Eye Movement-Related Eardrum Oscillations (EMREOs) Occur without Visual Input But Are Reduced during Closed Eyelids

Abstract

Recent investigations have shown that the tympanic membranes exhibit synchronous oscillations with each saccadic eye movement (Gruters et al., 2018), a phenomenon known as eye movement-related eardrum oscillations (EMREOs). However, the dependence of these saccade-associated EMREOs on ongoing visual activity remains to be elucidated. Given the direct projections from motor areas to primary auditory and visual cortices and the observation that EMREOs’ onset occurs concurrently with, or even precedes, saccades, we hypothesized that EMREOs would persist in the absence of visual stimulation. This report presents a study wherein 16 healthy male and female participants executed horizontal saccades under three distinct conditions: (1) in a well-lit environment, (2) in a darkened environment with eyes open, and (3) in a darkened environment with eyes closed. Ocular movements were quantified via electrooculography, and tympanic membrane oscillations were registered using in-ear microphones. The results demonstrated the presence of EMREOs concurrent with both visually guided and memory-guided saccades, although a late minor reduction in amplitude was observed in the “dark with open eyes” condition. Significant attenuation of EMREOs was evident when participants performed saccades with their eyelids closed, despite maintaining the same saccade amplitude and initial velocity. This amplitude reduction may reflect modulations in cortical states associated with predictive coding.

Significance Statement

Eye movement-related eardrum oscillations (EMREOs) occur synchronously with saccadic eye movements and persist even in the absence of visual input, indicating that visual stimuli do not solely drive these oscillations. However, the amplitude of EMREOs decreases slightly when saccades are made in darkness and strongly decreases with eyes closed, despite consistent saccade metrics. This attenuation suggests that EMREOs are modulated by cortical states associated with predictive coding rather than just motor commands or sensory feedback. These findings highlight a complex interaction between motor activity, sensory processing, and cortical predictive mechanisms.

Introduction

During natural conditions, changes in sensory organ activity often arise from self-initiated actions. This phenomenon, referred to as “active sensing” (Bajcsy, 1988; König and Luksch, 1998), contrasts with the traditional concept of “passive sensing,” which describes the perceptual processes triggered by input without any self-generated actions—e.g., light passively hitting the retina when the eyes are still. For instance, in the visual system, active sensing manifests when changes in retinal activity are induced by saccadic eye movements—movements that humans typically perform three to four times per second during normal visual perception (Ito et al., 2017). Numerous studies have documented the relationship between motor actions and sensory activity (Melloni et al., 2009; Wurtz, 2015), establishing this principle across various modalities, including olfaction (Wachowiak, 2011), audition (Morillon et al., 2015), and somatosensation (Gibson, 1962; Lederman et al., 1987; Blakemore et al., 1999). This principle of motor acts modulating sensory activity in a timely manner is observed not only in humans but also in other species such as rats, primates, fish, and bats, indicating the presence of a generalized mechanism in perception (Ulanovsky and Moss, 2008; Stamper et al., 2012; Hofmann et al., 2013; Krieger and Groh, 2015).

Moreover, active sensing is temporally coordinated across different modalities (Rizzo et al., 2020). Tomassini et al. (2015) demonstrated that visual perception is enhanced when the onset of a stimulus coincides with an unrelated hand movement. Similarly, Concha-Miranda et al. (2019) found that rats exhibited improved visual discrimination when the animal's arm movement triggered the onset of a visual stimulus. Even higher cognitive functions, such as working memory, are enhanced when subjects are allowed to initiate the onset of visual stimuli (Loyola-Navarro et al., 2022). However, the neuronal mechanisms underlying cross-modal active sensing remain largely unexplored. Recent findings by Gruters et al. (2018) indicated that the onset of eye movements is precisely synchronized with the brief oscillatory activity of the eardrum. In humans and monkeys engaged in saccadic eye movement tasks, these oscillatory eardrum movements commenced as early as 10 ms prior to the onset of the saccades and persisted for ∼50 ms. Notably, the amplitude of these eardrum oscillations was significantly correlated with the amplitude of the saccades, exhibiting larger amplitudes for horizontal displacements compared with vertical ones (Lovich et al., 2023b). Bröhl and Kayser (2023) further investigated the implications of eardrum oscillations for detecting near-threshold auditory stimuli, revealing that fluctuations in the amplitude of these oscillations were unrelated to detection performance. Studies support the idea of changes in the response of auditory neurons in the inferior colliculus when the eyes move, giving evidence of a neural relationship between eye movement and audition (Groh, et al., 2001).

