Perceiving a Continuous Visual World Across Voluntary Eye Blinks

Abstract

People blink their eyes every few seconds, but the changes in retinal illumination that accompany eyeblinks are hardly noticed. Furthermore, despite the loss of visual input, visual experience remains continuous across eyeblinks. Two hypotheses were investigated to account for these phenomena. The first proposes that perceptual information is maintained across a blink whereas the second proposes that perceptual information is not maintained but rather postblink perceptual experience is antedated to the beginning of the blink. Two experiments found no evidence for temporal antedating of a stimulus presented during a voluntary eyeblink. In a third experiment subjects judged the temporal duration of a stimulus that was interrupted by a voluntary eyeblink with that of a stimulus presented while the eyes were open. The duration of stimuli that were interrupted by eyeblinks was judged to be 117 ms shorter than that of stimuli presented while the eyes remained open, indicating that blink duration was not accounted for in the perception of stimulus duration. This suggests that perceptual experience is neither maintained nor antedated across eyeblinks, but rather is ignored, perhaps in response to the extraretinal signal that accompanies the eyeblink.

Full text

Perceiving a Continuous Visual World Across Voluntary Eye Blinks
David E. Irwin and Maria M. Robinson
Department of Psychology
University of Illinois
Running Head: Continuity Across Voluntary Eye Blinks
Corresponding Author:
David E. Irwin
Department of Psychology
University of Illinois
603 E. Daniel St.
Champaign IL, 61820
Phone: (217) 333-7746
FAX: (217) 244-5876
EMAIL: irwin@illinois.edu
In press, Journal of Experimental Psychology: Human Perception and Performance

Continuity Across Eye Blinks
!
2!
Abstract
People blink their eyes every few seconds but the changes in retinal illumination
that accompany eye blinks are hardly noticed. Furthermore, despite the loss of visual
input, visual experience remains continuous across eye blinks. Two hypotheses were
investigated to account for these phenomena. The first proposes that perceptual
information is maintained across a blink, while the second proposes that perceptual
information is not maintained but rather post-blink perceptual experience is antedated to
the beginning of the blink. Two experiments found no evidence for temporal antedating
of a stimulus presented during a voluntary eye blink. In a third experiment subjects
judged the temporal duration of a stimulus that was interrupted by a voluntary eye blink
with that of a stimulus presented while the eyes were open. The duration of stimuli that
were interrupted by eye blinks was judged to be 117 ms shorter than that of stimuli
presented while the eyes remained open, indicating that blink duration was not accounted
for in the perception of stimulus duration. This suggests that perceptual experience is
neither maintained nor antedated across eye blinks, but rather is ignored, perhaps in
response to the extraretinal signal that accompanies the eye blink.
Keywords: eye blinks; temporal antedating; perceptual memory

Continuity Across Eye Blinks
!
3!
Humans typically blink their eyes 12 – 15 times each minute, sometimes
reflexively in response to external stimulation, sometimes voluntarily in response to a
command, and most often spontaneously in the absence of any obvious evoking stimulus
(Stern, Walrath, & Goldstein, 1984). Although the kinematics of these different kinds of
eye blinks are somewhat different, in all cases vision is almost completely blocked by the
closed eyelids for approximately 100–150 ms (Riggs, Volkmann, & Moore, 1981).
Despite their frequency, magnitude, and duration, people rarely notice these blank
periods, even though dimming the lights in a room for the same duration is very
noticeable (Volkmann, Riggs, & Moore, 1980). Furthermore, despite the loss of visual
input, visual experience remains continuous across eye blinks. How does this occur?
Volkmann et al. (1980) demonstrated that the blank periods that occur during an
eye blink are not perceived because a central inhibitory signal suppresses vision and
thereby minimizes perception of the blackout. In their study a fiber-optic bundle was
placed against the roof of the mouth to present light through the back of the eyeball to the
retina. Participants wore opaque goggles to ensure that visual stimulation arrived only
through the back of the eye. Volkmann et al. found that visual sensitivity for brief
decrements in this visual stimulus was reduced by about 0.5 log units during a blink and,
to a lesser extent, 100 ms before and up to 200 ms after a blink as well, even though the
light source itself was never physically impeded. This demonstrated that visual
suppression during eye blinks is due, at least in part, to a central inhibitory mechanism.
Subsequent experiments showed that visual suppression occurs for reflexive blinks as
well as for voluntary blinks (Manning, Riggs, & Komenda, 1983). Possible neural loci
for this central inhibition include V1 (Gawne & Martin, 2000) and lateral occipital

