The moment of awareness influences the content of awareness in orientation repulsion

Through the neurally evolving process of dynamic contextual modulation of perceptual contents, it remains unclear how the content of awareness is determined. Here we quantified the visual illusion of orientation repulsion, wherein the target appears tilted against the surrounding ’ s orientation, and examined whether its extent changed when the target awareness was quickened by a preceding flanker. Independently of spatial cueing, repulsion was reduced when the flanker preceded the target by 100 ms compared with when they appeared simultaneously. We confirmed that the preceding flanker quickened the awareness of a nearby target relative to distant ones by 40 ms. Furthermore, the preceding flanker that was greater than 7 degrees away from the target still evoked such reduction of repulsion. These findings imply that the content of awareness is determined by the temporal interaction of two distinct processes: one controls the moment of awareness, and the other represents the perceptual content.


Introduction
Although our retinae are constantly exposed to excessive visual information, only the final outputs of content formation processes taking a few hundred milliseconds are consciously accessible ( Ögmen & Breitmeyer, 2006; see also Herzog et al., 2020;Shimojo, 2014).Previous psychophysical experiments using backward masking paradigms have corroborated this microgenetic view and revealed the time course of pre-conscious processing (e.g., Breitmeyer & Ögmen, 2006;Bachmann & Francis, 2014).For example, Breitmeyer et al. (2006) found that the stimulus onset asynchrony (SOA) where metacontrast masking occurs most effectively was longer for a contrast matching task than for a contour discrimination task even though the stimulus configurations for these two tasks were similar.These findings demonstrate that contours are formed before surfaces are filled-in, although both appear integrated as a single object to us in normal conditions without masking.Another example is from Ringach and Shapley's (1996) experiment on boundary completion.They quantified a shape discrimination performance for Kanizsa figures composed of four "pacmen" inducing modal or amodal completion.If a local pin-wheel mask interfered each of the inducers, the performance increased with SOA, reaching the asymptote at 117 ms.However, if another Kanizsa figure was added as a global mask, an SOA around 300 ms was necessary for the performance to reach the asymptote.These findings suggest two phases of boundary completion, the extraction of local features and the global integration of them.
However, through such internal dynamics implicated in a number of previous studies, it remains unclear what determines the perceptual content we will eventually become aware of and when it is determined.If we assume that certain aspects of the content internally undergoing neural evolution can phenomenally spill over onto the consciousness in some way or other (e.g., Block, 2011;Lamme, 2004; but see Cohen & Dennett, 2011;Knotts et al., 2019), determining the temporal boundary between pre-conscious and conscious processing will become a thorny problem.However, at least when the reportable content of awareness is focused on, the moment at which the perceptual content arises in one's awareness can be psychophysically defined (see e.g, Herzog et al., 2016Herzog et al., , 2020)).
In this study, we tentatively refer to this psychophysical moment as "the moment of awareness" and investigate its relationship with the content of awareness, without refuting the theory that some phenomenal content of awareness evolves over time.The reportable content of awareness may be finalized when the content formation process reaches a certain stable state, for example, when the prediction error is judged to be minimized (e.g., Clark, 2013;Di Lollo et al., 2000).In other words, the content formation process is privileged to determine the moment of awareness.Alternatively, another timer-like process may trigger some conscious access to the content, regardless of the current state of the content formation process (e.g., Bachmann, 2021;VanRullen & Koch, 2003).In other words, the moment of awareness influences the content of awareness.Therefore, it is intriguing to ask whether the content of awareness can be altered by manipulating the moment of awareness.
To manipulate the moment of awareness, we utilized the "temporal facilitation" phenomenon: after a flanker is presented, a stimulus presented at a location near the flanker is perceived earlier than another stimulus presented elsewhere, even if their onsets are physically simultaneous (Scharlau & Neumann, 2003;Shore et al., 2001); the difference in the moment of awareness between the cued and uncued stimuli is usually estimated in the form of a shift of the point of subjective simultaneity for a few tens of milliseconds in a temporal-order or simultaneity-judgment task (for a review, see Spence & Parise, 2010).This phenomenon has been mainly attributed to the acceleration of perceptual processing by spatially pre-focused attention to the flanker location and thus referred to as "prior entry" or "perceptual latency priming" (Scharlau, 2007;Shore et al., 2001).However, because acceleration is also exogenously triggered by a single flanker presented at the fovea and by multiple flankers exceeding the capacity of simultaneous attention, it may be a multifaceted phenomenon that cannot be accounted for only by spatial attention (Bachmann, 1989;Schneider & Bavelier, 2003).Therefore, in this article, we use a more general term, "temporal facilitation," for this phenomenon.
To quantify the content of awareness, we focused on the visual illusion of orientation repulsion, wherein a vertical target is perceived as tilted against the orientation of a surrounding inducer (e.g., Gibson, 1937;Westheimer, 1990; for a review, see Clifford, 2014), because the extent of this illusion is known to change as a function of processing time (Kaneko et al., 2017;Nakamura & Murakami, 2021, 2022;Takao et al., 2020).We quantified the degree of repulsion when the moment of awareness of the target for repulsion was facilitated by a preceding flanker.If the awareness of the target arises when the content formation process (in this case, contextual modulation of orientation) has been completed, a preceding flanker would not alter the repulsion.In contrast, if the awareness arises regardless of the current state of the contextual modulation process, an incompletely modulated representation would be consciously accessed, possibly altering the repulsion.

