;=\u0026thinsp;29.25, p \u0026lt; .001) with attenuated but still significant condition effects. The neutral condition reduced target fixation by 0.67 log-odds units relative to congruent (β = \u0026minus;0.67, SE\u0026thinsp;=\u0026thinsp;0.21, z\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;3.26, p = .001), while incongruent gestures produced a similar reduction of 0.74 log-odds units (β = \u0026minus;0.74, SE\u0026thinsp;=\u0026thinsp;0.20, z\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;3.67, p \u0026lt; .001). Spatial lag terms showed increased complexity during this window (Time \u0026times; Lagged Distance to Target: edf\u0026thinsp;=\u0026thinsp;4.31, χ\u0026sup2; = 706.46, p \u0026lt; .001), suggesting intensified competition dynamics. Condition-specific time-varying effects were non-significant across all three conditions. Pairwise comparisons revealed that congruent cues maintained an advantage over incongruent cues during the first half of the window (normalised time 0.043\u0026ndash;0.486) and over neutral cues during the second half (0.668\u0026ndash;0.940) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, Object columns).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eGAMM results for fixation analysis: condition effects across temporal windows. Reference level: congruent. Parametric coefficients are on the log-odds scale.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"14\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eModifier\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eObject\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eSpillover\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTerm\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEst/edf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSE/Ref.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ez/χ\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eEst/edf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eSE/Ref.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003ez/χ\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eEst/edf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eSE/Ref.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003ez/χ\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eParametric\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIntercept\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e24.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e9.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e29.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e11.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e24.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCond: Neutral\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;1.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;2.89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e\u0026minus;0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026minus;3.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.631\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCond: Incongruent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;1.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;3.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e\u0026minus;0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026minus;3.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e0.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e3.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLagged Target Fix\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e26.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e3.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e33.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e3.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e42.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSmooth\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTime \u0026times; Prior Target Fix\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e9.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.011\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e7.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e8.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e18.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.014\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e4.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e5.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e19.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.003\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTime \u0026times; Lag Dist Target\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e289.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e4.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e5.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e706.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e3.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e4.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e1096.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTime \u0026times; Lag Dist Comp\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e68.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e4.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e5.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e56.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e2.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e2.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e38.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTime (Overall)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e3.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.052\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.949\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e1.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e2.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e3.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.277\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTrial Order\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e12.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e5.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e6.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e11.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.083\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e7.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e8.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e36.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eParticipant (RE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e12.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e23.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e15.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e6.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e23.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e44.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e10.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e23.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e193.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eItem (RE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e21.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e29.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e18.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e25.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e29.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e323.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e27.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e29.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e557.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTime \u0026times; Cond: Cong\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.047\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e3.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e4.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e1.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.926\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e3.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e3.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e6.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.161\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTime \u0026times; Cond: Neut\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.952\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e4.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e5.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e3.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.706\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e1.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e1.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e1.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.368\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTime \u0026times; Cond: Inc\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.462\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e4.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e5.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e1.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.950\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e2.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e3.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e5.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.184\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eDuring the spillover window capturing the speaker's sentence offset to capture gaze patterns following the end of auditory input, the target preference reached its highest magnitude (β\u0026thinsp;=\u0026thinsp;11.03, SE\u0026thinsp;=\u0026thinsp;0.45, z\u0026thinsp;=\u0026thinsp;24.39, p \u0026lt; .001). Critically, the pattern of condition effects reversed compared to earlier windows. The neutral condition no longer differed significantly from congruent (β\u0026thinsp;=\u0026thinsp;0.08, SE\u0026thinsp;=\u0026thinsp;0.17, z\u0026thinsp;=\u0026thinsp;0.48, p = .631). In contrast, incongruent gestures produced a significant enhancement of target fixation relative to the congruent baseline (β\u0026thinsp;=\u0026thinsp;0.49, SE\u0026thinsp;=\u0026thinsp;0.16, z\u0026thinsp;=\u0026thinsp;3.12, p = .002). Pairwise comparisons confirmed this incongruent advantage over both congruent (normalised time 0.053\u0026ndash;0.184 and 0.648\u0026ndash;0.960) and neutral conditions (0.496\u0026ndash;0.980, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, Spillover columns).