Influence of a rhythmic context on the foreperiod effect: A behavioral and eye-tracker approach Rafael Román-Caballero, Elisa Martín-Arévalo, and Juan Lupiáñez Mind, Brain, and Behavior Research Center (CIMCYC), University of Granada, Granada, Spain; and Department of Experimental Psychology, University of Granada, Granada, Spain Introduction Attention fluctuates over time. In the long-term, it seems to wax and wane throughout the day concerning the sleep-wake cycle (Oken et al., 2006) and, when exercised in a sustained way, it is usual to observe a decrease in performance (i.e., vigilance decrement; Luna et al., 2018; Thomson et al., 2015). Several theories (Jones, 1976; Large & Jones, 1999; Fiebelkorn & Kastner, 2019) extend this idea and propose that attention acts as an oscillatory system that alternate, following an intrinsic period (1–4 Hz), between moments of enhanced perceptual sensitivity with instants of attentional disengagement. Moreover, fluctuations of attention can be synchronized with rhythmic environmental stimuli through entrainment (Dynamic Attending Theory, DAT; Jones, 1976). In particular, the regularity of the external rhythm is a key factor in the degree of attentional coupling. Periodic sequences of stimuli promote a tight phase coupling and allow the perceiver to anticipate the nature (e.g., modality, quality) and precise timing of a future event (Henry & Herrman, 2014). Therefore, the presence of a rhythmic context can benefit the perceptual discrimination and response preparation for stimuli occurring in phase with the rhythm (Henry & Herrman, 2014). On the other hand, the DAT proposes that the attentional oscillation is self-sustained, that is, the oscillator decays back to its intrinsic period when the external rhythm disappears. However, evidence about the lasting effect of entrainment is still scarce. A recent study (Trapp et al., 2020) suggests that the benefits of a rhythmic background can remain some minutes after it discontinues. One example of the impact of rhythmic contexts on perceptual performance can be found in Chang et al. (2019). They presented deviants tones (higher or lower in pitch) embedded in rhythmic (isochronous) or arrhythmic (non-isochronous) sequences of standard tones (500 Hz). Pitch deviations (500 ± Δ Hz) were near to the higher and lower discrimination thresholds. During the task, participants benefitted from rhythmic contexts and showed faster responses and superior pitch discrimination for target tones than in arrhythmic sequences. Moreover, electroencephalographic (EEG) signals from bilateral auditory cortices (filtered with a dipole source model) were entrained to rhythms. Delta (1–3 Hz) frequency bands showed power peaks at 2 Hz and harmonic frequencies (4 and 6 Hz) exclusively in rhythmic sequences, according to the isochronous presentation of stimuli each 500 ms. Rhythmic regularity also modulated delta phase, beta (15–25 Hz) power, and delta-beta coupling. More importantly, those EEG changes were correlated with changes in behavioral outcomes (i.e., pitch sensitivity and response speed). Rhythm is not the only type of temporal structure that can support attention (Nobre & van Ede, 2018). One of them is the passage of time itself when there is the certainty that the target will appear after an interval (i.e., when the target is presented in all trials). The probability that an event will occur given that it has not yet occurred increases over time and, therefore, the temporal expectation. As a consequence, reaction time (RT) decreases as the foreperiod (i.e., the interval between the cue and the target) becomes longer, which is usually known as the foreperiod effect (Niemi & Näätänen, 1981; Triviño et al., 2010). Previous studies have explored the interaction between rhythmic structures and the foreperiod effect. Martin et al. (2005) found clear effects of foreperiod and rhythm in a task in which the participants had to discriminate the color of circles. In half of the trials, a single white circle preceded the target (serving as a warning cue), leaving a silent interval of 1000, 1500, or 2000 ms. In the other half, that interval was filled with extra white circles forming an isochronous sequence (with an interonset interval, IOI, of 500 ms between each cue, and between the last of them and the target). Although in both conditions