Participants
Twenty naïve adults and one author (14 men and 7 women; 19-26 years of age) participated (throughout this study, the statistical significance did not change even if the data for the author were excluded).Before the experiment, the ethics committee of the Graduate School of Humanities and Sociology at the University of Tokyo approved this study, and we confirmed that all participants had normal or corrected-to-normal visual acuity and obtained written informed consent per the Declaration of Helsinki.Since three participants whose orientation discrimination performance did not reach the predetermined criterion (all JNDs < 5 • ) were excluded, data for the remaining 18 participants were reported here (the statistical significance of the results did not change even if the data for all participants were included).This sample size was determined by a simulation-based power analysis (Lakens & Caldwell, 2021) such that we could detect the interaction in a two-way repeated measures analysis of variance (rm-ANOVA) with a power of 0.9 and a significance level of 0.05.The expected effect size of the interaction was η 2 p = 0.42, which was calculated based on a previous study (Nakamura & Murakami, 2021).

Apparatus
The experiment was controlled using MATLAB R2019b and its extension package, Psychophysics Toolbox Version 3 (Brainard, 1997;Pelli, 1997).In a darkroom, visual stimuli were displayed on a gamma-corrected 22-in.CRT screen (Mitsubishi Electric RDF223H; resolution of 1600 × 1200 pixels; refresh rate of 60 Hz; mean luminance of 31 cd/m 2 ) via an AMD Radeon Pro WX 5100 graphic card installed on a computer (Dell Precision).The edge of the screen was obscured so that it would not serve as a reference frame.Participants viewed the stimuli with both eyes open from a viewing distance of 57 cm, constrained by a chinrest, and responded with a keyboard.

Stimuli
We used a method of single stimuli.Throughout each session, the participants were instructed to maintain their gaze on a central fixation point (a black bullseye subtending 0.8 × 0.8 deg of arc) and to report whether a target Gabor patch (spatial frequency of 2.0 cpd, Gaussian SD of 0.42 deg, Michaelson contrast of 99%, and orientation of ± 1, ±3, ±5, or ± 7 • relative to the vertical), which was T. Nakamura and I. Murakami presented 7.5 deg to the left or right of the fixation point, appeared tilted clockwise (CW) or counterclockwise (CCW) from the vertical.The inducer of orientation repulsion comprised eight Gabor patches (all with the same orientation, 20 • or − 20 • ) evenly spaced along a virtual circle (radius: 4.0 deg) concentric with each of the two possible target locations.To examine the effects of temporal facilitation, we also presented one or two flankers, each composed of five white dots (62 cd/m 2 ) evenly spaced along a virtual concentric circle (radius: 2.0 deg).In the "unilateral condition," only one flanker appeared at the side of the target; in the "bilateral condition," two flankers simultaneously appeared on both sides.