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTarget\u0026ndash;Competitor Competition Dynamics\u003c/p\u003e \u003cp\u003eTo characterise within-condition competition independent of between-condition differences, we modelled binary area-of-interest (AOI) outcomes restricted to target and competitor fixations (target\u0026thinsp;=\u0026thinsp;1, competitor\u0026thinsp;=\u0026thinsp;0), following the analytical approach of Stone et al. [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. This analysis focused on the temporal divergence term \u003cem\u003es(Time):RegionBin\u003c/em\u003e, which captures the time-course of target versus competitor differentiation within each condition and window. Pairwise difference curves generated via the \u003cem\u003eplot_diff()\u003c/em\u003e function (\u003cem\u003eitsadug\u003c/em\u003e package [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]) identified intervals where fixations to target and competitor significantly diverged from equality. All intervals are expressed in normalised time units (0\u0026ndash;1 within each window). Full model specifications are described in Supplementary Material S3.1; complete parametric and smooth term results for all nine models appear in S3.2 (Tables S1\u0026ndash;S9); model diagnostics and ACF plots are presented in S3.3 (Figures S7\u0026ndash;S15).\u003c/p\u003e \u003cp\u003eWhen gestures were congruent, target dominance emerged rapidly and persisted across all three windows. In the modifier window, the temporal divergence term was significant (edf\u0026thinsp;=\u0026thinsp;2.98, χ\u0026sup2; = 27.85, p \u0026lt; .001), with target fixations exceeding competitor fixations across a sustained interval spanning normalised time 0.034\u0026ndash;0.849, encompassing the vast majority of the pre-disambiguation window. In the object window, target dominance was again significant (edf\u0026thinsp;=\u0026thinsp;3.10, χ\u0026sup2; = 38.72, p \u0026lt; .001), with target exceeding competitor from normalised time 0.164 through window end (1.000). In the spillover window, target dominance was significant (edf\u0026thinsp;=\u0026thinsp;2.00, χ\u0026sup2; = 24.19, p \u0026lt; .001) across normalised time 0.003\u0026ndash;0.930(Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, upper row).\u003c/p\u003e \u003cp\u003eWhen gestures were incongruent, a biphasic trajectory emerged. In the modifier window, the temporal divergence term was significant (edf\u0026thinsp;=\u0026thinsp;2.00, χ\u0026sup2; = 19.85, p \u0026lt; .001), but critically, the direction was reversed compared with the congruent condition. Here, competitor fixations exceeded target fixations across nearly the entire window (normalised time 0.004\u0026ndash;0.950). In the object window, the divergence term remained significant (edf\u0026thinsp;=\u0026thinsp;2.46, χ\u0026sup2; = 28.50, p \u0026lt; .001) but revealed a crossover pattern. The competitor initially exceeded target during normalised time 0.003\u0026ndash;0.335, followed by target exceeding competitor from 0.627\u0026ndash;1.000 as disambiguating noun input accumulated. By the spillover window, recovery was complete: the divergence term was highly significant (edf\u0026thinsp;=\u0026thinsp;2.00, χ\u0026sup2; = 56.68, p \u0026lt; .001), with target exceeding competitor across the full window (0.003\u0026ndash;1.000, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, middle row).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eGAMM results for target-competitor divergence across conditions and temporal windows.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"14\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eCondition\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eModifier\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eObject\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eSpillover\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eTerm\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eedf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRef.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eχ\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eedf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eRef.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eχ\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eedf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eRef.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003eχ\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eCongruent\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):ValueLag1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.168\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e2423.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e12.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):RegionBin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e27.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e3.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e3.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e38.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e24.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.336\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e8.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.005\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e4.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.039\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eIncongruent\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):ValueLag1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1052.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e2902.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e3.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.070\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):RegionBin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e19.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e28.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e56.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.302\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.794\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.334\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eNeutral\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):ValueLag1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e10.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e1806.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e2.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e3.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e12.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.013\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):RegionBin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.107\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e7.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.028\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e3.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e4.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e23.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.097\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e2.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.104\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.377\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eWhen gestures were absent, processing was delayed. In the modifier window, the temporal divergence term failed to reach significance (edf\u0026thinsp;=\u0026thinsp;2.00, χ\u0026sup2; = 4.47, p = .107). A late-emerging tendency toward competitor dominance (normalised time 0.708\u0026ndash;0.970) was observed but did not reach significance. In the object window, significant divergence emerged (edf\u0026thinsp;=\u0026thinsp;2.00, χ\u0026sup2; = 7.12, p = .028) but was restricted to the latter half of the window (normalised time 0.486\u0026ndash;1.000). In the spillover window, the divergence term was significant (edf\u0026thinsp;=\u0026thinsp;3.75, χ\u0026sup2; = 23.32, p \u0026lt; .001). (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e lower row).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eCognitive Load: Pupillometric Evidence\u003c/p\u003e \u003cp\u003eBaseline-corrected pupil diameter served as a continuous index of cognitive load dynamics across gesture conditions. Pupillometric data were modelled using GAMMs with scaled-t family distributions to accommodate the heavy-tailed characteristics typical of pupil size data [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. A two-stage modelling approach addressed temporal autocorrelation: initial models estimated autoregressive coefficients (\u003cem\u003erho\u003c/em\u003e), which were subsequently incorporated in final models. All models included Time \u0026times; Condition smooth interaction terms (k\u0026thinsp;=\u0026thinsp;20 basis functions) to capture condition-specific temporal dynamics, along with gaze-position smooths to control for the relationship between pupil size measurements and eye position on the display. Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents the key model outputs; full model specifications appear in Supplementary Material S4.1; model diagnostics and ACF plots are presented in S4.2 (Figures S16\u0026ndash;S18).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eGAMM results for pupillometry across temporal windows.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"14\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eSection\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eModifier\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eObject\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eSpillover\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eTerm\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEst/edf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSE/Ref.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003et/F\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eEst/edf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eSE/Ref.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003et/F\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eEst/edf\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eSE/Ref.df\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003et/F\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eParametric\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIntercept\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e322.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e15.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e20.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e306.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e15.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e19.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e290.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e14.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e20.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCond: Incongruent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026minus;14.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e21.