Temporal contrast enhancement emerges from distinct pain and sound filtering mechanisms

20 The h uman brain continuously processes sensory inputs whose perceptual impact depends on 21 stimulus intensity, salience, and context. Temporal filtering shapes this processing by dynamically 22 modulating neural responses according to stimulus repetition , temporal structure, and predictive 23 contingencies. Higher-order percepts such as pain are likewise a subject to temporal filtering, as 24 exemplified by temporal contrast enhancement (TCE) , a stimulation paradigm eliciting pain 25 inhibition. However , g iven the prevalence of filtering mechanisms across sensory domains, it 26 remains unclear whether TCE is specific to pain or represents a supramodal filtering mechanism. 27 Here, we contrasted behavioral and neurophysiological responses, including 28 electroencephalography and pupillometry, in a TCE paradigm for painful heat versus 29 uncomfortable auditory stimulation . We sought to establish whether TCE generalizes across 30 modalities and bases on similar or distinct underlying neurophysiological processes. Human 31 participants took part in either a purely behavioral investigation (n = 33) or in a neurophysiological 32 and behavioral investigation (n = 29) . Both painful heat and loud sounds induced TCE effects, 33 suggesting a supramodal temporal filtering mechanism with modality-specific temporal dynamics. 34 Increasing pupil size and decreas ing power of neural alpha oscillations (~10 Hz) with higher 35 painful heat indicate bottom-up modulation of the autonomic nervous system and a release of 36 neural inhibition, respectively. Critically, expectation of pain – but not loud sound – induced an 37 alpha power increase, demonstrating to-down contributions to temporal filtering of pain. However, 38 no direct neurophysiological correlate s of subjectively experienced TCE effect s were found. 39 Findings suggests that a supramodal temporal filtering mechanism with modality -specific 40 neurophysiological dynamics shapes the processing of aversive stimulation. 41

Temporal Filtering, Temporal Contrast Enhancement, Offset Analgesia, Auditory 42 System, Pain, Sensory Processing 43 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 3 Abbreviations 44 ANS Autonomic nervous system 45 CT Constant trial 46 EEG Electroencephalography 47 eVAS Electronic visual analogue scale 48 FDR False discovery rate 49 fMRI Functional magnetic resonance imaging 50 OT Offset trial 51 TCE Temporal contrast enhancement 52 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 4

53 Human environments contain a plethora of sensory inputs. Sensory integration is a complex 54 process dependent on stimulus intensity, salience and context (1). Therefore, diverse filtering 55 mechanisms are required to selectively enhance specific aspects of stimuli, evaluate their 56 relevance, and modulate their perceptual processing (2,3). Commonly, these filtering mechanisms 57 are divided into spatial and temporal domains of perceived input , which can be observed in all 58 sensory systems (visual (4,5), auditory (6–8) tactile (9,10)). Well-adapted filtering mechanisms 59 help us make accurate and reliable judgements in constantly changing sensory environments (11). 60 Here, we test the hypothesis that temporal filtering in different modalities share s a common 61 perceptual and neurophysiological basis. 62 Convergences in temporal filtering have been documented across sensory domains. In the 63 context of auditory perception, temporal summation - alternatively referred to as temporal 64 integration (12) or temporal sound summation (13) - describes the phenomena of a perceived 65 loudness increase when a tone of constant volume is presented for a longer duration. Temporal 66 filtering of auditory percepts is of utmost importance, due to its relevance in vocalization and 67 speech recognition (14), as well as in avoiding uncomfortably loud sound . Temporal summation 68 likewise characterizes nociceptive perception (15), where adaptive temporal filtering is vital for 69 averting tissue injury and protecting the organism. Additional similarities can be observed in 70 filtering processes resulting in adaptation to perceived stimulus inputs. Such phenomena can be 71 observed in (acute) pain or auditory adaptation . When an individual is exposed to a constant 72 stimulus within the appropriate context and depending on the stimulus parameters, a reduction in 73 the perceived intensity of the stimulus can be observed (11,16,17). By contrast, when context 74 signals threat, through salience, magnitude, or affective associations, temporal filtering can 75 amplify perception indicated by pain intensification (17,18) or temporal sound summation (13), 76 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 5 highlighting the importance of these factors. Additional current evidence indicates temporal 77 filtering processes share similar mechanisms across modalities and sensory systems (19–22). 78 However, despite their high relevance, the extent to which temporal filtering in auditory and pain 79 processing is based on common perceptual and physiological mechanisms is not yet fully 80 understood. 81 Temporal contrast enhancement (TCE), a putative component of the endogenous pain 82 modulatory system, is conceptualized as a temporal filtering process whose mechanisms remain 83 unclear (23–25). Commonly referred to as offset analgesia, is characterized by a disproportionally 84 large reduction in perceived pain in response to a slight decrease in noxious stimulus intensity(25–85 27). A variety of mechanisms distributed along the neuroaxis have been postulated as potential 86 mediators of TCE . Several brain imaging studies using functional magnetic resonance imaging 87 (fMRI) showed increased activity of cortical structures such as the primary (24,25,28) and 88 secondary (28) somatosensory cortex, ventromedial (25) and dorsolateral (23,24) prefrontal cortex 89 and subcortical structures such as the putamen and nucleus accumbens (23), insula (24,25) and the 90 spinal cord (29) as evidenced by greater blood oxygenation levels during the TCE interval 91 compared to a control condition. Additionally, structures of the brainstem previously linked to 92 descending pain modulation such as the periaqueductal grey and rostro -ventromedial medulla 93 showed increased activity during TCE (25,28,30). 94 Interestingly, studies investigating the temporal dynamics of TCE using 95 electroencephalography (EEG) seem to be lacking in the field. General investigations of both tonic 96 and phasic pain observed in EEG have been conducted, with the main focus being oscillations in 97 the alpha range ( ~10Hz), showing a prominent pattern of reduced alpha power during tonic and 98 phasic painful stimuli (31–33). This might relate to the proposed functional inhibition of situational 99 irrelevant information (34–38), which is well-established by alpha power modulation in auditory 100 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 6 attention research (39–41). Additional temporally sensitive and objective measures, such as 101 pupillometry, have been used to characterize neurophysiological correlates of pain and auditory 102 processing. Changes in pupil diameter are closely related to autonomic nervous system activity 103 and have been linked to perceived stimulus intensity and arousal, particularly in response to painful 104 stimuli (42,43). In auditory research, pupil dilation has similarly been associated with loudness of 105 the stimuli, salience, and aversiveness (44–46), highlighting its relevance for tracking perceptual 106 and attentional processes. 107 Here, we hypothesize that the effects of TCE on noxious stimuli may not reflect pain-108 specific processes but rather could be indicative of a general temporal filtering mechanism in 109 response to salient stimuli . To determine whether observed effects in pain -research are modality 110 specific or merely reflect general arousal, prior studies have employed salience-matched control 111 stimuli; nevertheless the evidence still remains equivocal and the issue remains unresolved (31,47–112 50). We therefore conducted a multimodal experiment to determine whether TCE is specific to 113 pain. This was achieved by applying painful heat or uncomfortable auditory stimulation using a 114 typical TCE paradigm. The investigation encompassed behavioral and neurophysiological 115 responses (EEG, pupillometry), with the objective of improving our understanding of temporal 116 filtering mechanisms. 117