Procedure
In each trial, the inducer was presented for 633 ms, and at the midpoint of this time period, the target was flashed for 33 ms.In the "preceding condition," the flanker(s) was delivered 100 ms prior to the target onset and lasted for 33 ms; in the "baseline condition," the flanker(s) was presented simultaneously with the target (Fig. 1).The flanker was presented not only in the preceding condition but also in the baseline condition to equalize the roles played by the flanker in the psychophysical task (e.g., a spatial cue and a reference frame for reporting orientations).After the inducer disappeared, only the fixation point remained on the screen until the participants responded and an inter-trial interval of 1800 ms elapsed.Each participant completed 128 trials × 8 sessions × 2 days.All possible combinations of the conditions appeared at equal times in a random order within each session, and sufficient rest was provided between sessions.
If temporal facilitation influenced the appearance, the degree of repulsion would differ between the preceding and baseline conditions.Furthermore, if temporal facilitation depended on the focus of spatial attention, the change of repulsion would be stronger in the unilateral condition than in the bilateral condition.

Results
For each participant and condition, the percentage of trials in which the target appeared tilted CW was plotted against target orientation.Each data point was based on 32 observations.The data were fitted with a logistic function (the lapse rate was fixed at 0.02, and the slope was constrained between the CW-and CCW-inducer conditions) using Palamedes Toolbox (Kingdom & Prins, 2016).From the best-fit psychometric function, the target orientation corresponding to the 50% point of the ordinate was determined as the point of subjective equality (PSE) with the vertical.In addition, the just noticeable difference (JND) concerning the orientation was defined as half the distance between the target orientations corresponding to the 75% and 25% points of the ordinate.According to convention, the quantity of orientation repulsion was defined as half the difference between the two PSEs (the PSE for the CW-inducer condition minus the PSE for the CCW-inducer condition).
Repulsion was weaker in the preceding condition than in the baseline condition, F(1, 17) = 48.88,p <.001, η 2 p = 0.74, meaning that the preceding flankers reduced repulsion.Repulsion was also weaker in the bilateral condition than in the unilateral condition, F(1, 17) = 6.94, p =.017, η 2 p = 0.29, meaning that the addition of the flanker contralateral to the target, which was the only difference between the unilateral and bilateral conditions, reduced repulsion.Regarding the latter main effect, in hindsight, since the flanker provided information about the target location in half of the trials (i.e., the unilateral condition), the participants may have remained vigilant to the arrival of a flanker regardless of whether it was unilateral or bilateral (see Lin & Lu, 2016).In the presence of the contralateral flanker in addition to the ipsilateral one, perceptual load may have been higher and thus fewer resources might have been allocated to the task-irrelevant stimuli (inducers) in the bilateral condition (see Lavie & Tsal, 1994), possibly leading to weaker contextual modulation (see Freeman et al., 2001).In any case, since there was no significant interaction, F(1, 17) = 1.74, p =.205, η 2 p = 0.09, this effect of spatial configuration of the flanker was independent of temporal facilitation (the effect of the flanker timing).

Experiment 2
In Experiment 1, it was technically possible that the unilateral flanker in the preceding condition triggered both exogenous and endogenous types of spatial attention because the location of the impending target was 100% predictable.However, since endogenous attention takes roughly 300 ms to develop (Carrasco, 2011), it is unlikely that endogenous attention was fully activated to facilitate the processing of the target appearing 100 ms after the flanker.As a more stringent test of the involvement of endogenous spatial attention, in Experiment 2, we manipulated it independently of the flanker timing.The flankers were completely noninformative with respect to the target location throughout the experiment, whereas a central cue was provided to guide spatial attention.