A critical question regarding the neural mechanisms underlying the temporal synchronization between eye movements and eardrum oscillations is whether these eardrum oscillations necessitate ongoing visual activity or result from a purely top–down modulation associated with the motor commands of eye movements. Given that direct connections from motor areas to auditory and visual cortical regions have been demonstrated in rodents (Nelson and Mooney, 2016; Leinweber et al., 2017), we hypothesized that tympanic movements associated with eye movements would not rely on retinal visual activity. To test this hypothesis, we recruited subjects to perform voluntary saccades while measuring tympanic movements through intra-aural microphones and eye movements via electrooculography (EOG). We observed the anticipated tympanic movements associated with saccades, which were also evident during voluntary saccades conducted in complete darkness, with the eyes open. Additionally, we recorded tympanic movements during memory-guided saccades executed in darkness with the eyelids closed. Our findings indicate that tympanic oscillations associated with saccades can occur in the absence of visual input, although they were significantly diminished during saccades performed with closed eyelids.

Materials and Methods

Participants

Sixteen participants (11 females, 5 males) participated in this study. Eligible participants were adults (mean age 38.3 ± 16.7 years) with normal-hearing thresholds and adequate visual acuity to resolve two fixation points separated at 16° of the visual angle at 57 cm distance, without corrective lenses. All participants provided written informed consent, as approved by the Bioethics Committee of the Faculty of Medicine at the University of Chile (N° 013-2024). The consent form explicitly stated that participants could withdraw from the experiment at any time.

Experimental paradigm

The experiment was conducted in a sound-attenuated room. Participants were seated 57 cm from a screen displaying two fixed 2 cm diameter dots (2° of the visual angle), located at the center of gaze and the other separated horizontally to the right by 16° of the visual angle. Subjects performed horizontal saccades of 16° amplitude from the center fixation (outward) and then a similar saccade to the left (inward), shifting their gaze between the two fixation points at an interval of ∼2 s. The experiment consisted of three conditions, which were executed in an interleaved sequence. (1) Light open eyes (LOE): all participants performed alternate saccades between a central fixation dot and a second fixation dot located 16° to the right visual hemifield. The participants were instructed to pause for 2 s between each successive saccade. (2) Dark open eyes (DOE): fifteen subjects participated in this task, wearing opaque ski goggles adapted to prevent light passage, thus preventing external visual input. Participants were instructed to perform memory-guided saccades as if they were visualizing the two dots on the screen, keeping the same timing between saccades. (3) Dark closed eyes (DCE): 12 subjects participated in this task. This condition was identical to the DOE condition, except participants maintained their eyelids closed. Subjects underwent training in various conditions until they achieved consistent eye movement timing and amplitude. They performed these eye movements for 180 s under one condition, subsequently followed by an additional 180 s trial under a different condition. Conditions were interleaved throughout the task, ensuring a minimum of 9 min of total recording for each condition. Between conditions, participants were allowed to rest and briefly view the fixation points to refresh their memory of the target locations (Fig. 1).

Figure 1.

Experimental design. A, Participants were trained to make a simple horizontal saccade task, shifting the gaze between two dots separated by 16°. Background image was a natural image. B, Participants replicated the first task, wearing obscured goggles that impeded light passing, making memory-guided saccades. C, Subjects replicated the task in B, but with closed eyelids.