118 A total of 62 healthy volunteers were recruited in this study (behavioral investigation: n = 119 33, behavioral + neurophysiological investigation: n = 29). All participants tolerated the selected 120 stimulus intensities without difficulty and exhibited no signs of adverse events (see S1 and S2 121 Tables for details) . Figure 1 illustrates the study design. Both investigations used a similar 122 experimental approach to investigate behavioral responses to the TCE paradigm. Thus, the 123 following analysis consists of the combined behavioral data of both experiments. An overview of 124 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 7 stimulation parameters , as well as additional analyses for each individual experiment can be 125 accessed in the Materials and Methods Section and the Supporting Information (S3 – S5 Tables). 126 127 Figure 1. Study design. Schematic representation of the study design of both experiments . After preparation, participants 128 underwent a training & familiarization phase to accustom to stimuli and the operation of the electronic visual analogue scale 129 (eVAS). In Experiment 1 (top of the figure), stimulus intensities were individually calibrated. The experimental procedure was split 130 into two blocks (1 per modality) consisting of 8 trials with offset (OT) or constant (CT) trials in pseudorandomized sequence . In 131 each trial, participants rated their sensory experience using the eVAS . A two-minute pause was conducted after each trial and a 132 five-minute break was conducted after switching to the other modality . In Experiment 2 (bottom of the figure) fixed stimulus 133 intensities were used. The experimental procedure was split into four blocks (2 per modality) consisting of 10 trials with offset (OT) 134 or constant (CT) trials in pseudorandomized sequence. Each trial was followed by a two -minute break, and after each modality 135 block participants had a five-minute break. The first two trials of each block were fixed to include an OT and CT (being randomized 136 in order) and were used to collect the behavioral response. The remaining 8 trials did not include any eVAS ratings, only collecting 137 neurophysiological responses. 138 Temporal contrast enhancement is robust in heat-induced pain and 139 aversive auditory stimulation 140 Participants were either presented with a sine -tone via headphones (1000 Hz) or with a 141 thermal heat stimulus on the left forearm using a thermal contact stimulator. Stimulation intensity 142 was either constant across a duration of 35 s (constant trial) or increased in the time interval 143 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 8 between 10 and 20 s ( offset trial; Figure 2). First, we tested whether participants’ ratings of 144 experienced pain (heat) or discomfort (sound) would vary as a function of time interval (T1: 5-10 145 s, T2: 15-20 s, T3: 25-30 s), trial type (OT: offset trial, CT: constant trial) and modality (heat, 146 auditory). Given a significant main effect of modality and significant interactions involving 147 modality (see S3 Table for details), separate ANOVAs were conducted for each modality. 148 149 Figure 2: Behavioral Temporal Contrast Enhancement (TCE) response in two modalities. Means and standard errors of the 150 mean (SEM) for behavioral responses are displayed as obtained by an electronic visual analogue scale (eVAS). Thermal stimulation 151 (A) and auditory stimulation (B) both induced significant (*, p < .05) TCE effects in T3 when comparing offset (OT) (darker colors) 152 and constant (CT) (lighter colors) trials. The correlation of TCE effects (C) (CT-OT at T3) (green = behavioral experiment, black 153 = neurophysiological experiment) between both modalities was not significant. Boxplots of eVAS ratings (D) for each of the fo ur 154 conditions split into three relevant time intervals (T1: 5 -10s, T2: 15 -20s, T3: 25 -30s). Significant (p < .05) within -modality 155 comparisons indicated by solid lines, between-modality comparisons indicated by dashed lines. 156 ANOVAs revealed interactions of time interval x trial type for auditory (F (2, 122) = 157 102.39, p < 0.001,  p2 = 0.63) and thermal (F (2, 122) =186.55, p < 0.001,  p2 = 0.75) stimulation. 158 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 9 Comparing trials using false discovery rate (FDR) -corrected post -hoc testing for auditory 159 stimulation revealed no difference between OT and CT during T1 (p = 0.71). However, we found 160 significantly higher discomfort ratings for OT vs. CT in T 2 (p < 0.0 01) and a reversal of this 161 difference in T3 (p < 0.0 01), indicating that a return to the stimulus intensity of T1 induced a 162 significant TCE effect in the interval of interest (T3). Similarly, post hoc testing for thermal 163 stimulation showed no difference between OT and CT during T1(p = 0.30), but significantly higher 164 pain ratings for OT vs. CT in T2 (p < 0.0 01) and a reversal of this difference in T3 (p < 0.001). 165 These findings demonstrate that TCE – a significant and disproportionate reduction in perceived 166 intensity following a brief decrease in stimulus intensity – occurs for both, painful heat and 167 aversive auditory stimulation. 168 Having established behavioral TCE effects for thermal and auditory stimulation, we next 169 tested the association between the two. Across participants, TCE effects (i.e. the OT - CT 170 difference in T3) were not significantly correlated between auditory and thermal stimulation (r = 171 0.15, p = 0.25; BF₁₀ = 0.33), indicating a modest tendency of the results to be in favor of the null 172 hypothesis. Additionally, to test temporal filtering within each modality, we calculated dependent-173 samples t-tests comparing T1 and T3 within each trial type. For auditory stimulation, we found 174 significantly lower discomfort ratings in T 1 compared to T3 despite constant stimulation (CT; t 175 (61) = −5.69, p < 0.001, d z = 0.40, M DIFF = −14.32), indicating temporal loudness summation. 176 However, no difference between T1 and T3 was found for offset trials (OT; t (61) = 0.93, p = 0.36, 177 dz = 0.08, MDIFF = 2.46). To the contrary, for thermal stimulation, pain ratings decreased from T1 178 to T3 despite constant stimulation (CT, t ( 61) = 5.05, p < 0.001, d z = 0. 45, MDIFF = 14.53), 179 indicating temporal adaptation. Furthermore, a pronounced decrease in pain ratings from T1 to T3 180 was observed for offset trials (OT, t (61) = 12.17, p < 0.001, dz = 1.74, MDIFF = 62.57), resulting 181 in a strong TCE effect. A visual representation of changes over time using boxplots can be seen in 182 Fehler! Verweisquelle konnte nicht gefunden werden.. 183 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 10 Pupil size reflects perception of pain but not auditory discomfort 184 We next examined the event-related pupil dilation (Figure 3) to determine whether TCE is 185 accompanied by modality-dependent autonomic modulation, contrasting auditory versus thermal 186 stimulation. Similarly to the behavioral data, the main effect modality and all interaction effects 187 including modality reached significance (p < 0.001; S6 Table). For auditory stimulation, we found 188 a significant main effect of ‘time’ (F (2, 56) = 32.67, p < 0.001,  p2 = 0.54). FDR corrected post 189 hoc testing revealed significant reduction in the pupil size for T3 vs. T2 (p < 0.001), T3 vs. T1 (p 190 < 0.001) but not for T2 vs. T1 (p = 0.74). This indicates a gradual decrease in pupil size during the 191 trial. 192 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 11 193 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 12 Figure 3 Mean pupil response over time. Mean pupil diameter (mm) over time was measured for auditory (A) and thermal (B) 194 stimulation. Thermal stimulation induced significant changes (* , p < .05; cluster -based permutation test ) in the pupil diameter 195 when comparing offset trials (OT, darker colors) with constant trials (CT, lighter colors). No significant clusters could be identified 196 for auditory stimulation, despite visually observable changes to stimulus onset and offset. The timing of the stimulation intensity is 197 displayed as a solid (OT) and dashed (CT) line at the top of the figure and was similar for both modalities. Mean pupil diameter 198 (mm) for each condition (C) split into three time intervals (T1: 0-5s, T2: 15-20s, T3: 20-25s). Significant (p < .05) between-modality 199 comparisons indicated by dashed lines. 