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026minus;0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.517\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e\u0026minus;18.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e21.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e\u0026minus;0.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.382\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e\u0026minus;18.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e20.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e\u0026minus;0.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.353\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCond: Neutral\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e21.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.682\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e23.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e21.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e1.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.277\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e24.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e20.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e1.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.228\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eSmooth\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):Cond Cong\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e9.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.881\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e4.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e6.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.869\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e1.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e2.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e5.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.004\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):Cond Inc\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.649\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e3.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e4.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.601\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e10.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e12.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e3.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time):Cond Neut\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e12.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e14.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e11.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e13.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e1.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e.356\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e12.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e13.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e2.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e.006\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Xgaze, Ygaze)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e28.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e29.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e807.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e28.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e29.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e576.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e28.90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e29.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e1057.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003es(Time, Subject_Item)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4803.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4899.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2597.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e4809.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e4903.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e2495.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e4828.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e4902.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e2180.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eParametric condition effects on mean pupil diameter were non-significant across all three temporal windows. In the modifier window, neither neutral (t\u0026thinsp;=\u0026thinsp;1.21, p = .228) nor incongruent (t\u0026thinsp;=\u0026thinsp;0.89, p = .372) conditions differed from the congruent baseline in overall dilation level. Similarly, object and spillover windows showed no significant parametric differences (all |t| \u0026le; 1.15, all p \u0026ge; .251). The temporal dynamics of pupillary responses, captured by the Time \u0026times; Condition smooth interaction terms, revealed condition-specific processing patterns that emerged progressively across the three windows. In the modifier window, only the neutral condition exhibited a significant time-varying effect (edf\u0026thinsp;=\u0026thinsp;12.86, F\u0026thinsp;=\u0026thinsp;2.15, p = .004). Congruent (edf\u0026thinsp;=\u0026thinsp;7.21, F\u0026thinsp;=\u0026thinsp;0.43, p = .881) and incongruent (edf\u0026thinsp;=\u0026thinsp;1.01, F\u0026thinsp;=\u0026thinsp;0.21, p = .649) conditions showed no significant temporal dynamics, with the low edf for incongruent (1.01, approaching linearity) suggesting minimal temporal structure. Pairwise comparisons confirmed no significant intervals of differentiation between any condition pair during the modifier window.\u003c/p\u003e \u003cp\u003eCondition-contingent differences emerged during the object window, where linguistic disambiguation made it possible for the cognitive system to assess cue accuracy. The Time \u0026times; Condition smooth terms remained non-significant for all three conditions individually (congruent: edf\u0026thinsp;=\u0026thinsp;4.66, F\u0026thinsp;=\u0026thinsp;0.45, p = .869; incongruent: edf\u0026thinsp;=\u0026thinsp;3.09, F\u0026thinsp;=\u0026thinsp;0.62, p = .601; neutral: edf\u0026thinsp;=\u0026thinsp;11.23, F\u0026thinsp;=\u0026thinsp;1.11, p = .356), but pairwise difference curves identified a sustained interval during which the neutral condition produced significantly greater pupil dilation than the incongruent condition (normalised time 0.333\u0026ndash;0.899). No significant differences emerged for the congruent versus incongruent or congruent versus neutral comparisons during the object window (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, middle row).\u003c/p\u003e \u003cp\u003eThis pattern of cognitive load in the neutral condition exceeding that in the incongruent condition intensified during the spillover window. There, all three conditions exhibited significant Time \u0026times; Condition smooth effects, indicating that cognitive load dynamics continued to evolve post-articulation. The congruent condition showed a significant but relatively simple smooth (edf\u0026thinsp;=\u0026thinsp;1.63, F\u0026thinsp;=\u0026thinsp;5.21, p = .004), the incongruent condition showed a more complex pattern (edf\u0026thinsp;=\u0026thinsp;10.03, F\u0026thinsp;=\u0026thinsp;3.65, p \u0026lt; .001), and the neutral condition showed the most complex temporal structure (edf\u0026thinsp;=\u0026thinsp;12.28, F\u0026thinsp;=\u0026thinsp;2.21, p = .006). Pairwise comparisons again showed significantly greater pupil dilation for neutral than incongruent (normalised time 0.141\u0026ndash;0.697), replicating and extending the object-window finding into post-articulatory processing. No other pairwise differences reached significance in the spillover window (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, lower row).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study we combined the VWP with pupillometry to investigate how congruent, incongruent, and neutral deictic gestures modulate attentional allocation and cognitive load in professional simultaneous interpreters. The fixation and pupillometric measures capture different facets of this process. While fixation patterns track the spatiotemporal direction of the referential search, pupil dilation reflects the temporal dynamics of the processing effort required to sustain that search [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. The three gesture conditions produced distinct patterns in both fixation and pupil dilation. Congruent pointing guided anticipatory attention toward target referents. Incongruent gestures initially directed fixations toward competitor objects and produced transient increases in pupil dilation, but this conflict resolved as linguistic input accumulated. The neutral condition, in which no gestural cue was available, imposed the greatest cognitive load, exceeding even the cost associated with gestural incongruity.\u003c/p\u003e \u003cp\u003eCongruent gestures produced target dominance already in the modifier window. Target fixations surpassed competitor fixations from the modifier onset and maintained this advantage through both the object and spillover intervals. This early bias reflects a process in which visual cues constrain the referential domain before linguistic disambiguation becomes available [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]. Concurrent verbal working memory load delays predictive eye movements in both L1 and L2 speakers without altogether suppressing them [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]. The nature of load, therefore, is not inconsequential. Whereas verbal working memory load disrupts phonological-level prediction, visual working memory load preserves semantic-level anticipation [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. In our congruent condition, interpreters maintained anticipatory fixations throughout the task, indicating that anticipatory processing persisted under the concurrent demands of simultaneous interpreting. This observation is consistent with the pattern reported by Ito et al. [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e], showing how prediction survived concurrent load, and with that of Liu et al. [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e], suggesting that semantic-level anticipation is more resilient than phonological-level anticipation under resource constraints. By reducing referential ambiguity early in the unfolding sentence, congruent cues shortened the interval over which competitors remained active, consistent with models of situated comprehension in which visual scene information is incrementally recruited to constrain ongoing linguistic interpretation [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. The facilitatory effect of congruent cues aligns with findings in adverse listening conditions, where gestures compensate for degraded auditory input [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]. Our results extend this compensatory function to simultaneous interpreting, where resource limitations arise from concurrent task demands.