RTs decreased as the cue-target intervals became longer, in the rhythmic condition the reduction was more pronounced. The advantage of the isochronous context disappeared when its pace was drastically slowed down, with an IOI of 4000 ms (Martin et al., 2005; Experiment 2). However, one possibility is that the improved performance in the rhythmic condition could be driven by a higher alerting state due to the greater number of cues. In their third experiment, Martin et al. found that both isochronous sequences and complex rhythmic context lead to allocate attention to moments in time in contrast to random sequences. Similarly, Ellis and Jones (2010) observed a foreperiod effect with RTs in an auditory pitch discrimination task when a scrambled context preceded the target. That is, participants were faster responding to targets that appeared 1000 ms after than those that occurred 500 ms after (and even faster than those appearing after only 250 ms). Interestingly, the use of a complex but metrical rhythm (with a beat each 1000 ms) before the target reversed the foreperiod effect. It was interpreted as a consequence of the rhythm reducing the uncertainty in all the temporal moments, as it elicited increases in temporal expectation to 1000 ms and their subdivisions, 500 and 250 ms. Thus, the better precision to estimate short intervals than longer (250 vs. 1000 ms) would explain the increase of RTs over time. Therefore, simple and complex rhythmic contexts seem to influence the foreperiod effect, at least with response speed outcomes. Nevertheless, in the previous studies, the target always appeared in phase with the period of the entrained rhythm and just one period after the cueing sequence (e.g., 500 ms following the offset of the 500-ms isochronous sequence; Martin et al., 2005, Experiment 1). To probe the oscillatory nature of the rhythm entrainment and the extension of its effect over time, one possibility is to include targets out of phase (e.g., 250 or 750 ms in the last example) along with cue-target intervals of several periods (e.g., since 250 to 2250 ms). Another issue with the aforementioned studies is the presence of remarkably low perceptual demands (Ellis & Jones, 2010, Experiment 2; Martin et al., 2005, Experiment 1) or high (75% of accuracy; Ellis & Jones, 2010, Experiment 1). With those designs, a great percentage of participants might respectively have reached a ceiling effect or performed under the discrimination threshold. Thus, it seems relevant to adjust the demands to the perceptual skills of each participant to assess with enough sensitivity the influence of rhythm. Finally, along with behavioral measures, recent research has shown that temporal expectation can be observed through eye-tracker measures. In two visual tasks of sustained attention, Shalev and Nobre (2020) found that the discrimination and processing speed of the participants improved in the rhythmic blocks (isochronous sequence with an IOI of 3500 ms), compared to blocks where stimuli appeared after a variable interval (2500–4500 ms). In addition, rhythmic regularity reduced pupil size across the tasks, suggesting an arousal adjustment to the conditions of high temporal certainty and the achieving of a more efficient energetic state. Interestingly, pupil size also exhibited a foreperiod-like pattern in variable blocks, with transient increases of the pupil as a function of time (i.e., larger pupil size in intervals close to 4500 than 2500 ms). However, the existence of a single cue-target interval did not allow assessing that pupil-dilation index of the foreperiod effect under a rhythmic context. It has been proposed that pupil size changes can be a sensitive measure of temporal expectation as they vary according to the locus coeruleus-noradrenergic response (via parasympathetic and sympathetic pathways: pupil constriction and dilation, respectively; Wang & Munoz, 2015). A reduced uncertainty such as that in rhythmic contexts would promote reduced tonic arousal along with marked short-scale phasic increases contingent on relevant stimuli (Aston-Jones & Cohen, 2005; Shalev & Nobre, 2020). On the other hand, high uncertainty (i.e., arrhythmic contexts) favors a continuous state of high tonic arousal and slight transient increases. In another visual vigilance task, Dankner et al. (2017) found that saccades