Eye movements and tympanic recordings

Horizontal eye movements were recorded using EOG with a three-electrode configuration: one electrode was positioned laterally adjacent to each eye, while a reference electrode was placed on the forehead. The EOG signals were differentially amplified using a Tektronix 26A2 differential amplifier in a AC-coupled mode. This configuration enables us to estimate the stability and variation of eye position. Tympanic membrane movements were captured using in-ear microphones (ER10-C, Etymotic Research) with an amplification level of 40 dB. These signals were further amplified and digitized utilizing a National Instruments analog-to-digital (A/D) converter and filtered between 10 and 80 Hz. Data acquisition and analysis were performed using the custom-designed software developed in our laboratory, utilizing LabWindows/CVI, which enabled the simultaneous recording and concurrent monitoring of eye movements and tympanic membrane oscillations. EOG and ear microphone signals were digitized at a sampling rate of 2,000 Hz. Eye data were bandpass filtered with a lower cutoff frequency of 0.1 Hz and an upper cutoff frequency of 30 Hz for all participants, except one case where a lower cutoff of 0.5 Hz was necessary due to signal quality considerations. Given intersubject variability in the amplitude of the EOG signals, thresholds for saccade detection were adjusted between 250 and 600 µV. Data were segmented into 200 ms epochs with a temporal resolution of 0.5 ms, starting 50 ms before saccade onset. The onset of each saccade was determined using an algorithm that analyzes saccade velocity. This algorithm identifies the initiation of a saccadic eye movement by detecting a marked increase in the EOG signal, with a threshold set at 100 µV over a 20 ms interval. A threshold scheme previously developed in the laboratory was utilized. Comparable performance was observed between eye tracking and EOG recordings using this method (Madariaga et al., 2023).

Statistical analysis

Design

The experiment employed a within-subject design, incorporating a within-subject factor with the three conditions. EOG signals were acquired for each condition and participant, recorded across two measurement datasets (eye and ear), with a uniform temporal resolution of 0.5 ms. The data matrices were structured in a subject × time format. To ensure comparability across conditions, the number of columns (time points) was equalized, and the number of rows (subjects) was adjusted through NaN padding, without data imputation. At each time point t, statistical tests included only subjects with valid data across all conditions. To identify differences among the three conditions over time, we conducted a point-by-point repeated–measure ANOVA, with condition treated as the within-subject effect and subject as the random effect. The F statistic, degrees of freedom, and corresponding p value were extracted from each test. Partial eta squared (η_p²) was calculated as an index of the effect size, reflecting the proportion of within-subject variance explained by condition. When the ANOVA indicated a significant effect, dependent post hoc comparisons between pairs of conditions were performed at each time point using paired Student's t tests. This analysis was augmented by direct pairwise comparisons across the entire time range, irrespective of the ANOVA outcome, to provide a detailed description of the time course of each specific contrast.

Correction for multiple comparisons

Given the multiple tests performed over time, p values were adjusted using the Benjamini–Hochberg (false discovery rate, FDR) procedure with α = 0.05, applied separately to the test families (ANOVA and each paired contrast). Consequently, statistical significance was interpreted based on q values (q < 0.05), which control for the FDR without compromising sensitivity to sustained effects over time. For each paired contrast, Cohen's d_z was calculated, defined as the mean of the within-subject differences between conditions divided by the standard deviation of those differences. This index quantifies the standardized magnitude and direction of the difference. Conventional references were utilized to interpret small (≈0.2), medium (≈0.5), and large (≈0.8) values.

Results

EMREOs occur during regular saccades and memory-guided saccades in the dark

In this study, participants were instructed to perform horizontal saccades between two points separated by 16° with their eyes open and the room lights on. An average of 120 saccades was recorded for each direction, performed by each subject. All participants exhibited eye movement-related eardrum oscillations (EMREOs) under three conditions: lights on with eyes open (LOE), dark with eyes open (DOE), and dark with eyes closed (DCE). “Nevertheless, the precision of the saccades varied. We included only those saccades that met the quality criteria defined by our SAFIDE metric” (Madariaga et al., 2023). “Consequently, the number of valid saccades used to compute the average response for each individual varied.” Inward and outward valid saccades were merged for each subject to calculate the average for each condition. For the LOE condition, participants executed an average of 240 saccades (SD 74.9); for the DOE condition, an average of 238 saccades (SD 121.2); and for the DCE condition, an average of 165.7 saccades (SD 139.2). Figure 2 presents an example subject, illustrating the average tympanic membrane movement and EOG signals for both ears across the three conditions. In this example, EMREOs can be observed in all cases, although apparent differences in amplitude are also noted. The EMREOs observed during the LOE condition resemble those reported in previous studies (Gruters et al., 2018) in terms of amplitude and frequency, as well as opposing phases for leftward and rightward saccades. Note that EOG signals are inverted for outward (rightward) and inward (leftward) saccades, as these signals reflect changes in the cornea–retina dipole. When the eye moves, there is a brief delay in the response of the retina and cornea due to the inertia of the ocular muscles and the fluid dynamics within the eye. This results in a gradual change in the dipole as the eye accelerates and decelerates during saccadic eye movements. The eye movements for the dark conditions (DOE and DCE) display a longer time course, reflecting potential differences in saccade acceleration. EMREOs are also present in these memory-guided saccades, demonstrating that while eye movements are still coupled with tympanic movements, ongoing visual activity is not necessary to trigger these EMREOs. During DCE, the eye movement dynamics appear to be similar to those of the DOE condition, but there is an apparent reduction in the amplitude of the EMREOs.