Methods
The methods were identical to those used in Experiment 1, except for the following.Eighteen naïve adults and one author (13 men and 6 women; 20-28 years of age) participated.The gaze position of each participant was tracked monocularly with a video-based eye tracker (SR Research Eyelink 1000) at a sampling rate of 250 Hz.One participant did not actually perform any sessions due to calibration failures of eye tracking; thus, data for the remaining 18 participants were obtained.
Each trial started after fixation was maintained for 800 ms.Either one or both of two arrowhead-shaped cues (shaped like "<" and ">" each subtending 1.0 × 1.0 deg) was presented and remained on the screen until response (Fig. 3).In the "valid condition," either "<" or ">" was presented 0.88 deg to the left or right of the fixation point, respectively, and 100% predictive of the impending target's location (left or right, respectively).In the "neutral condition," both cues were simultaneously given; thus, the target location was unpredictable.The participants were informed of cue predictability in advance, and especially in the valid condition, they were  T. Nakamura and I. Murakami instructed to pay attention to the predicted target location without overt eye movement.The time course after the onset of the inducer, which was delivered 300 ms after the cue, was identical to that in Experiment 1, except that the flankers were bilaterally presented in all trials (completely noninformative regarding the target location) and that the inter-trial interval was 1500 ms.Each participant completed 128 trials × 8 sessions × 2 days.
Since directive arrow cues mainly guide endogenous attention (Posner, 1980;Green & Woldorff, 2012; but see Tipples, 2002), we had to check whether the valid cues actually guided attention toward the target location.To this end, the performance of an orientation discrimination task, namely reaction time (RT) and accuracy, was assessed for each participant in two additional sessions (128 trials each).The procedure was identical to that of the main sessions except for the following.First, only two orientations (7 • or − 7 • ) were used as the targets.Second, CW and CCW Gabor patches were circularly arranged and interleaved to cancel orientation repulsion while equating the spatial configuration with that used in the main sessions.Third, both speed and accuracy were explicitly emphasized during instruction.

Results
First, trials in which the gaze positions were deviated by 2.0 deg (either horizontally or vertically) or temporarily lost (e.g., due to a blink) during the period between the cue onset and target offset were excluded as fixation errors (approximately 0.5% and 1.4% for the neutral and valid conditions, respectively).The PSE and JND for each participant and condition were estimated in the same manner as in Experiment 1.We performed a 2 (preceding vs. baseline) × 2 (neutral vs. valid) rm-ANOVA for the orientation repulsion data.

Experiment 3
When two visual stimuli are presented in spatiotemporal proximity, the preceding one can quicken the awareness of the succeeding one (Bachmann, 1989;Scharlau, 2007).Although the preceding flanker in the above experiments may have caused similar temporal facilitation of the target, the occurrence of the facilitation needs validation in the units of time, which we did in Experiment 3 by utilizing a motion induction phenomenon (Hikosaka et al., 1993;von Grünau et al., 1996).In a typical motion induction paradigm, a flanker is initially flashed, and successively a static bar extending from the flashed position is presented; subjectively, the bar appears to expand quickly from the flashed position.This phenomenon is caused by temporal facilitation of local areas near the flanker but can occur simultaneously at multiple locations across the visual field, suggesting partial independence from spatial orienting (von Grünau et al., 1996).Instead of a bar, we presented three Gabor patches in a horizontal row to make only a minimal change in the stimulus configuration from that in the above experiments.One of the patches was located at the target location in the above experiments and accompanied by a flanker, and this patch was expected to appear first if it received the strongest temporal facilitation by the flanker.We manipulated the SOA among these Gabor patches and quantified temporal facilitation by a physical temporal offset necessary to cancel the illusory motion induced by the preceding flanker.As Hikosaka et al. (1993) posited, this motion induction paradigm yields less noisy data regarding temporal asynchrony than a temporal-order judgment because performance in the former can rely on the motion processing system inherently sensitive to spatiotemporal information, but performance in the latter is constrained by higherorder judgment processes having low temporal resolutions.Furthermore, for our purpose of quantifying the relative moment of awareness, the motion induction paradigm should be more suitable because the directional judgment is made based on the compelling perception of motion and thus less susceptible to response biases known to contaminate the temporal-order judgment (see Schneider & Bavelier, 2003: Shore et al., 2001).

Methods
The methods were identical to those used in Experiment 1, except for the following.Seventeen naïve adults and one author (8 men and 10 women; 19-27 years of age) participated.The resolution and refresh rate were set at 800 × 600 and 120 Hz, respectively.
At the locations where the inducers were presented in the above experiments, plaid patterns, each of which was made of the luminance average of a CW-and CCW-inducer, were circularly arranged; the two places that were nearest to the fixation point were left empty (Fig. 5a).Three vertical Gabor patches, each of which was identical to the target in the above experiments, were sequentially flashed (33 ms each) at 2.5, 5.0, and 7.5 deg eccentricities either in the inward or outward direction to the left or right of the fixation point.The SOA between the innermost and outermost targets, henceforth named "target SOA," were − 100, − 83, − 67, − 50, − 33, − 17, 0, 17, 33, 50, 67, 83, or 100 ms.The onset of the middle target was just at the midpoint between the innermost and outermost targets.Positive target SOAs denote that the innermost, middle, and outermost targets were presented in this order; thus, an outward motion would be perceived if illusory motion induction did not occur.When the target SOAs were negative, the presentation order of the targets was reversed (i.e., outermost, middle, and innermost).When the target SOA was zero, the three targets were simultaneously presented (Fig. 5b).Regardless of the target SOA, two bilateral flankers flashed either simultaneously with the outermost target ("baseline condition") or 100 ms prior to its onset ("preceding condition").The participants reported whether the perceived direction of motion was leftward or rightward.Each participant completed 104 trials × 6 sessions × 1 day.T. Nakamura and I. Murakami