200 For thermal stimulation, there was a significant interaction effect of ‘trial type’ and ‘time’ 201 (F (2, 56) = 64.35, p < 0.001,  p2 = 0.70). Comparing trial types at different time intervals , we 202 found significantly enhanced pupil size for OT vs. CT in T2 (p < 0.001) but not in T1 (p = 0.88) 203 and T3 (p = 0.08). In addition to these planned statistical analyses, we performed exploratory 204 cluster based permutation test (51), comparing OT and CT during the entire trial time course (see 205 Figure 3). 206 Neural alpha power reflects pain sensation and expectation 207 Finally, we tested whether commonalities and differences in behavioral TCE effects for 208 thermal versus auditory stimuli were reflected in neural oscillatory EEG responses (Figure 4). For 209 data analysis , we used baseline -corrected time frequency representations of oscillatory power , 210 averaged across selected channels and frequencies in the alpha range ( 7–13 Hz) for the defined 211 time bins of interest (for details, see Materials and methods). A visual representation of the event-212 related potential can be accessed in the Supporting Information (S8 Fig. and S9 Fig.). 213 We first calculated a combined repeated -measures ANOVA including both modalities 214 (painful heat & aversive sound), which can be accessed in the Supporting Information (S7 Table). 215 Unlike the pupillometry and behavioral results, interactions involving modality did not reach 216 significance (all p > 0.1). For interpretability, we nonetheless conducted two modality -specific 2 217 (trial type) x 3 (time interval) ANOVAs. For auditory stimulation , the analysis revealed a 218 significant main effect of ‘time’ (F (2, 56) = 9.92, p < 0.001,  p2 = 0.26), while thermal stimulation 219 displayed a significant interaction effect of ‘trial’ and ‘time’ (F (2, 56) = 5.91, p = 0.005,  p2 = 220 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 13 0.17). Comparing the different time intervals in auditory simulation with FDR corrected post-hoc 221 tests showed a significant difference in alpha power between T1 and T2 (p = 0.033), T2 and T3 (p 222 = 0.004), and T3 and T1 (p = 0.004), indicating an overall alpha power increase over time. The 223 interaction effect in thermal stimulation was driven by a significant difference between OT and 224 CT only for T2 (p = 0.002) but not for T1 (p > 0.05) and T3 (p > 0.05). 225 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 14 226 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 15 Figure 4: Mean alpha power over time for both modalities. Auditory (A) and thermal (B) stimulation-induced mean alpha power 227 (7–13 Hz) averaged across centro-occipital electrodes (P7, P3, O1, C3, POz, Pz, CPz, P8, P4, O2, C4) . Topographic maps show 228 spatial distributions of alpha power at time points 0, 10, 20 and 30 seconds (color limits: yellow = +2dB, blue = –2dB). Comparing 229 alpha oscillat ory power during Offset (OT) (darker colors) and Constant (CT) (lighter colors) trials using cluster -based 230 permutation tests resulted in significant (* , p < .05 ) differences only during thermal stimulation . Mean (alpha) power for each 231 condition (C) split into the three time intervals of interest (T1: 0-5s, T2: 10-15s, T3: 20-25s). Significant (p < .05) within-modality 232 comparisons indicated by solid lines, between-modality comparisons indicated by dashed lines. 233 Notably, significant alpha modulation in the T2 interval for thermal stimulation was 234 attributable to two distinct underlying mechanisms reflecting the occurrence and expectation of 235 the stimulus (Figure 5). To differentiate these, we performed post-hoc testing to compare a baseline 236 interval during T1 (5–10 s) to the time interval wherein a cluster-based permutation test revealed 237 a significant difference for OT vs. CT (10.6 - 17.8 s; see Figure 4). First, the increase in thermal 238 stimulation intensity induced an alpha power decrease in OT trials, resulting in significantly lower 239 alpha power compared to T1 (t (28) = −2.61, p = 0.015, dz = 0.48, MDIFF = −0.40). Second, an 240 omission of the expected increase in thermal stimulation intensity induced an alpha power increase 241 in CT trials compared to the baseline interval during T1 (t (28) = 3.34, p = 0.002, dz = 0.62, MDIFF 242 = 0.49). These findings demonstrate that temporal filtering of pain sensation (but not auditory 243 discomfort) reflects both stimulus -driven (bottom -up) and expectation -driven (top -down) 244 mechanisms. 245 246 Figure 5: Topographic distribution of alpha power. Stimulation-induced mean alpha power (7–13 Hz) averaged across centro-247 occipital electrodes (P7, P3, O1, C3, POz, Pz, CPz, P8, P4, O2, C4). Topographic maps show spatial distributions of alpha power 248 at the significant time point (11 sec – 17 sec, one topography per second) of a cluster-based permutation test (* p < .05) comparing 249 offset (OT) and constant (CT) trials (color limits: yellow = +2dB, blue = –2dB). 250 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 16

251 Temporal filtering operates across sensory modalities, including vision, audition, and pain 252 perception (20,21). Here, we aimed to delineate behavioral and (neuro -)physiological substrates 253 of temporal filtering and to assess its specificity to pain and unpleasantness . To this end, we 254 employed the phenomenon of temporal contrast enhancement , a temporal filtering mechanism 255 thought to underlie offset analgesia (26). TCE was elicited by pain-inducing heat and aversive 256 sounds, indicating a supramodal temporal filtering process with distinct temporal dynamics across 257 modalities. Pupil dilation, indexing autonomic nervous system activation, and alpha -band 258 oscillations, reflecting cortical inhibition, were both modulated by temporal increases in noxious 259 heat but not by auditory stimulation. These findings suggest pain-specific temporal filtering acting 260 on both the processing of painful input and its temporal expectation. 261 Modality-specific processes underlie temporal contrast enhancement 262 Consistent with prior evidence, we were able to successfully induce TCE using thermal 263 stimulation. Noxious heat is known to produce robust TCE effects, indicated by statistically 264 significant decreases in pain sensation following a short heat offset (52). To our knowledge, this 265 is the first experiment investigating TCE using non-painful aversive auditory stimuli, 266 demonstrating that TCE reflects a modality -general (i.e., supramodal) mechanism of afferent 267 contrast filtering rather than a nociception-specific process. In addition to domain-general effects, 268 we observed modality -specific patterns in subjective ratings , suggesting intricate processing 269 differences between modalities during the stimulation paradigm. When comparing different time 270 intervals (T1 vs. T3) during constant stimulation, we found pain ratings decreasing over time while 271 the opposite was observed for subjective auditory discomfort ratings . We argue that these results 272 represent two different temporal filtering mechanisms. First, adaptation to pain , which is well-273 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 17 known to be induced by tonic heat (16,53). Second, temporal sound summation, reflecting neural 274 integration of auditory energy over time (12,13). 275 Comparing ratings between time windows of interest for offset trials, we found a typical 276 decrease in perceived pain following heat offset (T3) compared to baseline (T1) . In the auditory 277 modality, however, the same contrast was not significant, suggesting a differently constituted TCE 278 effect. Here, TCE was driven by a gradual increase in TSS during constant stimulation, which was 279 absent during offset trials, thereby yielding a comparable overall TCE magnitude. Taken together, 280 these results indicate temporal pain inhibition but an absence of such an effect for auditory 281 stimulation. Importantly, the dissociation of underlying generators of TCE effects for pain and 282 auditory stimulation receives further support from the absence of a correlation between TCE 283 effects, suggesting partly independent generators of what appears like a common filtering 284 mechanism. 