Figure 2.

Data for one example subject. Clear EMREOs can be seen, lasting about two or three cycles of a circa 30 Hz oscillatory activity. The vertical gray line marks the point of the saccade onset. AC-coupled eye movement signal in brown (millivolt), left eardrum signal voltage in blue (microvolt), and right eardrum in red (microvolt). A, LOE condition for inward saccades; (B) LOE condition for outward saccades; (C) DOE condition for inward saccades; (D) DOE condition for outward saccades; (E) DCE condition for inward saccades; (F) DCE condition for outward saccades. Left- and right-ear signals typically show counterphase dynamics.

Amplitudes differences in EMREOs

Both EOG amplitudes and left and right microphone signals exhibited natural variability among subjects, which can be attributed to the positioning of the EOG electrodes and the seal achieved with the microphone in each ear. To facilitate comparison of the signals across different subjects and between the ears of the same subject, we normalized each curve to the maximum value obtained during the outward saccade in the lights-on eyes-open (LOE) condition. Subsequently, we computed the grand average and standard deviation of EMREO amplitude across all participants for each time point within the 200 ms epoch and constructed average waveforms for each condition (Fig. 3). The resulting curves indicate that EMREOs on both ears displayed higher amplitudes in the LOE (Fig. 3A,B) and DOE (Fig. 3C,D) conditions, in contrast, a reduction in amplitude in the DCE condition (Fig. 3E,F).

Figure 3.

Average EMREOs from all subjects. The normalized amplitudes for outward and inward saccades are presented across all conditions. The vertical gray line at time 0 indicates the onset of the saccades. Colored shadow areas depict the standard deviation at each time point. A, LOE condition for inward saccades; (B) LOE condition for outward saccades; (C) DOE condition for inward saccades; (D) DOE condition for outward saccades; (E) DCE condition for inward saccades; (F) DCE condition for outward saccades. The lighter-colored curves represent the standard deviation for each dataset.

EMREOs occur in the absence of ongoing visual activity, with reduced amplitude

Previous studies have demonstrated that EMREOs observed during left and right saccades exhibit similar time courses and amplitude characteristics, with tympanic movements of the left and right ears showing comparable patterns, characterized by a counterphase pattern (Gruters et al., 2018; Lovich et al., 2023a). Therefore, we statistically compared EMREOs in our dataset by aggregating the signals from both ears and saccades for all subjects. Specifically, we inverted the left-ear waveforms for rightward and right-ear waveforms for leftward saccades, aligning all data into a common phase. The statistical analysis is presented in Figure 4, illustrating each pairwise comparison. We found significant differences across all conditions using a point-by-point repeated–measure ANOVA, which allows for multiple comparisons with a significance level set at p < 0.01. Notably, comparisons between the LOE and DOE conditions revealed significant differences only in the amplitude of a third peak, which occurs ∼80 ms. The LOE response exhibited three cycles of the oscillation, while the DOE only exhibited two. This result implies that the amplitudes of the EMREOs for LOE and DOE do not differ significantly during the first two cycles of the oscillations, demonstrating that EMREOs can happen in the absence of ongoing visual activity. These differences were consistent with a larger size effect, as shown in the bottom panel of Figure 4.

Figure 4.