\u003c/p\u003e \u003cp\u003eInterestingly, incongruent gestures produced a bifurcated fixation pattern. Initial fixations favoured the gestured competitor throughout the modifier window and into the early object interval, indicating that misleading gestural cues were automatically integrated into the unfolding referential interpretation. This obligatory influence of gesture on speech processing is consistent with evidence that the two modalities form a coupled system during comprehension in which integration occurs regardless of cue validity [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. The time course of this erroneous bias mirrors established competitor dynamics in the VWP, where cohort activations rise from approximately 200 ms after target onset and decline to baseline by 500 ms as disambiguating input accumulates [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. In our data, target recovery was complete by the spillover window. Incongruent target fixations then surpassed congruent fixations during this interval, a post-conflict rebound in which the resolution of cross-modal discrepancy produced higher target fixation proportions than conditions where no conflict arose. Electrophysiological evidence from multimodal discourse studies indicates that speech-gesture integration recruits visuospatial working memory resources, indexed by heightened alpha and beta power suppression as the cognitive system reconciles cross-modal discrepancies [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]. The visual interference observed here highlights the dominance of the visual modality during simultaneous interpreting, where interpreters frequently failed to override misleading written cues despite professional standards, a pattern analogous to the Colavita visual dominance effect, in which visual stimuli can take precedence over auditory input in detection tasks [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn the neutral condition, target dominance emerged only during noun articulation and showed limited consolidation in the spillover interval. Without gestural cues, disambiguation depended entirely on accumulating phonological evidence from the unfolding speech signal. This delayed resolution converges with several lines of evidence on the costs of unimodal processing. Sekicki and Staudte [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e] found that the absence of referential gaze precluded early anticipatory eye movements, resulting in higher cognitive load as listeners relied on verbal constraints alone. Hostetter and Bahl [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e] showed that prohibiting gesture elevated verbal cognitive load in descriptive tasks, and Gieshoff [\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e] documented that audio-only conditions prolonged processing in simultaneous interpreting relative to multimodal input containing visible lip movements. The contrast between neutral and incongruent conditions is interesting. Both conditions lack a veridical visual cue, yet only the incongruent condition provides a spatial reference that the unfolding speech can eventually override.\u003c/p\u003e \u003cp\u003eThe pupillometric findings provide converging evidence for the distinction between cue-present and cue-absent processing. During the modifier window, no reliable differences in pupil dilation emerged across conditions, and the condition-specific time courses were largely flat. This uniformity indicates that gestural cues do not modulate cognitive load until the disambiguating noun allows the cue\u0026ndash;speech relationship to be evaluated, consistent with ERP evidence that gesture-speech integration effects are locked to the point of semantic disambiguation [\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e]. Across all three regions, no significant mean-level differences in pupil size emerged between conditions. The critical distinction lay instead in the temporal profile of the pupillary response. In the spillover window, the incongruent condition shifted from a near-flat trajectory during earlier windows to a markedly more dynamic pattern, possibly suggesting a delayed processing cost that surfaces only after the critical noun has passed. This temporal pattern suggests that cross-modal conflict incurs a processing cost that emerges not during the mismatch itself but during post-disambiguation reconciliation, as linguistic input overrides the misleading cue (cf. [\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e] for converging pupillemetric evidence that gesture-speech incongruence increases cognitive load). The neutral condition, by contrast, exhibited the most complex temporal dynamics across all windows together with the numerically largest mean dilation, consistent with evidence that auditory-only processing demands greater sustained working memory resources than multimodal encoding [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. The congruent condition showed the simplest spillover trajectory, suggesting that a valid gestural cue reduces not only the magnitude but also the temporal complexity of the cognitive load associated with reference resolution [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. The distinction, then, is not between high load and low load but between transient and sustained load. Cross-modal conflict generates a cost that terminates upon resolution, whereas unimodal processing without visual scaffolding imposes a maintenance cost that persists.\u003c/p\u003e \u003cp\u003eThe observed difference in temporal profiles between cue-present and cue-absent conditions is consistent with a cognitive offloading mechanism that connects several established theoretical frameworks. The Integrated-Systems Hypothesis posits that speech and gesture form a unified communicative system in which integration is obligatory, a view supported by evidence that gesture and speech recruit overlapping neural resources during comprehension [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]. Our temporal patterns are compatible with this account: both gesture conditions, whether congruent or incongruent, produced relatively simple pupillary trajectories during earlier processing windows, diverging from the more complex dynamics observed in the neutral condition. Dual-coding theory offers a complementary account: verbal and nonverbal systems operate as distinct but additive representational channels, distributing information across modalities and potentially reducing the load on any single channel [\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. In the neutral condition, the system operates within a single channel, where processing demands may accumulate as the candidate set is narrowed on phonological evidence alone [\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e]. At a mechanistic level, deictic gestures function as external pointers that bind objects to cognitive programs, allowing the processor to reference the environment directly rather than relying solely on internal representations [\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e]. When such external pointers are unavailable, referential maintenance falls more heavily on internal working memory [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThese findings carry substantive implications for applied human language processing. While Effort Models of simultaneous interpreting have historically prioritised the management of auditory and memory constraints [\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e], our results suggest that deictic gestures are not merely secondary; rather, they actively modulate the attentional dynamics of referent resolution. The absence of these cues measurably increases sustained cognitive demands, a conclusion that converges with evidence that audio-only interpreting conditions precipitate longer silent pauses due to increased load [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Furthermore, the influence of the visual modality is evident in the \"Colavita effect\" observed during interpreting with text, where visual stimuli can override auditory inputs during conflict, leading interpreters to prioritise visual information even when it contradicts the audio source [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. Consequently, these data validate professional standards advocating for direct visual access and underscore the cognitive risks of remote interpreting platforms: by restricting the visual channel, such platforms may inadvertently escalate fatigue by forcing the interpreter to rely on unimodal maintenance without the benefit of \"virtual presence\" [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Ultimately, this supports the broader principle that multimodal integration is the cognitive default, effectively reducing processing load in high-stakes communicative environments [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSeveral limitations warrant consideration. First, the sample comprised 24 professional conference interpreters, predominantly female, with French or Spanish as A languages and English as B or C language. This limits generalisability to other interpreter profiles and to broader populations who process multimodal speech under cognitive load [\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e]. Replication across populations in a broader community would clarify whether the observed offloading patterns are expertise-dependent or general. Second, the study focused exclusively on deictic pointing gestures, yet speakers produce iconic, beat, and non-manual cues that may engage different integration mechanisms [\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e]. Extending the paradigm to these cue types would inform models of multimodal comprehension beyond the referential function of pointing. Third, pupillometry captures only the temporal envelope of processing effort, not its neural substrates. Complementary methods, such as EEG time-frequency analysis [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e] could isolate the stages of conflict detection and referential updating. Fourth, the laboratory setting used scripted video stimuli with a fixed spatial layout that may have increased gestural salience relative to naturalistic conditions [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e], and future designs employing live interlocutors, varied spatial configurations, and competing visual distractors would test whether the cue-presence advantage generalises beyond controlled settings to the multimodal environments in which language comprehension ordinarily occurs.