285 Neurophysiological signatures of filtering perception of pain and 286 auditory discomfort 287 Alongside subjective behavioral responses, we collected objective neuro(physiological) 288 measures (pupillometry and EEG ) aiming to better understand the differences between the 289 observed TCE effects for auditory vs. thermal stimulation. Thermal stimulation induced significant 290 differences in pupil size between OT and CT in T2, the interval with the highest stimulus intensity. 291 Reactive increases in pupil diameter to pain -inducing stimuli may reflect autonomic nervous 292 system (ANS) activation, specifically sympathetic arousal associated with heightened awareness 293 and the fight -or-flight response, as observed in previous studies on pupillary responses to pain 294 (43,54–57). Our observed phasic increases in pupil diameter likely reflect bottom-up processes 295 induced by increases in stimulus temperature and are in line with previous research (58). However, 296 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 18 auditory stimulation did not display such trial type-specific differences, only eliciting significant 297 decreases of pupil diameter over time. This suggests that auditory discomfort did not elicit sizeable 298 modulation of activity in the ANS. Possible explanations for this difference could be factors such 299 as stimulus intensity or salience, with painful stimulation arguably resulting in a higher perceived 300 threat level compared to auditory stimulation , hence inducing far greater changes in pupil 301 diameter. 302 Consistent with the pupillary dynamics, time –frequency EEG analyses revealed a similar 303 modality-specific pattern: thermal stimulation elicited a significant alpha power reduction during 304 offset trials in T2, which was not observed for auditory stimulation. A reduction of alpha power in 305 response to painful stimulation is in line with previous reports of alpha decreases accompanying 306 increased pain perception (31,32,59). Moreover, when comparing T1 and T2 in OTs, we observed 307 a significant alpha desynchronization, reflecting bottom -up processes associated with increased 308 stimulus intensity. In contrast, CTs exhibited significant alpha synchronization (i.e., an alpha 309 power increase) that may reflect top -down processes, possibly induced by expectation of an 310 increase in painful stimulation. This agrees with prior research observing top -down processes 311 influencing alpha oscillations (60). For auditory stimulation, we hypothesized similar spatial and 312 temporal patterns of oscillatory activity ; however, no significant changes in alpha power were 313 observed for either stimulation pattern. Given the comparable stimulation paradigm employed in 314 both modalities, one could expect that participants would manifest a comparable increase in alpha 315 power, driven by the anticipation of a change in stimulus intensity. However, no such effects were 316 observed, which lends further support to the view that the neural modulation and representation of 317 auditory discomfort differ from that of painful stimulation. 318 Interestingly, we did not observe any neurophysiological changes mirroring the TCE effect 319 that we observed in behavior. Neither pupillometry, nor EEG displayed any significant 320 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 19 modulations by trial type (OT vs. CT) during T3. Based on literature showing a robust link between 321 tonic pain and reduced alpha oscillatory power, we hypothesized that the TCE effect would elicit 322 measurable changes in alpha activity —potentially an alpha increase reflecting the experienced 323 relief - as it has been reported for placebo analgesia (61,62). Contrary, we did not observe any 324 representation of the behavioral TCE effects in neural recordings. One plausible explanation could 325 be that the brain regions responsible for processing or mediating TCE are mainly located 326 subcortically, which would be less accessible to surface EEG. Previous fMRI studies have reported 327 changes to activity in the insula (24,25), structures of the brainstem (25,28), putamen and nucleus 328 accumbens (23) or the spinal cord (29). It might thus be that these structures are more relevant to 329 the construction of TCE compared to the more superficially located ones. Furthermore, one should 330 be cautious expecting identical findings when comparing fMRI to EEG, since both imaging 331 techniques show distinct sensitivities to neural and hemodynamic processes, temporal and spatial 332 resolution, and often can display different correlations depending on measured frequencies 333 (63,64). Additionally, subjective experiences such as changes in pain perception may not always 334 align with objective indices, which could be the case for TCE. Our pupillometry findings provide 335 partial support for this hypothesis, since both modalities displayed temporal filtering as evidenced 336 by decreases in pupil diameter over time. However, these temporal patterns did not correspond 337 with the observed patterns in the subjective domain, where only pain decreased over time, but 338 auditory discomfort increased. However, to our knowledge, studies investigating TCE employing 339

measures besides fMRI are scarce (65–67) making it difficult to assess the extent to 340 which subjective and physiological responses to TCE diverge and whether our findings reflect 341 such an incongruence. 342 We replicated robust behavioral TCE responses in both investigations using individually 343 calibrated and fixed -stimulus approaches. For the neurophysiological investigation, we adopted 344 the fixed-stimulus approach, as our first study and previous work indicate that it is better suited 345 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 20 for assessing neurophysiological outcomes (68,69). However, this approach may have led to a 346 lower perceived stimulus intensity, which could have potentially reduced our observed effects, 347 especially in pupillary reactions to auditory stimuli. Regardless, we could not have increased the 348 stimulation intensity further since this could have meant potential harm to the participants (70). 349 Nonetheless, we would attribute the smaller changes in pupil diameter to auditory stimulation, to 350 the underlying aversiveness and threat level that painful stimuli impose on the organism. In 351 addition, modality-specific patterns of temporal dynamics were observed, particularly represented 352 in the behavior of participants. An identical stimulation paradigm was employed for all trials, 353 enabling direct comparison between modalities and providing a common framework for 354 interpreting differences in temporal filtering dynamics. As demonstrated in previous research, 355 TCE has been consistently identified as a highly robust filtering mechanism in pain. TCE effects 356 can be induced with a variety of stimulus intervals, temperature changes and even repeated instead 357 of tonic stimuli (52,71,72). However, future research could benefit from the use of variable 358 stimulus sequences to further improve our understanding of these supramodal temporal filtering 359 dynamics, driving TCE across modalities. Additionally, future research should aim to characterize 360 the temporal dynamics of TCE in different sensory modalities without compromising spatial 361 resolution, ideally incorporating approaches that target subcortical regions potentially underlying 362 the effect. Further exploration of pain -specific paradigms, and of potential shared filtering 363 mechanisms across sensory modalities, may provide deeper insight into how such processes shape 364 perceptual experience in our multisensory environment. 365

366 In this study, we probed the specificity of temporal contrast enhancement to pain. We 367 successfully induced TCE using pain-inducing heat but also auditory stimulation, suggesting the 368 existence of a supramodal temporal filtering mechanism . However, divergences in temporal 369 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 21 filtering dynamics indicate that adaptation to pain , but temporal summation of sound underlie 370 contrast enhancement. Modality-specific modulations in pupil size and neural oscillations indicate 371 that painful heat – but not auditory – stimulation increases activity of the ANS and neural 372 representation of an occurring, as well as an expected but omitted increase in stimulation intensity. 373 These findings demonstrate the intricate dynamics of temporal filtering mechanisms and their 374 (neuro)physiological basis. 375