Results
For each participant and condition, the percentage of trials in which outward motion was perceived was plotted against the target SOA.Each data point was based on 24 observations.The data were fitted with a logistic function (the lapse rate was fixed at 0.02) to determine the PSE, that is, the physical delay of the outermost target relative to the innermost one when the illusory motion was just canceled (Fig. 6a).Compared to the PSE in the baseline condition, which was approximately the point of no physical motion, M = 3.7 ms, 95%CI = [− 0.8, 8.3], the PSE in the preceding condition significantly shifted toward the positive direction, M = 42.4 ms, 95%CI = [33.7,51.1], t(17) = 13.03,p <.001, d z = 3.07.Therefore, temporal facilitation indeed quickened the awareness of the outermost target by approximately 40 ms relative to that of the innermost target (Fig. 6b).This relative time is comparable to those in typical temporal facilitation paradigms in previous studies (see Spence & Parise, 2010).

Experiment 4
In the preceding condition of the above experiments, the sequential presentation of the flanker and target may give an impression of apparent motion.Previous studies on the flash-lag effect have demonstrated that updating properties within an already instantiated object requires less time than creating a new object (Kanai et al., 2009;Moore and Enns, 2004).In addition, apparent motion might involve the motion induction phenomenon (e.g., the situation in Experiment 3; see Downing & Treisman, 1997).Therefore, object correspondence between the flanker and target could cause temporal facilitation.However, the presence of the preceding flanker per se might trigger some location-unspecific process of temporal facilitation.To examine this location-unspecific effect on repulsion, we located a new type of flanker maximally far apart from the target in the final experiment.

Methods
The methods were identical to those used in Experiment 1, except for the following.Nineteen naïve adults and one author (8 men and 12 women; 19-28 years of age) participated.Since two participants whose orientation discrimination performance did not reach the predetermined criterion (all JNDs < 5 • ) were excluded, data for the remaining 18 participants were reported here (the statistical significance of the results did not change even if the data for all participants were included).
Instead of surrounding dots, four white bullseyes (each subtending 0.8 × 0.8 deg) were presented at remote locations as a flanker (Fig. 7a).They were flashed for 33 ms at the four corners of a virtual rectangle (25 deg wide and 10 deg high) centered on the fixation point.The SOA between the flanker and target was either 100 ms ("preceding condition") or 0 ms ("baseline condition").Each participant completed 128 trials × 8 sessions × 1 day.Three participants, who exhibited a PSE out of the range of ± 8 • in at least one condition, were requested to perform additional sessions (128 trials × 8 sessions) on another day; responses to more deviated target orientations (±9, ±11, and ± 13 • ) were mainly collected so that the psychometric function re-estimated from these doubled data would straddle the tentative PSE, but those orientations used on the first day (±1, ±3, ±5, and ± 7 • ) were also presented in one-fourth of all trials to strengthen the randomness of orientation choices across trials.

Results
The PSE and JND for each participant and condition were estimated in the same manner as in Experiment 1. Consistently with Experiments 1 and 2, a preceding flanker significantly reduced repulsion, t(17) = 3.65, p =.002, d z = 0.86 (Fig. 7b, 7c) but did not significantly modulate JND, t(17) = 0.16, p =.874, d z = 0.04, suggesting that even location-unspecific temporal facilitation can lead to the reduction of repulsion.

Summary of results
We found that orientation repulsion was reduced when the awareness of a target for repulsion was quickened by preceding flankers.This temporal facilitation effect on the reduction of repulsion was caused even by flankers that provided no information about the location of the impending target (Experiments 1 and 2).In Experiment 3, we confirmed that the flanker we used actually quickened the awareness of the target by approximately 40 ms.Furthermore, in Experiment 4, we found that even a remote flanker, four bullseyes presented far away from the target, reduced repulsion when they preceded the target.