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cb\u003eParticipants.\u003c/b\u003e Twenty-four professional conference interpreters working in Geneva were recruited from among staff interpreters at the United Nations Office in Geneva and members of the International Association of Conference Interpreters (AIIC) (22 women, two men; mean age\u0026thinsp;=\u0026thinsp;41.7 yrs, SD\u0026thinsp;=\u0026thinsp;10.8). Participants' A language was either French (n\u0026thinsp;=\u0026thinsp;12) or Spanish (n\u0026thinsp;=\u0026thinsp;12), with English as a working language. Professional experience averaged 11.6 years (SD\u0026thinsp;=\u0026thinsp;9.0). All participants had normal or corrected-to-normal vision. The gender distribution (92% female) reflects the demographic composition of conference interpreting, where women constitute approximately 70\u0026ndash;80% of the profession.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eEthical approval\u003c/strong\u003e \u003cp\u003e was obtained from the Ethics Committee of the Faculty of Translation and Interpreting at the University of Geneva (Approval Number: FTI-N./R\u0026eacute;f. 33), in accordance with the Declaration of Helsinki and institutional human research guidelines. Informed consent was obtained prior to participation, following a briefing about the study's objectives, procedures, potential discomforts (e.g., mild ocular fatigue from eye-tracking), and withdrawal rights without repercussion. Participants received a 50 Swiss Franc supermarket voucher as compensation.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eStimuli.\u003c/b\u003e \u003cem\u003eLinguistic Stimuli.\u003c/em\u003e The experiment comprised 30 experimental and two practice items. Each item consisted of a critical sentence with a \u0026ldquo;Prepositional phrase\u0026thinsp;+\u0026thinsp;Subject\u0026thinsp;+\u0026thinsp;Verb\u0026thinsp;+\u0026thinsp;Qualifier\u0026thinsp;+\u0026thinsp;Object\u0026rdquo; structure (e.g.: \u0026ldquo;In the desert, the explorer sees a large camel\u0026rdquo;) paired with a filler sentence (to allow ear-voice span management) following the structure, \u0026ldquo;There are +\u0026thinsp;Number\u0026thinsp;+\u0026thinsp;Noun\u0026thinsp;+\u0026thinsp;in the +\u0026thinsp;Qualifier\u0026thinsp;+\u0026thinsp;Noun Phrase\u0026rdquo; (e.g.: \u0026ldquo;There are two deserts on the vast continent\u0026rdquo;) to allow for additional interpreting time.\u003c/p\u003e \u003cp\u003eVerb-modifier and modifier-object collocations were verified using the British National Corpus. English-speaking participants recruited via Amazon Mechanical Turk (\u003cem\u003eN\u003c/em\u003e\u0026thinsp;=\u0026thinsp;160; 40 per list) assessed four stimulus lists (32 sentences each) for plausibility. ANOVAs with post-hoc \u003cem\u003et\u003c/em\u003e-tests identified optimal item pairings; the two lowest-plausibility pairs served as practice trials. The resulting 64 sentences (sentence length M\u0026thinsp;=\u0026thinsp;9.27 words, SD\u0026thinsp;=\u0026thinsp;0.7) were distributed across two matched lists, with matched word frequency and length confirmed with t-tests.\u003c/p\u003e \u003cp\u003eA second and confirmatory norming round was performed by 34 native English speakers (17 per list) to rate the plausibility on a 7-point Likert scale. Mean ratings (List A: M\u0026thinsp;=\u0026thinsp;5.53, SD\u0026thinsp;=\u0026thinsp;0.73; List B: M\u0026thinsp;=\u0026thinsp;5.99, SD\u0026thinsp;=\u0026thinsp;0.75) showed no significant difference between lists (Mann-Whitney U, p = .072).\u003c/p\u003e \u003cp\u003e \u003cem\u003eVisual Stimuli.\u003c/em\u003e Licensed images of the two objects named in each sentence (N\u0026thinsp;=\u0026thinsp;64) were obtained from Adobe Stock. They were selected based on resolution levels and background colours. Picture naming agreement was assessed via LimeSurvey, with participants providing the first descriptor elicited by each image. Responses with minor orthographic variations and semantically equivalent alternatives sharing lemmatic or phonological onset characteristics were retained (e.g., bike/bicycle, stairway/staircase). As the initial Amazon Mechanical Turk sample (N\u0026thinsp;=\u0026thinsp;50) yielded anomalous responses suggestive of automated image recognition tool usage, a second norming iteration explicitly prohibiting AI-assisted identification yielded 38 valid respondents. Final naming agreement was 93.22% (SD\u0026thinsp;=\u0026thinsp;8.5%, range: 63.16%\u0026ndash;100%).\"\u003c/p\u003e \u003cp\u003eEach trial presented a speaker on the left of the screen (covering two-thirds of the image) with two potential referents arranged vertically on the right, covering the remaining third (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). A right-handed male American English speaker recorded all sentences in a soundproof studio under controlled lighting conditions. Pointing gestures were performed with the non-dominant (left) hand, and initiated at the modifier onset. The speaker kept looking forward throughout the trials.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSentence duration averaged 6.98 s (\u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.66), with the critical part lasting 3.69 s (\u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.40). Modifier onset occurred at M\u0026thinsp;=\u0026thinsp;2.58 s (SD\u0026thinsp;=\u0026thinsp;0.36) post-video onset. Silent frames extended the total duration to 18 s. Videos were recorded against a green screen, replaced with a 25% grey background at 1920 \u0026times; 1080 resolution, 25 fps, and 44.1 kHz stereo audio.\u003c/p\u003e \u003cp\u003eAreas of interest (AOIs) were defined around each object, the speaker's head, and hand movement regions (See Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e) using SR Research EyeLink Experiment Builder (Version 2.5.90).\u003c/p\u003e \u003cp\u003e \u003cb\u003eConditions and Counterbalancing.\u003c/b\u003e Three gestural conditions were implemented with 10 items each: (a) congruent, i.e., pointing to the target after the modifier onset; (b) incongruent, i.e., pointing to the competitor; and (c) neutral, i.e., forward gaze without gestures.\u003c/p\u003e \u003cp\u003eSix versions of each item (3 conditions \u0026times; 2 vertical positions) controlled for positional preferences. A Latin square design distributed versions across six lists, rotated separately within French and Spanish interpreter subgroups.\u003c/p\u003e \u003cp\u003e\u003cb\u003eApparatus.\u003c/b\u003e Eye-tracking and pupillometric data were collected with an SR Research EyeLink Portable Duo desktop-mounted eye-tracker operating in Head Stabilised Tracking mode at a sampling rate of 1000Hz. Monocular recording tracked participants' left eye (right eye in one case due to superior calibration stability). Participants were seated approximately 60 cm from a monitor (1920 \u0026times; 1080 pixels) on a height-adjustable desk; a headrest maintained a consistent eye-to-tracker distance, while a chinrest was removed to permit unconstrained articulation during interpretation. Lighting conditions were kept constant. Audio from video stimuli and participants' interpreted output were recorded simultaneously on separate tracks of a stereo wave file.\u003c/p\u003e \u003cp\u003e\u003cb\u003eProcedure.\u003c/b\u003e Participants provided informed consent and completed an anonymised background questionnaire assessing demographic information and interpreting experience. The eye-tracker was calibrated using a nine-point grid followed by validation. Instructions were delivered in English, supplemented by experimenter clarification.\u003c/p\u003e \u003cp\u003eTwo practice trials preceded experimental items, after which participants could ask questions. Each trial began with a drift correction, and audio began approximately 1000 ms after video onset, allowing visual preview of each scene. Participants simultaneously interpreted video narratives from English into their A language (French or Spanish) while viewing the screen. Trials concluded at the video offset. No performance feedback was provided. The experimenter monitored eye-tracking quality throughout, reminding participants to avoid closing their eyes between trials. Total experimental duration was approximately 25 minutes.\u003c/p\u003e \u003cp\u003e \u003cb\u003eData preprocessing.\u003c/b\u003e \u003cem\u003eInterpreting validation.\u003c/em\u003e Following data acquisition, audio files were split to isolate the interpretation track, which was subsequently batch-transcribed using a Python script interfacing with the Amazon Web Services (AWS) Transcribe API for French and Spanish audio. All automated transcriptions were manually verified against the source audio by the authors, who corrected recognition errors and annotated disfluencies to ensure alignment accuracy.\u003c/p\u003e \u003cp\u003eThe validation protocol was designed to exclude trials in which eye-movement data might be confounded by misinterpretation or temporally misaligned deliveries. Trials were excluded in the case of: substantive lexical substitutions altering the proposition (e.g., \"robots \u003cem\u003ehave fictitious\u003c/em\u003e markets\" for \"robots predict future markets\"); omissions or mistranslations of the introductory locative, which shift the eye-tracking timeline by eliminating the temporal anchor; loss of the grammatical Subject, precluding determination of whether picture or utterance guided fixations; omissions or mistranslations of the critical Object; confusion of target with competitor referent (e.g., \"monkey\" for \"otter\"); severely incomplete, inaudible, or truncated sentences indicating temporal misalignment with the stimulus; or structural paraphrases conflating clauses or introducing extraneous information. Modifier adjectives (e.g., \"expensive,\" \"mischievous\") functioned as fillers; therefore, their omission alone did not trigger exclusion as they were not central to referential resolution. Minor syntactic reordering or synonymous wording preserving source semantics was accepted, accommodating natural variation in interpreting styles. This validation stage yielded 674 of 720 trials (93.6%) for subsequent preprocessing.\u003c/p\u003e \u003cp\u003e\u003cem\u003eFixation Data Preprocessing.\u003c/em\u003e Gaze data were inspected in SR Research Data Viewer (Version 4.4.1) and preprocessed in R (v 4.3.2 [\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e]) using VWPre (v 1.2.4 [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]). Onset and offset timestamps for linguistically defined segments (Context, Subject, Verb, Modifier, Object, and numeral-containing follow-up sentence) were extracted from AWS-generated transcripts, converted to message markers, and written to raw EDF files for millisecond-accurate temporal segmentation.