were inhibited prior to the target onset at predictable compared with less predictable moments, given by the use of fixed (2000 ms) and variable (1000, 1500, 2000, or 2500 ms) IOI, respectively. The prestimulus saccadic inhibition (PSSI) was absent in most participants with attention-deficit/hyperactivity disorder (ADHD) and it was reduced in those (both neurotypical and with ADHD) with poorer sustaining attention scores. Thus, the PSSI-predictability effect is a sensitive measure of temporal expectation that can be related to suppression mechanisms to prevent image motion on the retina. By reducing eye motion for a short interval of time prior to the target, it may prevent transient visual distortions and enhance acuity. Even more, it has been observed an increase in saccade rates 200 ms after the saccadic suppression, which correlated with PSSI: the higher the PSSI, the higher the post-stimulus saccade rate (Dankner et al., 2017). This post-stimulus rate increase was interpreted as a rebound-like effect that could have pervasive consequences if the target appeared in an unpredictable moment, a few hundred milliseconds later. The PSSI effect has not yet been investigated with non-visual paradigms, such as auditory foreperiod tasks, as in the present research project. The present study aims to investigate the modulating effect of rhythms over the foreperiod effect, as well as its extension over time, in behavioral outcomes (RTs and accuracy) and eye-tracker measures (pupil size and PPSI-predictability effect). We will use a pitch discrimination task with a sequence of standard tones before the target onset. In one block, the IOI will be fixed at 500 ms forming a rhythmic context, while in another block the IOI will be randomly varied between 250 and 750 ms (arrhythmic context). To adjust the perceptual demands in each participant, at the same time that we leave a large proportion of correct responses, a 1-up/4-down procedure will be used to estimate the 84.1% correct pitch discrimination threshold (Levitt, 1971). Moreover, we will use variable cue-target intervals up to 2250 ms and targets in phase with the rhythmic sequence (i.e., n 500 ms after the last tone of a 500-ms isochronous sequence) and targets out of phase (i.e., n 500 +/− 250). This manipulation will allow observing whether attention is devoted more to the moments predicted by the entrained rhythm than those out of phase. Hypotheses We expect a classic foreperiod effect will appear in response speed and accuracy. That is, the later the target onset, the faster the response and the better discrimination (higher accuracy) of the participants. Moreover, and crucially for this study, the presence of a rhythmic context before the target will enhance the perceptual sensitivity and the preparation to the critical moment, and will modulate the foreperiod effect. The presentation of several isochronous standard tones will lead to better performance at moments in phase with the entrained rhythm (500, 1000, 1500, and 2000 ms after the last standard tone). Moreover, we predict that although those effects will extend to the later in-phase moments (1000, 1500, and 2000 ms), it is probable that the alignment with the rhythm will decay over time. Regarding eye-tracker measures, we also expect a modulation of the rhythmic context over the foreperiod effect. Specifically, we hypothesize a decrease of the pupil size at the beginning of the sequence of standard tones along with an increase before the target onset, that will be more marked in the rhythmic block. Furthermore, whether the PSSI is a cross-modal effect, it would be possible to find a similar block modulation over the PSSI: a higher saccade rate just before (−100 ms) the appearance of in-phase targets during the variable-IOI block than in the fixed-IOI block. Finally, the impact of the rhythmic context might decay with the pass of time, being greater in the first periods than in the last ones. Method Participants We have selected the three-way interaction [block (2) × period (4) × phase (2); for details, see below] to estimate the sample size in the present experiment. In a previous pilot study with 11 participants, we found an effect size of f2 = 0.12. Using the Superpower R package (Caldwell et al., 2020), we have found that at least 43 participants will be necessary for a small-to-medium