376 Behavioral investigation of the TCE effect in auditory stimulation 377 Participants 378 Participants (n = 33, 22f; age = 23.7, SD = 6.0) were included if they subjectively reported 379 being healthy and pain-free on the day of the procedure. Exclusion criteria were chronic pain (> 3 380 months) within the last two years, diagnosed systemic, neurological, cardiovascular or psychiatric 381 diseases and being with diagnosed hearing loss or tinnitus. All participants were asked not to take 382 any painkillers, consume alcohol or undertake any strenuous physical activity 24 hours before 383 participating in the study. To characterize the study sample age, sex at birth, body mass (kg), 384 stature (cm), handedness and general fear of heat pain and fear of loud noises on a numerical rating 385 scale from 0 (no fear) to 100 (highest possible fear) were recorded for each participant. To measure 386 noise sensitivity the Weinstein Noise Sensitivity Scale (WNSS) was used (73). In addition, the 387 Pain Vigilance and Awareness Questionnaire (PVAQ) was used to measure attention to pain and 388 assesses awareness, consciousness, vigilance, and observation of pain (74) and the Pain 389 Catastrophizing Scale (PCS) was used to assess the catastrophizing behavior (75). 390 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 22 The study was approved by the Ethics Committee of the University of Lübeck (2023-633) 391 and conducted in accordance with the Declaration of Helsinki. The methodology was preregistered 392 on the Open Science Framework (OSF; https://osf.io/643kg). 393 Equipment 394 All heat stimuli were applied with a thermal contact stimulator (TCS; André Dufour, 395 University of Strasbourg, France). The probe of the TCS has a total stimulation zone of 9 cm2 (five 396 equal stimulation zones, each 0.74 x 2.4cm = 8.88cm2) and was applied to the left forearm. The 397 probe weighs 440g and the TCS has a temperature range of 0°C to 60°C, adjustable at 0.1°C 398 intervals. The maximum temperature rise and fall rate is 100°C/second. For auditory stimulation, 399 a 1000-Hz sine wave tone sampled at 44.100Hz, generated using Adobe Audition (Adobe Systems 400 Software, Dublin, Republic of Ireland ), was applied. The sound was presented using over -ear 401 headphones (PXC 550 -II, Sennheiser, Wedemark, Germany) . Probing pain intensity was 402 conducted with a Python -based eVAS (76,77). The eVAS was displayed using a computer and 403 ranged from 0 "no sensation" to 200 "worst heat pain imaginable", while a value of 100 represented 404 the “pain threshold” for thermal stimulation . For auditory stimulation, the same eVAS was used 405 but it displayed different anchors, 0 representing “no sound audible”, 200 being “maximum 406 discomfort imaginable” and 100 representing the “discomfort threshold”. 407 Familiarization and calibration procedure 408 Before attending the main part of the experiment, participants were familiarized to the 409 rating procedure and stimulus intensities. For this, they received three different stimulus 410 intensities, a high (48°C, 100dB) , low (33°C, 49dB) and an intermediate (40°C, 82dB) level of 411 intensity for thermal and auditory stimulation, respectively. Afterwards, the stimulation intensities 412 for each modality were calibrated, using a staircase procedure (S10 Fig. in the Supporting 413 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 23 Information). Participants were asked to continuously rate their sensory experience using the 414 eVAS scale. The calibration consisted of ascending stimulation intensities, where each stimulation 415 was applied for ten seconds , followed by ten seconds of either no stimulation (0dB) or baseline 416 temperature (32°C) depending on the applied modality. The stimulation started at either 33°C or 417 49dB and increased in 1°C or 3dB steps up until the highest stimulation intensity of 48°C or 100dB 418 was reached. This procedure was repeated for each modality with the modalities being alternated. 419 The stimulation intensities were then derived from the second iteration of each modality by using 420 stimulation intensities that produced perceived pain intensities of 150/200 and 175/200 on the 421 eVAS. These values were used for the initial stimulation intensity (T1) and increase in stimulation 422 intensity (T2) in offset trials (see Experimental paradigm). Mean heat pain that corresponded to 423 150 points on the 0- to 200-point eVAS was 46.4°C (SD 0.8°C) and 47.4°C (SD 0.8°C) for 175/200 424 points respectively. Mean sound stimulation parameters for 150/200 and 175/200 eVAS ratings 425 were 90.8dB (SD 8.0dB) and 95.0dB (SD 7.5dB), respectively. 426 Experimental paradigm 427 A TCE paradigm with three successive periods (T1-T2-T3) consisting of an OT and a CT 428 was performed (78). CTs were administered for a duration of 35 seconds, during which continuous 429 heat or sound stimulation was applied at a calibrated intensity corresponding to 150/200 on the 430 eVAS. The OTs consisted of an intensity corresponding to 150/200 (T1), followed by ten seconds 431 of 175/200 (T2) and then decreased back to the stimulation intensity of T1 for 15 seconds (T3). 432 Rise and fall rates were kept constant (100°C/s). Each trial was performed four times (4x OT, 4x 433 CT) with a break of 2 min in between the trials . The order of trials was pseudorandomized in a 434 counterbalanced manner. 435 Participants continuously rated the experienced pain intensity throughout each trial and 436 were instructed to attend carefully and indicate even very subtle sensations or changes. This 437 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 24 procedure was repeated once for each modality, resulting in a total of 16 trials per participant. All 438 trials were completed for one modality before the other modality was tested . This transition was 439 preceded by a 5-min interval of no stimulation to reduce possible carry-over effects. 440 Statistical analysis 441 Sample size calculation was based on the previously described OA effect size (78). With 442 an estimated effect size of dz = 0.76 (27), an estimated group size of n = 25 participants for paired 443 comparisons (two-sided t-test, α = 5%) was required to reach statistical power of 95% (G*Power, 444 University of Düsseldorf) (79). To preserve planned power, improve precision, and mitigate sex 445 imbalance, we prospectively oversampled to n=33 (80). 446 Statistical analyses were performed using R Studio (RStudio version 2024.04.11 with R 447 version 4.5.0, R Foundation for Statistical Computing, Vienna, Austria) (81) and MATLAB (The 448 Mathworks Inc., 2024) (82) Parametric data is presented in means ( x̄ ) with standard deviations 449 (SD) and nonparametric data in median (M) with ranges (R) or absolute and relative frequencies. 450 The average eVAS ratings were obtained for each time interval (T1, T2 and T3). The first 5s of 451 T1, T2 and T3 and the last 5s of T3 were not considered in our analysis. This is an approach we 452 have chosen before to consider the delay in pain response and extract stable pain ratings (83,84). 453 We calculated a 2x3 repeated measures ANOVA with the factors ‘trial’ (CT, OT) and ‘time’ (T1, 454 T2, T3). If there were significant findings, FDR corrected (85,86) t-tests were performed. The level 455 of significance was set at p < 0.05. We deviated from our initial analysis plan (dependent t-tests) 456 in the preregistration to achieve more coherence in context of the analysis of the second study. An 457 additional ANOVA analysis including both modalities in a 2 x 2 x 3 ANOVA can be accessed in 458 the Supporting Information (S4 Table). 459 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 25 Neurophysiological expression of TCE in auditory and thermal 460 stimulation 461 In this experiment, we aimed to investigate neural and autonomic nervous system responses 462 using a similar stimulation paradigm and the same modalities as explained above . For this, we 463 collected EEG data, pupillometry data and behavioral ratings. The study was approved by the 464 Ethics Committee of the University of Lübeck (2023 -633) and conducted in accordance with the 465 Declaration of Helsinki. Again, the methodology was preregistered on the Open Science 466 Framework (OSF; https://osf.io/v37mp). 