\u003c/p\u003e \u003cp\u003eSamples were mapped to four rectangular Areas of Interest (AOIs): Picture-up, Picture-down, Hand, and Head (see Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). The two picture AOIs were subsequently reclassified as Target or Competitor based on condition-specific position arrangements. Centroid coordinates for target and competitor pictures were extracted for distance calculations.\u003c/p\u003e \u003cp\u003eCritical sentences were segmented into three analytical windows: Modifier (qualifier region), Object (noun region), and Spillover (noun offset to extra sentence onset), with +\u0026thinsp;200 ms adjustment for oculomotor delays. Valid gaze data percentages were computed for each time window by determining the proportion of samples with valid coordinates not marked as track loss (blinks, saccades, off-screen gaze). Trials with \u0026lt;\u0026thinsp;50% valid gaze data within critical windows were excluded. This threshold was determined through systematic evaluation of retention rates across 5% increments from 50% to 80%. Crucially, the multi-window design imposed a conjunctive constraint: each trial comprised three critical windows, and retention required all of them to simultaneously satisfy the threshold. At 80%, only 43% of trials survived; at 70%, 64%; whereas 50% preserved 84.2% of trials. We therefore selected 50% as the optimal balance between data quality and retention.\u003c/p\u003e \u003cp\u003eFollowing validation and sparse-sample elimination, 606 of 720 trials (84.2%) were retained. For each 1-ms sampling interval, we computed: (i) binary indicators for gaze within Target or Competitor AOIs; (ii) Euclidean distance from current gaze position to Target and Competitor centroids; and (iii) lag-one versions of all binary and distance measures to capture temporal dependencies essential for GAMM autoregressive modelling [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAs window durations varied across trials, time resampling was performed separately within each window using Python's \u003cem\u003edecimate\u003c/em\u003e function, a downsampling method preserving signal integrity. For each window, the shortest series was identified, and all others resampled to match, producing uniform window-specific grids across Subject and Item. Samples were assigned normalised time indices (0\u0026thinsp;=\u0026thinsp;window onset; 1\u0026thinsp;=\u0026thinsp;window offset), enabling comparison on standardised proportional scales.\u003c/p\u003e \u003cp\u003e \u003cem\u003ePupillometry Preprocessing.\u003c/em\u003e Raw pupil traces were processed in R (v 4.3.2 [\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e]) following the PupilPre pipeline (v 0.6.2 [\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e]) and protocols outlined by Math\u0026ocirc;t and Vilotijević [\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e]. EyeLink sample reports were filtered to retain only trials that passed the interpreting validation. Timestamps were adjusted relative to \"Context Onset\" markers to establish baseline correction anchors.\u003c/p\u003e \u003cp\u003eOff-screen samples (beyond monitor boundaries) were flagged. Blinks were replaced with missing values, including 50 ms padding on either side to account for partial blinks. Unmarked blinks and high-velocity artefacts were identified via a robust median-absolute-deviation procedure (100 ms padding, 2 \u0026times; MAD threshold) and treated likewise; automatic corrections underwent manual review. Following van Rij [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e], missing values were neither interpolated nor smoothed.\u003c/p\u003e \u003cp\u003eBaseline correction subtracted the mean pupil diameter from the 500 ms interval preceding context onset from all subsequent measurements. Data were segmented into three linguistically defined epochs (Modifier, Object, Spillover) using embedded time markers with +\u0026thinsp;200 ms latency adjustment. Applying the same overall 50% validity threshold described above, requiring all four sub-windows (baseline plus three epochs) to simultaneously satisfy the criterion, yielded 545 of 720 trials (75.7%) for pupillometric analysis.\u003c/p\u003e \u003cp\u003eCleaned epoch-specific data were resampled using Python's \u003cem\u003edecimate\u003c/em\u003e function following identical procedures to fixation preprocessing, producing uniform window-specific grids with normalised time indices (0\u0026ndash;1) for cross-trial comparison.\u003c/p\u003e \u003cp\u003e \u003cb\u003eStatistical Analysis.\u003c/b\u003e All analyses employed generalised additive mixed models (GAMMs) via the \u003cem\u003emgcv\u003c/em\u003e package in R, selected for their capacity to accommodate non-linear, spatio-temporal dependencies and pronounced temporal autocorrelations inherent in high-resolution eye-tracking and pupillometric data [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e]. Results are reported with effective degrees of freedom (\u003cem\u003eedf\u003c/em\u003e), reference degrees of freedom (\u003cem\u003eRef.df\u003c/em\u003e), chi-square or F-statistics, and \u003cem\u003ep\u003c/em\u003e-values for smooth terms.\u003c/p\u003e \u003cp\u003e \u003cem\u003eTarget fixation across conditions.\u003c/em\u003e Binary target fixation outcomes were modelled with logistic GAMMs across three temporal windows (Modifier, Object, Spillover). Following Brown-Schmidt et al.[\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e] and Cho et al. [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e], autocorrelation was addressed by incorporating the lag-one value of the response variable, accounting for the substantial influence of prior fixations on subsequent gaze positions. Lag-one distances between gaze position and Target/Competitor centroids were additionally integrated, capturing the dynamic role of spatial proximities in modulating gaze behaviour [\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e]. Model formulae included: Condition as parametric factor; Time-by-Condition tensor product smooths; and Subject \u0026times; Item factor smooths for random effects. Models were fitted using \u003cem\u003ebam()\u003c/em\u003e with discrete optimisation. Model diagnostics confirmed robust estimation: residual plots exhibited random scatter with no heteroscedasticity; ACF analyses revealed minimal persistence beyond lag 1; smooth term approximations confirmed appropriate basis dimensions (\u003cem\u003ek\u003c/em\u003e-index\u0026thinsp;\u0026gt;\u0026thinsp;0.9). A tensor product interaction \u003cem\u003eti(Trial, Time)\u003c/em\u003e was additionally included to control for trial-by-time trends.\u003c/p\u003e \u003cp\u003e \u003cem\u003eTarget Versus Competitor Fixation.\u003c/em\u003e Drawing upon the GAMM framework for deriving divergence points in VWP research [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e], we modelled binary AOI outcomes (1\u0026thinsp;=\u0026thinsp;target fixation, 0\u0026thinsp;=\u0026thinsp;competitor fixation) to assess temporal competition dynamics. Models incorporated normalised time, a lagged fixation predictor (\u003cem\u003eValueLag1\u003c/em\u003e) accounting for gaze inertia, smooth terms for temporal non-linearities, and random effects for Subject \u0026times; Item clustering. Primary focus rested on the smooth term \u003cem\u003es(Time):RegionBin\u003c/em\u003e, capturing temporal divergence between target and competitor\u0026mdash;a key metric for competition dynamics. Pairwise difference curves generated via \u003cem\u003eplot_diff()\u003c/em\u003e identified temporal intervals of significant AOI divergence. All models demonstrated excellent fit (adjusted \u003cem\u003eR\u003c/em\u003e\u0026sup2; \u0026ge; 0.994; deviance explained\u0026thinsp;\u0026ge;\u0026thinsp;98.7%) with robust diagnostics (\u003cem\u003ek\u003c/em\u003e-index\u0026thinsp;\u0026asymp;\u0026thinsp;0.97\u0026ndash;0.99, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05; negligible residual autocorrelation).\u003c/p\u003e \u003cp\u003e \u003cem\u003ePupillometric Analysis.\u003c/em\u003e Baseline-corrected pupil diameter was modelled with \u003cem\u003escaled-t\u003c/em\u003e (\u003cem\u003escat\u003c/em\u003e) family GAMMs to accommodate heavy-tailed distributions typical of pupillometric data [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. Two-stage modelling addressed autocorrelation: initial models estimated autoregressive coefficients (\u003cem\u003erho\u003c/em\u003e), subsequently incorporated via the \u003cem\u003erho\u003c/em\u003e parameter with trial-start event markers for pre-whitening residuals (van Rij et al., 2019). Model formulae included: Condition as parametric factor; Time-by-Condition smooths (\u003cem\u003ek\u003c/em\u003e\u0026thinsp;=\u0026thinsp;20); gaze position smooths s\u003cem\u003e(Xgaze, Ygaze)\u003c/em\u003e controlling for pupil size changes associated with eye position; and Subject \u0026times; Item factor smooths for random effects. Models were fitted using \u003cem\u003ebam()\u003c/em\u003e with discrete optimisation and multi-threading for computational efficiency. Pairwise difference plots via \u003cem\u003eplot_diff()\u003c/em\u003e identified temporal intervals of condition divergence. All models demonstrated excellent fit (adjusted \u003cem\u003eR\u003c/em\u003e\u0026sup2; \u0026ge; 0.989; deviance explained\u0026thinsp;\u0026ge;\u0026thinsp;90.7%) with adequate diagnostics (\u003cem\u003ek\u003c/em\u003e-index\u0026thinsp;\u0026asymp;\u0026thinsp;0.78\u0026ndash;1.00, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026ge;\u0026thinsp;0.065; minimal residual autocorrelation beyond lag 1).\u003c/p\u003e \u003cp\u003eNotebookLM assisted with reference consolidation, AWS Transcribe enabled batch audio transcription and timestamp extraction, and Claude supported refactoring and optimisation of analytical scripts.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData and Code Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThe datasets and analysis code for this study are publicly available on the Open Science Framework (OSF) at\u0026nbsp;\u003c/strong\u003ehttps://osf.io/jcpv8.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors thank Edward A. Gibson for his advice on collecting data via Amazon Mechanical Turk.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eD.P. and K.G.S. designed the study. K.G.S. verified verbal and visual stimuli and contacted participants for online norming and lab studies. D.P. collected and analysed the data. D.P. wrote the first draft. K.G.S. revised the manuscript. Both authors approved the final version.