effect size of f2 = 0.09 (similar to that observed in the pilot) with an alpha of .05 and a power of .80. Accordingly, we have chosen a sample size of 45 participants due to the possibility of some missing participants as a consequence of technical issues or misunderstandings of the instructions. All the participants will be students from the University of Granada that will receive 15€ for participating. All will have a self-reported normal or corrected-to-normal vision and will be naïve as to the purpose of the experiment. The study will be conducted following the ethical guidelines laid down by the University of Granada (536/CEIH/2018), in accordance with the ethical standards of the 1964 Declaration of Helsinki (last update: Seoul, 2008). Outlier detection will be based on performance (i.e., mean reaction times and accuracy) identified as poor in terms of meeting all the following indices: standard deviation from the mean of the sample (> 2.5), studentized deleted residuals (> 3), and Cook’s Di (> 4/n). Apparatus and Stimuli The stimulus presentation, timing, and data collection will be controlled by a program written using E-Prime 2.0 (Schneider et al., 2002), run on a standard Pentium 4 PC. During the task, a light gray fixation cross will be presented on the center of a 17-inch widescreen monitor with a 1,280 × 1024-pixel resolution and 60 Hz of refresh rate. Auditory stimuli will be sine wave tones created with Audacity. Tones will last 60 ms with 10 ms rise and 10 ms fall times. Standard tones will have 500 Hz of frequency and will be presented at 70 dB. Whereas, comparison higher and lower tones will have 500 ± Δ Hz and 78 dB, where Δ represents the 84.1% correct pitch discrimination threshold estimated with a 1-up/4-down procedure (Levitt, 1971). We will use a video-based eye-tracker (EyeLink 1000 Plus, SR Research, Ontario, Canada) to measure pupil diameter monocularly as well as microsaccades and blinks at 1000 Hz. Procedure The study will be carried out after preregistration, during June of 2021. The experiment will take place in a dimly-light room and using glasses or contact lenses if needed. The session will last approximately 1 hour and 15 minutes. Prior to the experimental task, participants will complete a questionnaire (Appendix A) with items about demographic information, the fatigue and vigor scales of the Spanish abbreviated version of the Profile of Mood States questionnaire (Andrade et al., 2013; original version, Grove & Prapavessis, 1992), musical training background, and exclusion criteria: vision or hearing problems, neurological or psychiatric illness, diabetes, drug abuse, or chronic use of psychoactive medication. After that, participants will sit 64 cm from the monitor, and a chin rest will be used to keep their head still. First, the individual pitch discrimination threshold will be estimated using a 1-up/4-down procedure (84.1% correct pitch discrimination, following Levitt, 1971). In each trial, a standard tone will be presented followed by a comparison tone 500 ms after. The participants will be required to judge whether they perceive the second stimulus higher or lower in pitch than the standard (pressing “8” or “2” keys, respectively). The estimation of the discrimination threshold will comprise two blocks, one for the upper threshold and beginning with a 515-Hz tone and a lower threshold block beginning with a tone of 485 Hz of frequency (the order of the blocks will be counterbalanced). We will estimate the discrimination threshold by averaging the frequency of all the peaks and valleys in each type of block. Peaks and valleys are the frequency values at which a change in the perceptual trend occurs (for instance, when the individual perceives that the second tone has a lower pitch after a series of trials in which it was judged as higher). To reach a reliable discrimination threshold in each participant, the 1-up/4-down procedure will finish after the occurrence of ten peaks or valleys in each block. After the estimation of the discrimination threshold, participants will carry out the pitch discrimination task. The instructions will appear on the screen and will be explained to the participant by the experimenter, who will also answer any questions about the procedure. In each trial, the participants will hear five, six, or seven standard tones