467 Participants 468 A total of 29 healthy participants (sex = 19f; age = 24.6, SD = 5.7) were used for analysis. 469 In- and Exclusion criteria were similar for both experiments. None of the volunteers that 470 participated in the behaviorally focused experiment were recruited for the second experiment. 471 Equipment 472 The same stimulation equipment as in the previous experiment was used for all procedures. 473 Behavioral data were collected using an eVAS similar in style and anchors, but the visual 474 presentation, data collection and stimulus control was achieved using MATLAB (The Mathworks 475 Inc., 2024). For visual presentation purposes, the psychophysics toolbox (87) was used. The EEG 476 was recorded at 24 passive scalp electrodes (SMARTING, mBrainTrain, Belgrade, Serbia) at a 477 sampling rate of 500 Hz (DC to 250 Hz bandwidth), referenced against electrode FCz. Electrode 478 impedances were kept below 10 kΩ. The amplifier was attached to the EEG cap (Easycap, 479 Herrsching, Germany) and the EEG data were transmitted via Bluetooth to a nearby computer, 480 which recorded the data using the Smarting Streamer (Version 3.4.2). For pupillometry, a Tobii 481 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 26 X3-120 Eye Tracker (Tobii Technology Inc., Stockholm, Sweden) was used with sampling rate of 482 120Hz. Pupillometry data were directly recorded via MATLAB. 483 Familiarization and calibration procedure 484 Contrary to the first experiment, we did not individually calibrate the stimuli. This 485 approach was chosen due to the robust effects seen in the first experiment . Furthermore, the 486 application of constant stimulation intensities ensures uniform stimulus input across all 487 participants, a feature that assumes greater significance in the context of neurophysiological 488 investigations (68). We derived stimulus intensities from mean values of the stimulation 489 parameters of the behavioral experiment. The initially derived stimulation parameters from the 490 first experiment that were used for thermal stimulation led to major pain habituation effects 491 resulting in close to zero pain felt by the participants after a few trials. Due to this, we increased 492 the temperature by 1.5°C (initially 4 6°C and 4 7°C) to reduce these habituation effects . The 493 participants (n = 11) that received the stimulation protocol prior to this change were excluded from 494 the analysis. In order to account for the missing practice using the eVAS, the participants 495 underwent a brief familiarization procedure. This procedure comprised five stimuli in ascending 496 order, either starting from 44.5°C for thermal stimulation and increasing in 1°C steps or starting 497 from 80dB and increasing in 5dB steps. Each stimulus was presented for 10 seconds and then 498 followed by a 10-second pause (matching the calibration procedure from the first experiment). 499 Experimental paradigm 500 The TCE paradigm was the same as the one used for the behavioral investigation but 501 consisted of fixed stimulation intensities of 47.5°C or 95dB for T1, 48.5°C or 100dB for T2 and 502 47.5°C or 95dB for T3, respectively. Each modality was tested twice, and each stimulation block 503 consisted of ten trials . The first two trials per block were designated as an OT or CT, with the 504 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 27

of collecting behavioral data. Participants were tasked with continuously rating their 505 perceived stimulus intensity. For the remaining eight trials, participants were instructed to maintain 506 their gaze on a fixation cross, focusing on their sensory experience without moving. The order of 507 trials was pseudorandomized to a matching number of trials per block. Each “modality block” was 508 followed by a 5-min break and then alternated to the other modality. For an overview of the 509 procedure, consult Figure 1. 510 Statistical analysis 511 The determination of sample size was based on the effect sizes derived from the first 512 experiment. For the auditory modality, an effect size of d z = 0.64 (thermal d z = 1.09) , with 513 statistical thresholds set at alpha = 0.05 and beta = 0.1 (90% power), indicated an estimated sample 514 size of 28 participants . Due to the mentioned changes in temperature (see Familiarization and 515 calibration procedure) and the exclusion of 11 participants that were previously recorded the final 516 sample size for analysis was n = 29. 517 Statistical analyses were performed using R Studio (RStudio version 2024.04. 11 with R 518 version 4.5.0, R Foundation for Statistical Computing, Vienna, Austria) (81) and MATLAB (82). 519 Parametric data is presented in means (x̄ ) with standard deviations (SD) and nonparametric data 520 in median (M) with ranges (R) or absolute and relative frequencies. We calculated a 2x3 repeated-521 measures ANOVA with the factors ‘trial’ (CT , OT) and ‘time’ (T1, T2, T3). If there were 522 significant effects, FDR corrected (85,86) post-hoc t-tests were performed. The level of 523 significance was set at p < 0.05. An additional ANOVA analysis including both modalities in a 2 524 x 2 x 3 ANOVA can be accessed in the Supporting Information ( S5 Table). For behavioral data 525 analysis, the extracted time intervals matched the ones chosen in the first study. For the analysis 526 of EEG and pupillometry data, we shifted the extracted time intervals to right after onset of the 527 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 28 stimulus, since we did not expect any delay in response compared to the behavioral data, resulting 528 in T1 being 0-5s, T2 being 10-15s and T3 being 20-25s. 529 EEG preprocessing and analysis 530 The continuous EEG data were high -pass ( 1 Hz) and low -pass filtered ( 100 Hz), re -531 referenced to the average reference across all electrodes, and epoched from –5 to +40s relative to 532 the onset of auditory/thermal stimulation. An independent component analysis (ICA) was used to 533 remove components related to eye -blinks, eye-movements and muscle activity . Remaining 534 artefactual epochs were removed afterwards by visual inspection. All data analyses were carried 535 out in Matlab (R2024b), using custom scripts and the Fieldtrip toolbox (88). Time-frequency 536 oscillatory power representations of single -trial EEG data were obtained using Fast Fourier 537 Transform (FFT) with multi-tapering (DPSS, discrete prolate spheroidal sequences) for a moving 538 time window (length: 2s; moving in steps of 0.1s through the trial) for frequencies 1–80Hz in steps 539 of 1 Hz with 2Hz spectral smoothing. 540 Pupillometry preprocessing 541 Similarly to EEG analysis, we used Fieldtrip (88) to conduct the necessary preprocessing 542 steps for the pupillometry data. First, samples reflecting physiologically implausible pupil changes 543 were identified using a velocity -based criterion: data points whose dilation speed exceeded three 544 standard deviations above the mean velocity within each trial were marked as invalid. Next, short 545 gaps of missing data (< 500 ms) were interpolated using a cubic spline method to reconstruct brief 546 blink-related signal loss, following the recommendations of Sebastiaan Mathôt and Kret and Sjak-547 Shie (89,90). Longer gaps were left unaltered to avoid introducing artificial signal. Finally, pupil 548 size data from the left and right eye were averaged to obtain a single mean for subsequent analyses. 549 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 29

550 We thank Anna M. Hagemann for her contribution and assistance during data collection. 551 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 30