A–C, EMREOs comparison of pairwise conditions for all participants. A point-by-point repeated–measure ANOVA was conducted, with condition treated as the within-subject effect and subject as the random effect. Significant differences (p < 0.01) between EMREOs are depicted as colored blocks at the bottom of each plot. The F statistic, degrees of freedom, and corresponding p value were extracted from each test. D–F, Partial eta squared (η_p²) was calculated as an index of the effect size, reflecting the proportion of within-subject variance explained by condition. Cohen's d_z values are plotted in a time-resolved manner below each condition comparison.

In contrast, comparisons between the LOE and DCE conditions showed significant differences in the first peak values, demonstrating that while EMREOs still occurred in both conditions, closing the eyelids significantly reduced the amplitude of tympanic oscillations. Similarly, we also observed that EMREOs during the DOE condition exhibited significantly larger amplitudes than those observed under the DCE conditions. In these two conditions, subjects performed memory-guided saccades in the absence of light, and closing the eyelids led to reduced tympanic oscillations. This result suggests that, in addition to the motor-related processes associated with coupling eye movements to tympanic oscillations, an additional factor should be considered to explain the differences observed between the DOE and DCE conditions.

Peak timing differences in EMREOs does not explain amplitude differences

The average data presented in Figure 4 reveal large values of the standard deviation, which was unexpected given that the data were normalized across subjects. Additionally, Figure 4 shows apparent differences in the average phase of the tympanic oscillations manifested as different peak timing. One possible explanation for this variability is that, across subjects, the peak time of the oscillations exhibits jitter. This was confirmed when we plotted the individual averages for one condition (LOE) across all subjects (Fig. 5A), although this also happened with the other conditions. We found that the curves were largely similar across subjects but exhibited apparent phase shifts. Therefore, we also compared the amplitudes of the EMREOs for the different conditions by measuring the peak values. As shown in the normalized averages for all subjects, we found significant differences between the LOE and other conditions. However, the differences were higher for the DCE condition than the LOE and DOE conditions (Fig. 5B). These results showed that when subjects were analyzed collectively, the oscillatory signals between LOE and DOE were comparable. However, individual subject analysis revealed that the DOE signal was slightly, yet significantly, smaller than the LOE signal.

Figure 5.

Comparison of peak-to-peak EMREOs signals for all subjects. A, Average EMREOs for all subjects for the LOE condition during outward saccades. B, The T test comparing EMREOs at the three conditions. Each point indicates the peak-to-peak amplitude for the average EMREO oscillation for each object, for all conditions. The red circle indicates the mean and SD. Empty circles depict outliers. Significance levels are noted in panel B.

One possible explanation for the differences in EMREO amplitude is that saccade amplitude and velocity may have varied across conditions, particularly since the saccades in the DOE and DCE conditions were memory-guided. We compared saccade signals across the three conditions to investigate whether differences in eye movement dynamics could account for the changes in EMREO amplitude (Fig. 6A–C). We began by applying a baseline correction to align all AC-coupled saccade traces to 0 mV at a −50 ms timestamp. Subsequently, we calculated the mean and standard deviation for each condition of the saccade waveforms. As illustrated in Figure 6A, the saccade peak in the LOE condition occurred ∼50 ms after movement onset, whereas in the DOE and DCE conditions, peak latencies exceeded 100 ms. This finding indicates that saccades performed without visual input were slower. Additionally, the amplitude and variability (as reflected by the standard deviation) were more significant in the DOE and DCE conditions (Fig. 6B,C), likely due to the inherent variability associated with memory-guided eye movements.

Figure 6.

A–C, Eye movement average and standard deviation. Normalization was computed with the peak amplitude for the left ear in the LOE condition. The normalized amplitudes for outward (red) and inward (blue) saccades are presented across all conditions. The vertical gray line at time 0 indicates the onset of the saccades. A, LOE condition; (B) DOE condition; (C) DCE condition. The lighter-colored curves represent the standard deviation for each dataset. D–F, Inward and outward saccades signals pooled together. Significant differences (p < 0.01) between saccade time courses are depicted as colored bars at the bottom of each plot. As in Figure 4, Cohen's d_z values are plotted in a time-resolved manner below each condition comparison.