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e\u0026Ouml;zy\u0026uuml;rek, A., Willems, R. M., Kita, S. \u0026amp; Hagoort, P. On-line integration of semantic information from speech and gesture: insights from event-related brain potentials. \u003cem\u003eJ. Cogn. Neurosci.\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e, 605\u0026ndash;616 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcNeill, D. Speech and gesture integration. \u003cem\u003eNew Directions for Child and Adolescent Development\u003c/em\u003e 11\u0026ndash;27 (1998). (1998).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSekicki, M. \u0026amp; Staudte, M. Eye'll Help You Out! How the Gaze Cue Reduces the Cognitive Load Required for Reference Processing. \u003cem\u003eCogn. Sci.\u003c/em\u003e \u003cb\u003e42\u003c/b\u003e, 2418\u0026ndash;2458 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArikan, G., Boddy, P. \u0026amp; Coventry, K. R. The relative importance of language, gaze, and gesture in deictic reference. \u003cem\u003eJ. Experimental Psychology: Learn. Memory Cognition\u003c/em\u003e. \u003cb\u003e51\u003c/b\u003e, 1662\u0026ndash;1681 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBangerter, A. Using pointing and describing to achieve joint focus of attention in dialogue. \u003cem\u003ePsychol. Sci.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 415\u0026ndash;419 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKelly, S. D., \u0026Ouml;zy\u0026uuml;rek, A. \u0026amp; Maris, E. Two sides of the same coin: speech and gesture mutually interact to enhance comprehension. \u003cem\u003ePsychol. Sci.\u003c/em\u003e \u003cb\u003e21\u003c/b\u003e, 260\u0026ndash;267 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, Y. C. \u0026amp; Coulson, S. How iconic gestures enhance communication: an ERP study. \u003cem\u003eBrain Lang.\u003c/em\u003e \u003cb\u003e101\u003c/b\u003e, 234\u0026ndash;245 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHostetter, A. B. When do gestures communicate? A meta-analysis. \u003cem\u003ePsychol. Bull.\u003c/em\u003e \u003cb\u003e137\u003c/b\u003e, 297\u0026ndash;315 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHervais-Adelman, A. G. \u0026amp; Babcock, L. The neurobiology of simultaneous interpreting: Where extreme language control and cognitive control intersect. \u003cem\u003eBilingualism\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 740\u0026ndash;751 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChristoffels, I. K., de Groot, A. M. B. \u0026amp; Kroll, J. F. Memory and language skills in simultaneous interpreters: The role of expertise and language proficiency☆. \u003cem\u003eJ. Mem. Lang.\u003c/em\u003e \u003cb\u003e54\u003c/b\u003e, 324\u0026ndash;345 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRennert, S. Visual Input in Simultaneous Interpreting. \u003cem\u003eMeta\u003c/em\u003e \u003cb\u003e53\u003c/b\u003e, 204\u0026ndash;217 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeeber, K. G. Multimodal Processing in Simultaneous Interpreting in The Handbook of Translation and Cognition (eds Schwieter, J. W. \u0026amp; Ferreira, A.) 461\u0026ndash;475 (Wiley, (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGieshoff, A. C. The impact of visible lip movements on silent pauses in simultaneous interpreting. \u003cem\u003eInterpreting\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 168\u0026ndash;191 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoser-Mercer, B. Remote Interpreting: Issues of Multi-Sensory Integration in a Multilingual Task. \u003cem\u003eMeta\u003c/em\u003e \u003cb\u003e50\u003c/b\u003e, 727\u0026ndash;738 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAIIC. AIIC Guidelines on Distance Interpreting. (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDargue, N., Sweller, N. \u0026amp; Jones, M. P. When our hands help us understand: A meta-analysis into the effects of gesture on comprehension. \u003cem\u003ePsychol. Bull.\u003c/em\u003e \u003cb\u003e145\u003c/b\u003e, 765\u0026ndash;784 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMacoun, A. \u0026amp; Sweller, N. Listening and watching: The effects of observing gesture on preschoolers\u0026rsquo; narrative comprehension. \u003cem\u003eCogn. Dev.\u003c/em\u003e \u003cb\u003e40\u003c/b\u003e, 68\u0026ndash;81 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeeber, K. G. Multimodal Input in Simultaneous Interpreting: An eye-tracking experiment in \u003cem\u003eProceedings of the 1st International Conference TRANSLATA, Translation \u0026amp; Interpreting Research: yesterday - today - tomorrow.\u003c/em\u003e (ed. Zybatow, L., Petrova, A. \u0026amp; Ustaszewski, M.) 343\u0026ndash;347Peter Lang, (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuettig, F., Rommers, J. \u0026amp; Meyer, A. S. Using the visual world paradigm to study language processing: A review and critical evaluation. \u003cem\u003eActa. Psychol.\u003c/em\u003e \u003cb\u003e137\u003c/b\u003e, 151\u0026ndash;171 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTanenhaus, M. K., Spivey, M. J., Eberhard, K. M. \u0026amp; Sedivy, J. C. Integration of visual and linguistic information in spoken language comprehension. \u003cem\u003eScience\u003c/em\u003e \u003cb\u003e268\u003c/b\u003e, 1632\u0026ndash;1634 (1995).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCooper, R. M. The control of eye fixation by the meaning of spoken language. \u003cem\u003eCogn. Psychol.\u003c/em\u003e \u003cb\u003e6\u003c/b\u003e, 84\u0026ndash;107 (1974).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllopenna, P. D., Magnuson, J. S. \u0026amp; Tanenhaus, M. K. Tracking the Time Course of Spoken Word Recognition Using Eye Movements: Evidence for Continuous Mapping Models. \u003cem\u003eJ. Mem. Lang.\u003c/em\u003e \u003cb\u003e38\u003c/b\u003e, 419\u0026ndash;439 (1998).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKamide, Y., Altmann, G. T. M. \u0026amp; Haywood, S. L. The time-course of prediction in incremental sentence processing: Evidence from anticipatory eye movements. \u003cem\u003eJ. Mem. Lang.\u003c/em\u003e \u003cb\u003e49\u003c/b\u003e, 133\u0026ndash;156 (2003).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAltmann, G. T. M. \u0026amp; Kamide, Y. Incremental interpretation at verbs: restricting the domain of subsequent reference. \u003cem\u003eCognition\u003c/em\u003e \u003cb\u003e73\u003c/b\u003e, 247\u0026ndash;264 (1999).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrasad, S., Puri, S., Kapiley, K., Rafeekh, R. \u0026amp; Mishra, R. Looking Without Knowing: Evidence for Language-Mediated Eye Movements to Masked Words in Hindi-English Bilinguals. \u003cem\u003eLanguages\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e, 32 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFridman, C. \u0026amp; Meir, N. Dynamics of competition and co-activation in trilingual lexical processing: An eye-tracking study. \u003cem\u003eBilingualism\u003c/em\u003e, 1\u0026ndash;16 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbashidze, D., Schmidt, A., Trofimovich, P. \u0026amp; Mercier, J. Integration of visual context in early and late bilingual language processing: evidence from eye-tracking. \u003cem\u003eFront. Psychol.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e, 1113688 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWieling, M. Analyzing dynamic phonetic data using generalized additive mixed modeling: A tutorial focusing on articulatory differences between L1 and L2 speakers of English. \u003cem\u003eJ. PHONETICS\u003c/em\u003e. \u003cb\u003e70\u003c/b\u003e, 86\u0026ndash;116 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePorretta, V., Kyr\u0026ouml;l\u0026auml;inen, A. J., van Rij, J. \u0026amp; J\u0026auml;rvikivi, J. Visual World Paradigm Data: From Preprocessing to Nonlinear Time-Course Analysis in \u003cem\u003eIntelligent Decision Technologies 2017\u003c/em\u003e (ed. Czarnowski, I., Howlett, R. J. \u0026amp; Jain, L. C.) 268\u0026ndash;277Springer International Publishing, (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIto, A. \u0026amp; Knoeferle, P. Analysing data from the psycholinguistic visual-world paradigm: Comparison of different analysis methods. \u003cem\u003eBehav. Res. Methods\u003c/em\u003e. \u003cb\u003e55\u003c/b\u003e, 3461\u0026ndash;3493 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaayen, H., Vasishth, S., Kliegl, R. \u0026amp; Bates, D. The cave of shadows: Addressing the human factor with generalized additive mixed models. \u003cem\u003eJ. Mem. Lang.\u003c/em\u003e \u003cb\u003e94\u003c/b\u003e, 206\u0026ndash;234 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCho, S. J., Brown-Schmidt, S., de Boeck, P. \u0026amp; Naveiras, M. Space-time modeling of intensive binary time series eye-tracking data using a generalized additive logistic regression model. \u003cem\u003ePsychol. Methods\u003c/em\u003e. \u003cb\u003e27\u003c/b\u003e, 307\u0026ndash;346 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJust, M. A. \u0026amp; Carpenter, P. A. The intensity dimension of thought: pupillometric indices of sentence processing. \u003cem\u003eCan. J. experimental Psychol. = Revue canadienne de psychologie experimentale\u003c/em\u003e. \u003cb\u003e47\u003c/b\u003e, 310\u0026ndash;339 (1993).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeatty, J. Task-evoked pupillary responses, processing load, and the structure of processing resources. \u003cem\u003ePsychol. Bull.\u003c/em\u003e \u003cb\u003e91\u003c/b\u003e, 276\u0026ndash;292 (1982).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKahneman, D. \u0026amp; Beatty, J. Pupil diameter and load on memory. \u003cem\u003eScience\u003c/em\u003e \u003cb\u003e154\u003c/b\u003e, 1583\u0026ndash;1585 (1966).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEngelhardt, P. E., Ferreira, F. \u0026amp; Patsenko, E. G. Pupillometry reveals processing load during spoken language comprehension. \u003cem\u003eQ. J. Experimental Psychol.\u003c/em\u003e \u003cb\u003e63\u003c/b\u003e, 639\u0026ndash;645 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKlingner, J., Tversky, B. \u0026amp; Hanrahan, P. Effects of visual and verbal presentation on cognitive load in vigilance, memory, and arithmetic tasks. \u003cem\u003ePsychophysiology\u003c/em\u003e \u003cb\u003e48\u003c/b\u003e, 323\u0026ndash;332 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeatty-Mart\u0026iacute;nez, A. L., Tamargo, G., Dussias, P. E. \u0026amp; R. E. \u0026amp; Phasic pupillary responses reveal differential engagement of attentional control in bilingual spoken language processing. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 23474 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHy\u0026ouml;n\u0026auml;, J., Tommola, J. \u0026amp; Alaja, A. M. Pupil dilation as a measure of processing load in simultaneous interpretation and other language tasks. \u003cem\u003eQ. J. Experimental Psychol. Sect. A-Human Experimental Psychol.\u003c/em\u003e \u003cb\u003e48\u003c/b\u003e, 598\u0026ndash;612 (1995).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeeber, K. G. \u0026amp; Kerzel, D. Cognitive load in simultaneous interpreting: Model meets data. \u003cem\u003eInt. J. Biling.\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e, 228\u0026ndash;242 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStone, K., Lago, S. \u0026amp; Schad, D. J. Divergence point analyses of visual world data: applications to bilingual research. \u003cem\u003eBilingualism\u003c/em\u003e \u003cb\u003e24\u003c/b\u003e, 833\u0026ndash;841 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Jacolien, M., Wieling, R. \u0026amp; Harald Baayen \u0026amp; Hedderik van Rijn. \u003cem\u003eitsadug: Interpreting Time Series and Autocorrelated Data Using GAMMs\u003c/em\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Rij, J., Hendriks, P., van Rijn, H., Baayen, R. H. \u0026amp; Wood, S. N. Analyzing the Time Course of Pupillometric Data. \u003cem\u003eTrends Hear.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 2331216519832483 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheung, R. W., Hartley, C. \u0026amp; Monaghan, P. Better early than late: the temporal dynamics of pointing cues during cross-situational word learning. \u003cem\u003eLang. Cognition\u003c/em\u003e. \u003cb\u003e16\u003c/b\u003e, 1960\u0026ndash;1986 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIto, A., Corley, M. \u0026amp; Pickering, M. J. A cognitive load delays predictive eye movements similarly during L1 and L2 comprehension. \u003cem\u003eBilingualism\u003c/em\u003e \u003cb\u003e21\u003c/b\u003e, 251\u0026ndash;264 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, S., Hu, W. \u0026amp; Liu, X. Different effects of verbal and visual working memory loads on Language prediction. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 22432 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrocker, M. W., Knoeferle, P. \u0026amp; Mayberry Situated sentence processing: The coordinated interplay account and a neurobehavioral model. \u003cem\u003eBrain Lang.\u003c/em\u003e \u003cb\u003e112\u003c/b\u003e, 189\u0026ndash;201 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoller, J. \u0026amp; Levinson, S. C. Multimodal Language Processing in Human Communication. \u003cem\u003eTrends Cogn. Sci.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 639\u0026ndash;652 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMomsen, J., Gordon, J., Wu, Y. C. \u0026amp; Coulson, S. Event related spectral perturbations of gesture congruity: Visuospatial resources are recruited for multimodal discourse comprehension. \u003cem\u003eBrain Lang.\u003c/em\u003e \u003cb\u003e216\u003c/b\u003e, 104916 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChmiel, A., Janikowski, P. \u0026amp; Lijewska, A. Multimodal processing in simultaneous interpreting with text. \u003cem\u003eTarget\u003c/em\u003e \u003cb\u003e32\u003c/b\u003e, 37\u0026ndash;58 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHostetter, A. B. \u0026amp; Bahl, S. Comparing the cognitive load of gesture and action production: a dual-task study. \u003cem\u003eLang. Cognition\u003c/em\u003e. \u003cb\u003e15\u003c/b\u003e, 601\u0026ndash;621 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGieshoff, A. C. Does it help to see the speaker\u0026rsquo;s lip movements? An investigation of cognitive load and mental effort in simultaneous interpreting. \u003cem\u003eTCB\u003c/em\u003e \u003cb\u003e4\u003c/b\u003e, 1\u0026ndash;25 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHolle, H. \u0026amp; Gunter, T. C. The role of iconic gestures in speech disambiguation: ERP evidence. \u003cem\u003eJ. Cogn. Neurosci.\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e, 1175\u0026ndash;1192 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorett, L. M., Roche, J. M., Fraundorf, S. H. \u0026amp; McPartland, J. C. Contrast Is in the Eye of the Beholder: Infelicitous Beat Gesture Increases Cognitive Load During Online Spoken Discourse Comprehension. \u003cem\u003eCognitive Science\u003c/em\u003e \u003cb\u003e44\u003c/b\u003e, (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClark, J. M. \u0026amp; Paivio, A. Dual coding theory and education. \u003cem\u003eEducational Psychol. Rev.\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e, 149\u0026ndash;210 (1991).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePing, R. \u0026amp; Goldin-Meadow, S. Gesturing saves cognitive resources when talking about nonpresent objects. \u003cem\u003eCogn. Sci.\u003c/em\u003e \u003cb\u003e34\u003c/b\u003e, 602\u0026ndash;619 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClough, S., Brown-Schmidt, S., Cho, S. J. \u0026amp; Duff, M. C. Reduced on-line speech gesture integration during multimodal language processing in adults with moderate-severe traumatic brain injury: Evidence from eye-tracking. \u003cem\u003eCortex\u003c/em\u003e \u003cb\u003e181\u003c/b\u003e, 26\u0026ndash;46 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBallard, D. H., Hayhoe, M. M., Pook, P. K. \u0026amp; Rao, R. P. Deictic codes for the embodiment of cognition. \u003cem\u003eBehav. Brain Sci.\u003c/em\u003e 20, 723\u0026thinsp;\u0026ndash;\u0026thinsp;42; discussion 743\u0026thinsp;\u0026ndash;\u0026thinsp;67 (1997).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGile, D. \u003cem\u003eBasic concepts and models for interpreter and translator training\u003c/em\u003e (John Benjamins, 2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDong, Y. \u0026amp; Li, P. Attentional control in interpreting: A model of language control and processing control. \u003cem\u003eBilingualism\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 716\u0026ndash;728 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e\u0026Ouml;zy\u0026uuml;rek, A. Hearing and seeing meaning in speech and gesture: insights from brain and behaviour. \u003cem\u003ePhilos. Trans. R. Soc. Lond. B Biol. Sci.\u003c/em\u003e \u003cb\u003e369\u003c/b\u003e, 20130296 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR Core Team. R: A Language and Environment for Statistical Computing. (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrown-Schmidt, S., Cho, S. J., Fenn, K. M. \u0026amp; Trude, A. M. Modeling spatio-temporal patterns in intensive binary time series eye-tracking data using Generalized Additive Mixed Models. \u003cem\u003eBrain Research\u003c/em\u003e \u003cb\u003e1854\u003c/b\u003e, (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKyr\u0026ouml;l\u0026auml;inen, A. J., Porretta, V., van Rij, J. \u0026amp; J\u0026auml;rvikivi, J. \u003cem\u003ePupilPre: Tools for Preprocessing Pupil Size Data\u003c/em\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMath\u0026ocirc;t, S. \u0026amp; Vilotijević, A. Methods in cognitive pupillometry: Design, preprocessing, and statistical analysis. \u003cem\u003eBehav. Res. Methods\u003c/em\u003e. \u003cb\u003e55\u003c/b\u003e, 3055\u0026ndash;3077 (2023).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"speech processing, simultaneous interpreting, deictic gestures, Visual World Paradigm, eye-tracking, cognitive load","lastPublishedDoi":"10.21203/rs.3.rs-9294880/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9294880/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSpoken language processing requires the integration of acoustic signals with visual information. Deictic gestures, such as pointing, may facilitate this integration by constraining the referential domain before linguistic disambiguation takes place. However, it remains unclear how multimodal cues modulate processing under increased cognitive load. Simultaneous interpreting provides a task environment for the examination of speech processing under substantial cognitive load. In this study, we combined the Visual World Paradigm (VWP) with pupillometry to investigate how congruent, incongruent, and neutral pointing gestures influence attention and cognitive load in 24 professional simultaneous interpreters. Eye-tracking revealed that congruent gestures elicited early anticipatory visual target fixations. Incongruent gestures directed gaze toward visual competitors before rapid correction, while the absence of gestures delayed visual target identification. Counterintuitively, the gesture-absent condition produced greater pupil dilation than the incongruent condition, while congruent and incongruent conditions did not differ significantly in cognitive load. These results challenge the assumption that cue accuracy drives processing cost and instead suggest that the presence of a visual cue, even a misleading one, is the critical variable that reduces cognitive demands.\u003c/p\u003e","manuscriptTitle":"The Point of Pointing: Deictic Gestures Modulate Attentional Shifts and Cognitive Load in Simultaneous Interpreting","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-21 15:50:01","doi":"10.21203/rs.3.rs-9294880/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"268019228570422235786728492091415650831","date":"2026-05-19T07:35:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"270393748273098339766209476558910707949","date":"2026-05-04T14:50:22+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-05-04T10:24:27+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-08T17:31:40+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-02T04:16:06+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-02T04:16:05+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-04-01T17:23:04+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"bfacde8a-bc67-4535-9c77-e58a97dc8978","owner":[],"postedDate":"April 21st, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"268019228570422235786728492091415650831","date":"2026-05-19T07:35:25+00:00","index":134,"fulltext":""},{"type":"reviewerAgreed","content":"270393748273098339766209476558910707949","date":"2026-05-04T14:50:22+00:00","index":114,"fulltext":""},{"type":"reviewersInvited","content":"42","date":"2026-05-04T10:24:27+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":66528788,"name":"Biological sciences/Neuroscience"},{"id":66528789,"name":"Biological sciences/Psychology"},{"id":66528790,"name":"Social science/Psychology"}],"tags":[],"updatedAt":"2026-05-04T10:38:24+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-21 15:50:01","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9294880","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9294880","identity":"rs-9294880","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}