Psychological processes underpinning the temporal facilitation
In this study, orienting spatial attention improved the speeded judgment of orientation, as assessed by RT divided by the correct response rate (Experiment 2), but did not alter the appearance of orientation repulsion.Irrespective of spatial attention, the presence of a preceding flanker per se reduced repulsion (Experiments 1, 2, and 4).Since the preceding flanker quickened the awareness of the nearby target relative to the distant one (Experiment 3), location-specific but attention-independent processes, such as object correspondence between the flanker and target, are arguably involved to some extent.
However, the fact that even a remote flanker was able to reduce repulsion (Experiment 4) suggests that some location-unspecific process, possibly temporal attention triggered by the preceding flanker, can also quicken the awareness of the target.In parallel with spatial attention, temporal attention can be guided endogenously (Correa et al., 2004;Coull & Nobre, 1998) and exogenously (Duyar et al., 2023;Posner et al., 1973).Previous studies have demonstrated that both types of temporal attention indeed quicken the awareness, although these studies assessed temporal facilitation based on different paradigms (Hilkenmeier et al., 2012;Seifried et al., 2010).In our experiments, however, it is unlikely that endogenous attention was involved in the reduction of repulsion because endogenous attention requires longer time than the SOA (100 ms) between the flanker and target we used (Hackley et al., 2009, but see Yeshurun & Tkacz-Domb, 2021).In any case, to dissociate between these two types of attention more precisely, future studies should manipulate the temporal contingency of the flanker and target (Lawrence & Klein, 2013; see also McCormick et al., 2018).
Whether spatially pre-focused attention is necessary to cause temporal facilitation is controversial (Bachmann, 1989;Schneider & Bavelier, 2003), but spatial-attention-dependent and independent components of temporal facilitation have not been clearly dissociated because both yield virtually the same psychophysical effects such as the shift of the point of subjective simultaneity.Therefore, to our knowledge, our findings are the first to demonstrate the dissociation in terms of their differential relationship with the content of awareness, that is, the degree of orientation repulsion.

The moment of awareness influences the content of awareness
A reduction in orientation repulsion means that the target appears more faithful to the retinal input when our awareness of the target comes earlier than usual.According to a previous study, repulsion is reduced by backward masking only when the inducer is presented together with or after the target, and this temporal asymmetry has been explained by a quantitative model (Nakamura & Murakami, 2021).Our present findings accord with this previous computational model and suggest that the internal representation of the target orientation is initially faithful to the retinal input and evolves to become tilted against the inducer due to contextual modulation.Based on this idea, some timer-like processes, possibly temporal attention, may operate to convert the internal representation into explicit content for awareness even when the temporal evolution is incomplete.In other words, a perceptual content is forced to be finalized at the moment of awareness determined by another process that is insensitive to the current state of the content formation process.This idea resonates with the perceptual retouch theory (Bachmann, 1984(Bachmann, , 2021; see also Aru et al., 2020), which argues that, for the observer to become aware of a perceptual content, neural activities within cortical areas tuned for visual features must be sufficiently modulated by reticulo-thalamic content-nonspecific systems regulating the level of consciousness (e.g., Tononi & Edelman, 1998).Assuming that content-nonspecific modulation is delayed relative to content-specific activity by tens of milliseconds, this theory explains various spatiotemporal effects in psychophysics both qualitatively and quantitatively (Bachmann, 2021;Kirt & Bachmann, 2013).For example, backward masking occurs because the content-nonspecific modulation triggered by the target "retouches" the content-specific representation of the mask.Although the present study cannot specify the neural substrate for the reduced repulsion by temporal facilitation, the temporal interaction between distinct processes responsible for the moment and content of awareness, similar to this "perceptual retouch," may determine what we are aware of in orientation repulsion.In addition, the temporal facilitation by the preceding flanker that we quantified in Experiment 3 (approximately 40 ms) is consistent with the timescales of the evolution of orientation tunings in the macaque's V1 neurons (Ringach et al., 1997(Ringach et al., , 2003)).A recent psychophysical simulation incorporating these evolution timescales also well explains the duration dependence of the orientation tuning of illusory repulsion in the human (Takao et al., 2020).Therefore, we speculate that the preceding flanker promotes the conscious access to the internal representation of target orientation in the middle of the ongoing interactions among V1 neurons dynamically changing their orientation tunings.
Although the representational dynamics prior to awareness are often inferred with neurophysiological recording techniques (e.g., Alilović et al., 2021), our psychophysical approaches are advantageous.We propose that a direct link between the internal representation and the content of awareness can be elucidated only when we can take advantage of the observer's conscious access to a premature representation in a psychophysically traceable manner, as done in this study.