with an IOI of 500 ms or a random value between 250 and 750 ms, depending on the block (fixed vs. variable block; see Figure A in the OSF project https://osf.io/cw93r/). After the sequence of standard tones, a comparison tone will appear with an IOI randomly selected from 250, 500, 750, 1000, 1250, 1500, 1750, 2000, and 2250 ms. The participants will be required to discriminate as fast and accurately as possible whether the comparison tone has a higher or lower pitch than the standard tones, pressing “z” or “m” (the meaning of what indicates each key will be counterbalanced across participants). The discrimination task will begin with two practice blocks of eight trials each. Feedback for incorrect responses will be provided only during practice blocks. Then, participants will complete one block of 252 trials with fixed IOIs of 500 ms for the sequence of standard tones or with variable IOIs between 250 and 750 ms (the order will be counterbalanced across participants). As there are nine possible IOIs between the last standard tone and the target, and two pitches, there will be 14 trials for each type (with the same IOI and the same pitch). Subsequently, participants will be asked about the difficulty of the last task block (Appendix B). After 1 or 2 minutes of rest, they will carry out a second task block, with variable IOIs if the first block was with fixed IOIs or with fixed if the previous were variable. Finally, the participants will complete the question about the difficulty. The experimental task will compromise 28 trials for each experimental condition. Design and Statistical Analyses The experiment has a (2 × 4 × 2) three-factor repeated-measures design: block (fixed vs. variable), period (500–750/1000–1250/1500–1750/2000–2250 ms), and phase (in-phase: 500, 1000, 1500, and 2000 ms; vs. out-of-phase target: 750, 1250, 1750, and 2250 ms). For behavioral measures, we will conduct repeated-measures ANOVAs with mean RTs and accuracy as dependent variables. Moreover, paired t-tests will be used for comparison between conditions. We expect a higher number of misses in the cue-target interval of 250 ms, and, subsequently, a reduced proportion of trials to be analyzed. Regarding eye-tracker measures, blink intervals will be identified following Hershman et al.’s procedure (2018) and will be removed from the analyses. For analyses with trial time series, raw pupil size will be converted to z scores, by calculating the mean and standard deviation of each separate block for each participant. Then, the time series before the target onset of both blocks will be compared with permutation t-tests based on 10,000 samples. In addition, the slopes of the increase in the pupil size during foreperiod will be analyzed with a regression model, including block (fixed vs. variable) as a categorical factor. For saccade analyses, gaze-position data will be segmented into epochs from –500 to 500 ms relative to the target onset. For saccades detection, we will use a similar procedure to Dankner et al. (2017): saccades are identified by eye moments that exceed a threshold of 4 standard deviations from the median velocity during 7 ms (seven consecutive eye-position samples). The dependent variable will be the saccade rate at –100 to 0 ms relative to the onset of targets for PSSI and the saccade rate at 0 to 500 ms for the post-stimulus effect. We will exclude every trial in which a blink interval overlaps with that interval or in which a saccade larger than 3° occurs at that interval. Permutation t-tests for each time point based on 10,000 samples will be used. In addition, we will use two repeated-measures ANOVAs with block (fixed vs. variable) and phase (in phase vs. out of phase) as factors: one with the mean saccade rates at the pre-stimulus interval (–100 to 0 ms) as dependent variable, and another with the mean saccade rates at the post-stimulus interval 200 to 500 ms). References Andrade, E., Arce, C., De Francisco, C., Torrado, J., & Garrido, J. (2013). Abbreviated version in Spanish of the POMS questionnaire for adult athletes and general population. Revista de Psicología del Deporte, 22(1), 95–102. Aston-Jones, G., & Cohen, J. D. (2005). An integrative theory of locus coeruleus- norepinephrine function: Adaptive gain and optimal performance. Annual Review of Neuroscience, 28(1), 