To statistically compare saccade dynamics across conditions, we aligned the waveforms by inverting the traces of outward saccades, employing the same methodology used for the EMREO data. This approach allowed us to pool data for each condition, regardless of saccade direction. Again, we used a point-by-point repeated–measure ANOVA with a significance level set at p < 0.01 (Fig. 6D). The analysis revealed that saccade waveforms were initially similar across all conditions, with significant differences emerging ∼30 ms after saccade onset. A saccade of 16° typically lasts ∼45 ms (Baloh et al., 1975), indicating that the velocity of the AC-coupled EOG signal shown here does not correlate with the EMREO signal which peak occurred at ∼25 ms. In addition, the differences in EMREO amplitude and velocity were more significant between the DOE and DCE conditions than between the LOE and DOE conditions, especially during the first 60 ms. In contrast, the AC-coupled EOG signals under DCE and DOE conditions exhibited greater similarity, while its EMREOS signals showed large differences. Therefore, these kinematic differences cannot account for the reduced EMREO amplitude observed in these three conditions.

Discussion

The findings of this study demonstrate that EMREOs occur during both regular and memory-guided saccades, even in the absence of visual stimuli. Participants performed horizontal saccades of 16° under three conditions: lights on with eyes open (LOE), dark with eyes open (DOE), and dark with eyes closed (DCE). EMREOs were consistently observed across all conditions, with notable differences in amplitude, particularly showing higher amplitudes during LOE and DOE compared with reduced amplitudes in DCE. We also found that while EMREOs exhibited similar time courses and amplitude characteristics during saccades, the absence of visual input (as in DOE and DCE) resulted in longer time courses and differences in saccade dynamics. There were significant differences in EMREO amplitudes and phases between conditions, indicating that while visual activity is not necessary for EMREO generation, the closure of eyelids significantly diminishes EMREO amplitude. We provide strong corroboration for previous findings concerning EMREOs observed under various conditions. Notably, our results align with those initially reported by Gruters et al. (2018) and further explored by King et al. (2023) and Lovich et al. (2023a,b), all of which documented similar amplitude and frequency characteristics of EMREOs during horizontal saccades, particularly when guided by visual stimuli. Our findings are consistent with Sotero Silva et al. (2025), who demonstrated that the amplitudes of EMREOs can vary significantly between ipsilateral and contralateral saccades.

In contrast to the majority of previous studies that have predominantly focused on outward saccades, our investigation incorporated a comparative analysis of inward saccades. Notably, the inward saccades produced tympanic oscillations that were comparable in both amplitude and frequency to those observed in outward movements. This could indicate a potential mechanism that facilitates the alignment of changes in the visual scene with corresponding shifts in the auditory environment. However, it is worth noting that a recent study by Lovich et al. (2023b) found differences in EMREOs depending on the starting location of the saccade. Their findings concluded that these signals carry information about both the horizontal and vertical components of eye movements, displaying signatures related to changes in eye position as well as the absolute position of the eyes within the orbits. This highlights the complexity of the relationship between eye movements and tympanic oscillations, suggesting that further research is needed to elucidate these interactions fully.

What about the role of EMREOs in sensory processing? Bröhl and Kayser (2023) investigated the relationship between EMREOs and the detection of near-threshold auditory stimuli in human participants. Their research revealed that, across three experiments, the detection of sounds presented to the left versus the right ear was unaffected by the timing of sound presentation relative to saccade onset nor by the amplitude or direction of the saccades. This led to the conclusion that EMREO amplitude does not correlate with detection performance, at least in contexts where monaural cues are sufficient. Given the direct connections from motor areas to auditory and visual cortical regions (Nelson and Mooney, 2016; Leinweber et al., 2017), we hypothesized that tympanic movements associated with eye movements would not depend on retinal visual activity. These studies refer to direct projections from motor areas to the sensory cortex in rodents. In mammals, direct connections have only been identified between the motor and primary sensory cortex (Jones and Powell, 1969; Stepniewska et al., 1993). Nevertheless, this does not preclude the possibility that the motor cortex can indirectly, or through other areas such as the superior colliculus or frontal eye fields, send projections to the auditory or visual systems. Furthermore, we conjecture that these signals do not represent a corollary discharge (efferent copy) of a motor command but rather a modulatory signal that can modify timing and synchronous activity in these systems (Ito et al., 2011).