552 1. Zhu Y, Nachtrab G, Keyes PC, Allen WE, Luo L, Chen X. Dynamic salience processing in 553 paraventricular thalamus gates associative learning. Science. 2018 Oct 26;362(6413):423–9. 554 2. Choi I, Lee JY, Lee SH. Bottom-up and top-down modulation of multisensory integration. 555 Curr Opin Neurobiol. 2018 Oct;52:115–22. 556 3. Noppeney U. Perceptual Inference, Learning, and Attention in a Multisensory World. Annu 557 Rev Neurosci. 2021 July 8;44:449–73. 558 4. Fredericksen RE, Hess RF. Estimating multiple temporal mechanisms in human vision. 559 Vision Research. 1998 Apr 1;38(7):1023–40. 560 5. Snowden RJ, Hess RF. Temporal frequency filters in the human peripheral visual field. 561 Vision Research. 1992 Jan 1;32(1):61–72. 562 6. Mehta AH, Feng L, Oxenham AJ. Neural auditory contrast enhancement in humans. Proc 563 Natl Acad Sci U S A. 2021 July 20;118(29):e2024794118. 564 7. Allen D. Spatial summation in the evoked cortical response to auditory stimuli. Perception & 565 Psychophysics. 1968 Nov 1;4(6):355–6. 566 8. Sollini J, Mill R, Sumner CJ. Spatial Processing Is Frequency Specific in Auditory Cortex 567 But Not in the Midbrain. J Neurosci. 2017 July 5;37(27):6588–99. 568 9. Ordás CM, Alonso-Frech F. The neural basis of somatosensory temporal discrimination 569 threshold as a paradigm for time processing in the sub-second range: An updated review. 570 Neurosci Biobehav Rev. 2024 Jan;156:105486. 571 10. Lieber JD, Bensmaia SJ. The neural basis of tactile texture perception. Curr Opin Neurobiol. 572 2022 Oct;76:102621. 573 11. Willmore BDB, King AJ. Adaptation in auditory processing. Physiol Rev. 2023 Apr 574 1;103(2):1025–58. 575 12. Pedersen CB, Salomon G. Temporal integration of acoustic energy. Acta Otolaryngol. 576 1977;83(5–6):417–23. 577 13. Buus S, Florentine M, Poulsen T. Temporal integration of loudness, loudness discrimination, 578 and the form of the loudness function. J Acoust Soc Am. 1997 Feb 1;101(2):669–80. 579 14. Mauk MD, Buonomano DV. The neural basis of temporal processing. Annu Rev Neurosci. 580 2004;27:307–40. 581 15. Chen KK, Rolan P, Hutchinson MR, de Zoete RMJ. Reliability of Temporal Summation of 582 Pain in Healthy and Clinical Populations: A Systematic Review and Meta-Analysis. 583 European Journal of Pain. 2025;29(8):e70097. 584 16. Hashmi JA, Davis KD. Effects of temperature on heat pain adaptation and habituation in 585 men and women. Pain. 2010 Dec;151(3):737. 586 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 31 17. Koyama Y, Koyama T, Kroncke AP, Coghill RC. Effects of stimulus duration on heat 587 induced pain: the relationship between real-time and post-stimulus pain ratings. Pain. 2004 588 Feb;107(3):256. 589 18. Weissman-Fogel I, Dror A, Defrin R. Temporal and spatial aspects of experimental tonic 590 pain: Understanding pain adaptation and intensification. European Journal of Pain. 591 2015;19(3):408–18. 592 19. Sathian K, Lacey S. Cross-Modal Interactions of the Tactile System. Current Directions in 593 Psychological Science [Internet]. 2022 July 25 [cited 2025 Sept 30]; Available from: 594 https://journals.sagepub.com/doi/10.1177/09637214221101877 595 20. Grothe B, Klump GM. Temporal processing in sensory systems. Current Opinion in 596 Neurobiology. 2000 Aug 1;10(4):467–73. 597 21. Di Stefano N, Spence C. Perceiving temporal structure within and between the senses: A 598 multisensory/crossmodal perspective. Atten Percept Psychophys. 2025 July 1;87(5):1811–599 38. 600 22. Cariani P, Baker JM. Survey of temporal coding of sensory information. Front Comput 601 Neurosci. 2025;19:1571109. 602 23. Zhang S, Li T, Kobinata H, Ikeda E, Ota T, Kurata J. Attenuation of offset analgesia is 603 associated with suppression of descending pain modulatory and reward systems in patients 604 with chronic pain. Mol Pain. 2018;14:1744806918767512. 605 24. Nahman-Averbuch H, Martucci KT, Granovsky Y, Weissman-Fogel I, Yarnitsky D, Coghill 606 RC. Distinct brain mechanisms support spatial vs temporal filtering of nociceptive 607 information. Pain. 2014 Dec;155(12):2491–501. 608 25. Yelle MD, Oshiro Y, Kraft RA, Coghill RC. Temporal filtering of nociceptive information 609 by dynamic activation of endogenous pain modulatory systems. J Neurosci. 2009 Aug 610 19;29(33):10264–71. 611 26. Yelle MD, Rogers JM, Coghill RC. Offset analgesia: A temporal contrast mechanism for 612 nociceptive information. Pain. 2008 Jan;134(1):174–86. 613 27. Grill JD, Coghill RC. Transient analgesia evoked by noxious stimulus offset. Journal of 614 Neurophysiology. 2002;87(4):2205–8. 615 28. Derbyshire SWG, Osborn J. Offset analgesia is mediated by activation in the region of the 616 periaqueductal grey and rostral ventromedial medulla. Neuroimage. 2009 Sept;47(3):1002–6. 617 29. Sprenger C, Stenmans P, Tinnermann A, Büchel C. Evidence for a spinal involvement in 618 temporal pain contrast enhancement. NeuroImage. 2018 Dec 1;183:788–99. 619 30. Brooks J, Tracey I. REVIEW: From nociception to pain perception: imaging the spinal and 620 supraspinal pathways. Journal of Anatomy. 2005;207(1):19–33. 621 31. Peng W, Hu L, Zhang Z, Hu Y. Changes of Spontaneous Oscillatory Activity to Tonic Heat 622 Pain. PLOS ONE. 2014 Mar 6;9(3):e91052. 623 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 32 32. Nickel MM, Ta Dinh S, May ES, Tiemann L, Hohn VD, Gross J, et al. Neural oscillations 624 and connectivity characterizing the state of tonic experimental pain in humans. Hum Brain 625 Mapp. 2020 Jan;41(1):17–29. 626 33. Nir RR, Sinai A, Moont R, Harari E, Yarnitsky D. Tonic pain and continuous EEG: 627 prediction of subjective pain perception by alpha-1 power during stimulation and at rest. Clin 628 Neurophysiol. 2012 Mar;123(3):605–12. 629 34. Jensen O, Bonnefond M, VanRullen R. An oscillatory mechanism for prioritizing salient 630 unattended stimuli. Trends in Cognitive Sciences. 2012 Apr 1;16(4):200–6. 631 35. Jensen O, Mazaheri A. Shaping Functional Architecture by Oscillatory Alpha Activity: 632 Gating by Inhibition. Front Hum Neurosci [Internet]. 2010 Nov 4 [cited 2025 Sept 30];4. 633 Available from: https://www.frontiersin.org/journals/human-634 neuroscience/articles/10.3389/fnhum.2010.00186/full 635 36. Clements GM, Gyurkovics M, Low KA, Kramer AF, Beck DM, Fabiani M, et al. Dynamics 636 of alpha suppression index both modality specific and general attention processes. 637 NeuroImage. 2023 Apr 15;270:119956. 638 37. Strauß A, Wöstmann M, Obleser J. Cortical alpha oscillations as a tool for auditory selective 639 inhibition. Front Hum Neurosci [Internet]. 2014 May 28 [cited 2025 Apr 8];8. Available 640 from: https://www.frontiersin.org/journals/human-641 neuroscience/articles/10.3389/fnhum.2014.00350/full 642 38. Klimesch W, Sauseng P, Hanslmayr S. EEG alpha oscillations: The inhibition–timing 643 hypothesis. Brain Research Reviews. 2007 Jan 1;53(1):63–88. 644 39. Wöstmann M, Herrmann B, Maess B, Obleser J. Spatiotemporal dynamics of auditory 645 attention synchronize with speech. Proceedings of the National Academy of Sciences. 2016 646 Apr 5;113(14):3873–8. 647 40. Wöstmann M, Alavash M, Obleser J. Alpha Oscillations in the Human Brain Implement 648 Distractor Suppression Independent of Target Selection. J Neurosci. 2019 Dec 649 4;39(49):9797–805. 650 41. Schneider D, Herbst SK, Klatt LI, Wöstmann M. Target enhancement or distractor 651 suppression? Functionally distinct alpha oscillations form the basis of attention. European 652 Journal of Neuroscience. 2022;55(11–12):3256–65. 653 42. Kyle BN, McNeil DW. Autonomic Arousal And Experimentally Induced Pain: A Critical 654 Review of the Literature. Pain Research and Management. 2014;19(3):536859. 655 43. Chapman CR, Oka S, Bradshaw DH, Jacobson RC, Donaldson GW. Phasic pupil dilation 656 response to noxious stimulation in normal volunteers: relationship to brain evoked potentials 657 and pain report. Psychophysiology. 1999 Jan;36(1):44–52. 658 44. Liao HI, Kidani S, Yoneya M, Kashino M, Furukawa S. Correspondences among pupillary 659 dilation response, subjective salience of sounds, and loudness. Psychon Bull Rev. 2016 Apr 660 1;23(2):412–25. 661 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 33 45. Zekveld AA, Koelewijn T, Kramer SE. The Pupil Dilation Response to Auditory Stimuli: 662 Current State of Knowledge. Trends Hear. 2018;22:2331216518777174. 663 46. Wang CA, Munoz DP. A circuit for pupil orienting responses: implications for cognitive 664 modulation of pupil size. Current Opinion in Neurobiology. 2015 Aug 1;33:134–40. 665 47. Horing B, Sprenger C, Büchel C. The parietal operculum preferentially encodes heat pain 666 and not salience. PLOS Biology. 2019 Aug 12;17(8):e3000205. 667 48. Valentini E, Halder S, McInnerney D, Cooke J, Gyimes IL, Romei V. Assessing the 668 specificity of the relationship between brain alpha oscillations and tonic pain. NeuroImage. 669 2022 July 15;255:119143. 670 49. Lee IS, Necka EA, Atlas LY. Distinguishing pain from nociception, salience, and arousal: 671 How autonomic nervous system activity can improve neuroimaging tests of specificity. 672 Neuroimage. 2020 Jan 1;204:116254. 