403–450. https://doi.org/10.1146/annurev.neuro.28.061604.135709 Caldwell, A., Lakens, D., DeBruine, L., & Love, J. (2020). Superpower: Simulation-based power analysis for factorial designs. R package version 0.0.3. https://cran.r-project.org/web/packages/Superpower/ Chang, A., Bosnyak, D. J., & Trainor, L. J. (2019). Rhythmicity facilitates pitch discrimination: Differential roles of low and high frequency neural oscillations. NeuroImage, 198, 31–43. https://doi.org/10.1016/j.neuroimage.2019.05.007 Dankner, Y., Shalev, L., Carrasco, M., & Yuval-Greenberg, S. (2017). Prestimulus inhibition of saccades in adults with and without attention-deficit/hyperactivity disorder as an index of temporal expectations. Psychological Science, 28(7), 835–850. https://doi.org/10.1177/0956797617694863 Ellis, R. J., & Jones, M. R. (2010). Rhythmic context modulates foreperiod effects. Attention, Perception, & Psychophysics, 72(8), 2274–2288. https://doi.org/10.3758/BF03196701 Fiebelkorn, I. C., & Kastner, S. (2019). A rhythmic theory of attention. Trends in Cognitive Sciences, 23(2), 87–101. https://doi.org/10.1016/j.tics.2018.11.009 Grove, J. R., & Prapavessis, H. (1992). Preliminary evidence for the reliability and validity of an abbreviated profile of mood states. International Journal of Sport Psychology, 23(2), 93–109. Henry, M. J., & Herrmann, B. (2014). Low-frequency neural oscillations support dynamic attending in temporal context. Timing & Time Perception, 2(1), 62–86. https://doi.org/10.1163/22134468-00002011 Hershman, R., Henik, A., & Cohen, N. (2018). A novel blink detection method based on pupillometry noise. Behavior Research Methods, 50(1), 107–114. https://doi.org/10.3758/s13428-017-1008-1 Jones, M. R. (1976). Time, our lost dimension: Toward a new theory of perception, attention, and memory. Psychological Review, 83, 323–355. Large, E. W., & Jones, M. R. (1999). The dynamics of attending: How people track time-varying events. Psychological Review, 106(1), 119–159. https://doi.org/10.1037/0033-295X.106.1.119 Levitt, H. (1971). Transformed up‐down methods in psychoacoustics. The Journal of the Acoustical Society of America, 49, 467–477. Luna, F. G., Marino, J., Roca, J., & Lupiáñez, J. (2018). Executive and arousal vigilance decrement in the context of the attentional networks: The ANTI-Vea task. Journal of Neuroscience Methods, 306, 77–87. https://doi.org/10.1016/j.jneumeth.2018.05.011 Martin, T., Egly, R., Houck, J. M., Bish, J. P., Barrera, B. D., Lee, D. C., & Tesche, C. D. (2005). Chronometric evidence for entrained attention. Attention, Perception & Psychophysics, 67(1), 168–184. https://doi.org/10.3758/BF03196701 Niemi, P., & Näätänen, R. (1981). Foreperiod and simple reaction time. Psychological Bulletin, 89, 133–162. https://doi.org/10.1037/0033-2909.89.1.133 Nobre, A. C., & Van Ede, F. (2018). Anticipated moments: temporal structure in attention. Nature Reviews Neuroscience, 19(1), 34–48. https://doi.org/10.1038/nrn.2017.141 Oken, B. S., Salinsky, M. C., Elsas, S. M. (2006). Vigilances, alertness, or sustained attention: physiological basis and measurement. Clinical Neurophysiology, 117(9), 1885–1901. http:// dx.doi.org/10.1016/j.clinph.2006.01.017. Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime (Version 2.0). [Computer software and manual]. Pittsburgh, PA: Psychology Software Tools Inc. Shalev, N., & Nobre, A. C. (2020). Eyes wide open: regulation of arousal by temporal expectations. bioRxiv. https://doi.org/10.1101/2020.07.24.219428 Thomson, D. R., Besner, D., & Smilek, D. (2015). A resource-control account of sustained attention: Evidence from mind-wandering and vigilance paradigms. Perspectives on Psychological Science, 10(1), 82–96. https://doi.org/10.1177/174569161455 Trapp, S., Havlicek, O., Schirmer, A., & Keller, P. E. (2020). When the rhythm disappears and the mind keeps dancing: sustained effects of attentional entrainment. Psychological Research, 84(1), 81–87. https://doi.org/10.1007/s00426-018-0983-x Triviño, M., Correa, Á., Arnedo, M., & Lupiáñez, J. (2010). Temporal orienting deficit after prefrontal damage. Brain, 133(4), 1173–1185. https://doi.org/10.1093/brain/awp346 Wang, C. A., & Munoz, D. P. (2015). A circuit for pupil orienting responses: implications for cognitive modulation of pupil size. Current Opinion in Neurobiology, 33, 134–140. https://doi.org/10.1016/j.conb.2015.03.018