Our findings support this hypothesis, as EMREOs observed in the dark exhibited similar characteristics when subjects performed memory-guided saccades compared with visually guided saccades. Notably, we found that EMREOs in the dark were only minimally statistically different from those occurring under light conditions. This result indicates that ongoing visual activity is not necessary for EMREOs to be triggered simultaneously by eye movements, suggesting the involvement of a common motor process or another central generating signal (Gruters et al., 2018).

It has been reported that memory-guided saccades performed in the dark are slower than those executed in a lit environment (Felßberg and Dombrowe, 2018) and exhibit significantly reduced peak velocities, accompanied by markedly increased durations (Smit et al., 1987). Our AC-coupled recording data are consistent with these studies. In humans, a 16° saccade typically lasts ∼45–55 ms (Bahill et al., 1975; Leigh and Zee, 2015), by which time most of the EMREO response is completed. A key finding in our study is that, despite these potential differences, EMREOs during light and dark conditions with the eyes open were similar, diverging only slightly around the top cycles of the oscillations. Our results align with a recent report by Abbasi et al. (2025), which demonstrated that EMREO characteristics were comparable under both light and dark conditions. In our study, we also demonstrate that in darkness, closing the eyelid results in a significant reduction of these oscillations. Despite these significant differences in speed, the EMREOs were much more similar across the two open eyes conditions. Furthermore, while the eye movement recordings were notably more comparable between the DOE and DCE conditions, the amplitude of the EMREOs in the latter case was significantly smaller. Thus, differences in saccade velocity mainly occur during time windows in which the EMREO has already dissipated. Interestingly, the pattern of similarity among EMREOs across the three conditions mirrors the pattern of saccade velocity only during the period when the EMREO is still present. With closed eyelids, it is expected that the kinetics of eye movement would slow down; indeed, it has been shown that sustained eye closure inhibits the activity of omnipause neurons, which hypothetically should weaken the premotor burst neurons that control saccades (Shaikh et al., 2010).

Our results demonstrated that EMREOs persist in the absence of visual input but are significantly reduced when the eyelids are closed. This pattern suggests a modulation by neural mechanisms associated with active sensing and predictive processing. During active sensing, motor commands not only generate movement but also prepare sensory systems to anticipate expected stimuli. When the eyes move in darkness, an internal prediction of the visual target may still be generated because of the expectation, maintaining the coupling between oculomotor and auditory systems. However, when the eyelids are closed, this predictive process likely diminishes due to the absence of visual expectation, resulting in reduced EMREO amplitude. This interpretation aligns with evidence that the cortical state and arousal modulate sensory processing. During eye closure, neural connectivity patterns change significantly, with increased alpha power reflecting disengagement from external stimuli (Barry and De Blasio, 2017). In contrast, with eyes open (even in darkness), the brain may still generate visual predictions related to target localization. Given that the visual and auditory systems cooperate in cross-modal perception (Shams and Beierholm, 2010), this suggests that EMREOs may represent a peripheral correlate of these predictive mechanisms, enhancing spatial localization and aiding multisensory integration.

Such modulation likely originates from central predictive processes, as proposed in previous work (Gruters et al., 2018). Within the predictive coding framework, the brain continuously generates and updates internal models to minimize the discrepancy between expected and actual sensory input (Rao and Ballard, 1999; Clark, 2013). When visual input is unavailable, prediction errors decrease, leading to reduced top–down modulation. This interpretation is consistent with studies showing that motor-related signals can precisely modulate early sensory processing in both auditory and visual cortices (Ito et al., 2011, 2022; Concha-Miranda et al., 2019; Leszczynski et al., 2023). Therefore, during eyelid closure, the absence of visual prediction may weaken central modulatory signals related to eye movements, leading to diminished EMREO amplitude, whereas in dark conditions, the brain still anticipates visual changes, maintaining a stronger motor–sensory interaction.

In summary, our study investigated the potential influences of saccade dynamics on the amplitude of EMREOs, concluding that differences in saccade velocity and amplitude alone do not fully explain the observed reductions in EMREO amplitude in the DCE condition. We also demonstrated that ongoing visual activity is not necessary for EMREOs to occur; however, the precise mechanisms that modulate this signal, particularly in relation to saccade speed and duration, require further exploration. Overall, these findings enhance our understanding of the relationship between eye movements and auditory processing, suggesting that EMREOs may function as a mechanism for integrating visual and auditory sensory activity.