673 50. Mouraux A, Iannetti GD. Nociceptive laser-evoked brain potentials do not reflect 674 nociceptive-specific neural activity. J Neurophysiol. 2009 June;101(6):3258–69. 675 51. Gerber EM. permutest [Internet]. 2025 [cited 2025 Oct 5]. Available from: 676 https://de.mathworks.com/matlabcentral/fileexchange/71737-permutest 677 52. Hermans L, Calders P, Van Oosterwijck J, Verschelde E, Bertel E, Meeus M. An Overview 678 of Offset Analgesia and the Comparison with Conditioned Pain Modulation: A Systematic 679 Literature Review. Pain Physician. 2016 July;19(6):307–26. 680 53. Greene LC, Hardy JD. Adaptation of thermal pain in the skin. Journal of Applied 681 Physiology. 1962 July;17(4):693–6. 682 54. Chouchou F, Fauchon C, Perchet C, Garcia-Larrea L. An approach to the detection of pain 683 from autonomic and cortical correlates. Clinical Neurophysiology. 2024 Oct 1;166:152–65. 684 55. Eisenach JC, Curry R, Aschenbrenner CA, Coghill RC, Houle TT. Pupil responses and pain 685 ratings to heat stimuli: Reliability and effects of expectations and a conditioning pain 686 stimulus. Journal of Neuroscience Methods. 2017 Mar 1;279:52–9. 687 56. Chapman CR, Bradshaw DH, Donaldson GW, Jacobson RC, Nakamura Y. Central 688 noradrenergic mechanisms and the acute stress response during painful stimulation. J 689 Psychopharmacol. 2014 Dec;28(12):1135–42. 690 57. Szabadi E. Modulation of physiological reflexes by pain: role of the locus coeruleus. Front 691 Integr Neurosci [Internet]. 2012 [cited 2025 Dec 10];6. Available from: 692 http://journal.frontiersin.org/article/10.3389/fnint.2012.00094/abstract 693 58. Chapman CR, Oka S, Bradshaw DH, Jacobson RC, Donaldson GW. Phasic pupil dilation 694 response to noxious stimulation in normal volunteers: relationship to brain evoked potentials 695 and pain report. Psychophysiology. 1999 Jan;36(1):44–52. 696 59. Hauck M, Domnick C, Lorenz J, Gerloff C, Engel AK. Top-down and bottom-up modulation 697 of pain-induced oscillations. Front Hum Neurosci. 2015 July 2;9:375. 698 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 34 60. Peng W, Babiloni C, Mao Y, Hu Y. Subjective pain perception mediated by alpha rhythms. 699 Biological Psychology. 2015 July 1;109:141–50. 700 61. Huneke NTM, Brown CA, Burford E, Watson A, Trujillo-Barreto NJ, El-Deredy W, et al. 701 Experimental Placebo Analgesia Changes Resting-State Alpha Oscillations. PLoS One. 2013 702 Oct 11;8(10):e78278. 703 62. Li L, Wang H, Ke X, Liu X, Yuan Y, Zhang D, et al. Placebo Analgesia Changes Alpha 704 Oscillations Induced by Tonic Muscle Pain: EEG Frequency Analysis Including Data during 705 Pain Evaluation. Front Comput Neurosci. 2016 May 10;10:45. 706 63. Laufs H, Kleinschmidt A, Beyerle A, Eger E, Salek-Haddadi A, Preibisch C, et al. EEG-707 correlated fMRI of human alpha activity. NeuroImage. 2003 Aug 1;19(4):1463–76. 708 64. Murta T, Leite M, Carmichael DW, Figueiredo P, Lemieux L. Electrophysiological 709 correlates of the BOLD signal for EEG-informed fMRI. Hum Brain Mapp. 2015 710 Jan;36(1):391–414. 711 65. Szikszay TM, Adamczyk WM, Panskus J, Heimes L, David C, Gouverneur P, et al. 712 Psychological mechanisms of offset analgesia: The effect of expectancy manipulation. PLoS 713 One. 2023;18(1):e0280579. 714 66. Nahman-Averbuch H, Dayan L, Sprecher E, Hochberg U, Brill S, Yarnitsky D, et al. Sex 715 differences in the relationships between parasympathetic activity and pain modulation. 716 Physiol Behav. 2016 Feb 1;154:40–8. 717 67. Van Den Houte M, Van Oudenhove L, Bogaerts K, Van Diest I, Van den Bergh O. 718 Endogenous Pain Modulation: Association with Resting Heart Rate Variability and Negative 719 Affectivity. Pain Med. 2018 Aug 1;19(8):1587–96. 720 68. Adamczyk WM, Szikszay TM, Nahman-Averbuch H, Skalski J, Nastaj J, Gouverneur P, et 721 al. To Calibrate or not to Calibrate? A Methodological Dilemma in Experimental Pain 722 Research. The Journal of Pain. 2022 Nov 1;23(11):1823–32. 723 69. Nahman-Averbuch H, Coghill RC. Pain-autonomic relationships: implications for 724 experimental design and the search for an “objective marker” for pain. Pain. 2017 725 Nov;158(11):2064–5. 726 70. Chan HS. Occupational noise exposure; criteria for a recommended standard [Internet]. 727 National Institute for Occupational Safety and Health. Division of Biomedical and 728 Behavioral Science.;National Institute for Occupational Safety and Health. Education and 729 Information Division.; 1998 [cited 2023 July 22]. (DHHS publication). Available from: 730 https://stacks.cdc.gov/view/cdc/6376 731 71. Szikszay TM, Adamczyk WM, Luedtke K. The Magnitude of Offset Analgesia as a Measure 732 of Endogenous Pain Modulation in Healthy Participants and Patients With Chronic Pain: A 733 Systematic Review and Meta-Analysis. The Clinical Journal of Pain. 2019 Feb;35(2):189. 734 72. Szikszay TM, Adamczyk WM, Lévénez JLM, Gouverneur P, Luedtke K. Temporal 735 properties of pain contrast enhancement using repetitive stimulation. European Journal of 736 Pain. 2022;26(7):1437–47. 737 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 35 73. Zimmer K, Ellermeier W. Eine deutsche Version der Lärmempfindlichkeitsskala von 738 Weinstein. Zeitschrift für Lärmbekämpfung. 1997;44:107–10. 739 74. McCracken LM. Pain Vigilance and Awareness Questionnaire (PVAQ). APA PsycTests. 740 1997; 741 75. Meyer K, Sprott H, Mannion AF. Cross-cultural adaptation, reliability, and validity of the 742 German version of the Pain Catastrophizing Scale. J Psychosom Res. 2008 May;64(5):469–743 78. 744 76. Morin C, Bushnell MC. Temporal and qualitative properties of cold pain and heat pain: a 745 psychophysical study. Pain. 1998 Jan 1;74(1):67–73. 746 77. Gouverneur P, Li F, Luebke L, Szikszay TM, Roelen SD, Krajewski J, et al. eVAS: A user-747 friendly electronic Visual Analogue Scale. JOSS. 2025 Mar 20;10(107):6876. 748 78. Szikszay TM, Adamczyk WM, Luedtke K. The magnitude of offset analgesia as a measure 749 of endogenous pain modulation in healthy participants and patients with chronic pain. 750 Clinical Journal of Pain. 2019;35(2):189–204. 751 79. Faul F, Erdfelder E, Buchner A, Lang AG. G*Power 3: A flexible statistical power analysis 752 program for the social, behavioral, and biomedical sciences. Behavior Research Methods. 753 2007;39(2):175–91. 754 80. Anderson CJ. Central Limit Theorem. In: Weiner IB, Craighead WE, editors. The Corsini 755 Encyclopedia of Psychology [Internet]. 1st ed. Wiley; 2010 [cited 2025 Dec 10]. p. 1–2. 756 Available from: https://onlinelibrary.wiley.com/doi/10.1002/9780470479216.corpsy0160 757 81. R Core Team. R: A Language and Environment for Statistical Computing [Internet]. Vienna, 758 Austria: R Foundation for Statistical Computing; 2024. Available from: https://www.R-759 project.org/ 760 82. The Mathworks Inc. MATLAB version: 24.1.0.2578822 (R2024a) [Internet]. Natick, 761 Massachusetts, United States: The MathWorks Inc.; 2024. Available from: 762 https://www.mathworks.com 763 83. Luebke L, von Selle J, Adamczyk WM, Knorr MJ, Carvalho GF, Gouverneur P, et al. 764 Differential Effects of Thermal Stimuli in Eliciting Temporal Contrast Enhancement: A 765 Psychophysical Study. The Journal of Pain. 2024 Jan 1;25(1):228–37. 766 84. Szikszay TM, Melz N, Von Glasenapp B, Adamczyk WM, Luedtke K. Effects of stimulation 767 area and temperature rates on offset analgesia. PAIN Reports. 2022 Nov;7(6):e1043. 768 85. Benjamini Y, Yekutieli D. The Control of the False Discovery Rate in Multiple Testing 769 under Dependency. The Annals of Statistics. 2001;29(4):1165–88. 770 86. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful 771 Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B 772 (Methodological). 1995;57(1):289–300. 773 87. Brainard DH, Vision S. The psychophysics toolbox. Spatial vision. 1997;10(4):433–6. 774 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint 36 88. Oostenveld R, Fries P, Maris E, Schoffelen JM. FieldTrip: Open Source Software for 775 Advanced Analysis of MEG, EEG, and Invasive Electrophysiological Data. Computational 776 Intelligence and Neuroscience. 2011;2011(1):156869. 777 89. Mathôt S, Vilotijević A. Methods in cognitive pupillometry: Design, preprocessing, and 778 statistical analysis. Behav Res. 2023 Sept 1;55(6):3055–77. 779 90. Kret ME, Sjak-Shie EE. Preprocessing pupil size data: Guidelines and code. Behav Res. 780 2019 June 1;51(3):1336–42. 781 .CC-BY 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 17, 2025. ; https://doi.org/10.64898/2025.12.16.694610doi: bioRxiv preprint