Processing time and subjective time
Due to the neural processing latency, the time of awareness of an event content is inevitably delayed relative to the time of retinal stimulation by the event.However, the moment of awareness within the neural time course is not equal to the moment of the event's occurrence in one's consciousness (Holcombe, 2014;Johnston & Nishida, 2001); in the laboratory, the subjective time quantified by a temporal-order or simultaneity-judgment task may be postdictively reconstructed after accumulating sensory evidence reaches the perceptual threshold (Amano et al., 2016).Since we were interested in the moment of awareness along the neural processing time, we did not directly examine the subjective time to avoid this epistemological problem regarding the disjunction between the processing time and subjective time.Instead, we quantified RT (Experiment 2), which is presumably linked to the time of straddling the perceptual threshold (Amano et al., 2016), and the motion induction phenomenon (Experiment 3), which presumably quantifies the relative time of signal entry into a neural process monitoring the spatiotemporal relationship (Hikosaka et al., 1993).In doing so, we found that both were facilitated by the preceding flanker.By contrast, the subjective moment of awareness of orientation repulsion in terms of our temporal phenomenology is an open question.

Conclusions
Orientation repulsion was reduced when the preceding flanker quickened the awareness of the target.This demonstrates that the content of repulsion is dynamically formed prior to awareness, and that the preceding flanker promotes our conscious access to a premature representation even if contextual modulation has not yet been completed.The present study captures the temporal relationship between the conscious and preconscious contents in the case of orientation repulsion, and can be a model case for psychophysical approaches to visual consciousness.

Funding
This work was supported by JSPS KAKENHI Grant Numbers 21J20400 and 22KJ0555 to TN, and 18H05523 and 23H01052 to IM.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.The time course of a single trial in Experiment 1.Only the bilateral condition is shown (in the unilateral condition, the flanker is presented only at the side of the target).The depictions here are framed in colors only for illustrative purpose.

Fig. 2 .
Fig. 2. Results of Experiment 1.(a) Inter-participant mean of orientation repulsion.Error bars indicate 95% within-participant confidence intervals (Morey, 2008).(b) Each individual's repulsion data for the preceding against baseline conditions.Markers above the diagonal line indicate that repulsion was reduced by temporal facilitation.

Fig. 3 .
Fig. 3. Snapshots of visual stimuli for the neutral (left) and valid (right) conditions in Experiment 2. Only the frames at which the target is presented in the baseline condition are shown here.

Fig. 4 .
Fig. 4. Results of Experiment 2. (a) Inter-participant mean of orientation repulsion in the main sessions.(b) Each individual's repulsion data for the preceding against baseline conditions.Inter-participant mean of (c) RT and (d) accuracy in the additional sessions are also shown.Error bars indicate 95% within-participant confidence intervals.

Fig. 5 .
Fig. 5. Procedure of Experiment 3. (a) Time course of a single trial.Only the condition in which the SOA between the outermost and innermost targets ("target SOA") is 67 ms is shown.(b) Space-time relationship between the flanker and three targets in the conditions wherein the target SOA is negative (left), zero (center) and positive (right).In these schematics, the targets are in the right hemifield and the x and t axes indicate eccentricity and time, respectively.

Fig. 6 .
Fig. 6. Results of Experiment 3. (a) Best-fit psychometric functions for a representative participant.Triangle markers indicate actual data points.(b) Inter-participant mean of PSE.Error bars indicate 95% within-participant confidence intervals.

Fig. 7 .
Fig. 7. Stimuli and results of Experiment 4. (a) A snapshot of visual stimuli for the baseline condition.Only the frame at which the target is presented is shown here.(b) Inter-participant mean of orientation repulsion.Error bars indicate 95% within-participant confidence intervals.(c) Each individual's repulsion data for the preceding against baseline conditions.