Continuity Across Eye Blinks
!
4!
cortex, particularly area V5/MT (Bristow, Frith, & Rees, 2005), where neural activation
is reduced during an eye blink.
While visual suppression helps explain why blank periods during blinks are not
perceived, it is less clear why visual continuity is perceived across eye blinks. In the
present paper we examined two possible hypotheses for this perceptual experience. The
first is that perceptual experience is maintained in memory across the eye blink. For
example, Bristow, Frith, and Rees (2005) proposed that a short-term mnemonic signal
associated with the blink motor command maintains the pre-blink percept during the
blank period that occurs during an eye blink. Based on an fMRI study they proposed that
a region in medial parieto-occipital cortex (the human equivalent of the macaque
V6/V6A complex, area PO) might subserve this purpose, because neural activity in the
presence (vs. absence) of a blink was greater when a visual stimulus was present
compared to when it was absent.
The second hypothesis we examined is based on one that has been proposed to
explain the perception of visual continuity across saccadic eye movements. It suggests
that perceptual experience is not maintained across a saccade or an eye blink, but rather
that perceptual experience is antedated to the beginning of these eye movement events.
Evidence for this kind of temporal antedating has been found for saccadic eye
movements via the phenomenon of saccadic chronostasis, which refers to the fact that
people consistently overestimate the duration of a stimulus presented during a saccade in
comparison with the same stimulus seen at fixation (e.g., Yarrow, Haggard, Heal, Brown,
& Rothwell, 2001; Yarrow, Johnson, Haggard, & Rothwell, 2004; Yarrow, Whiteley,
Haggard, & Rothwell, 2006). In a typical experiment of this kind, subjects saccade to a

Continuity Across Eye Blinks
!
5!
target that changes during the saccade, then they judge whether the new target stimulus
was presented for a longer or shorter time than a subsequently presented reference
stimulus. The point of subjective equality (PSE) for the target stimulus is found and
compared to the same task performed at fixation. The PSE is consistently shorter in the
saccade condition than in the fixation condition, implying that people overestimate the
duration of the post-saccadic stimulus. The overestimation is directly related to the
duration of the saccade (i.e., it is longer for long saccades than for short saccades), which
suggests that the perceived onset of the saccade target is effectively antedated to a
moment just prior to saccade onset. The interpretation of Yarrow and colleagues is that
the brain simply assumes that the post-saccadic stimulus was present when the saccade
was initiated, yielding a perception of continuity across the saccade. Given that temporal
antedating occurs across saccades, it seems possible that it might occur for eye blinks as
well. This would be consistent with the results of Hari, Salmelin, Tissari, Kajola, and
Virsu (1994), who found that blinks influenced activation in posterior parietal cortex
shortly after blink offset; they suggested that this activation was essential for stable visual
perception across eye blinks, as though perceptual experience were being antedated to the
beginning of the blink (see also, Bodis-Wollner, Bucher, & Seelos, 1999).
We investigated the perceptual maintenance and temporal antedating hypotheses
in three experiments in which subjects judged the temporal duration of a stimulus that
was interrupted by a voluntary eye blink with that of a stimulus presented while the eyes
were open. Of interest was whether blink duration would be taken into account when
judging the temporal duration of a stimulus, as predicted by these hypotheses. It is
important to note that perceptual continuity need not necessarily rely on “filling in” the

Continuity Across Eye Blinks
!
6!
blank period during an eye blink, however. That is, perceived continuity across eye
blinks might occur as a result of the perceptual system simply ignoring the visual
interruption caused by the eye blink. Finding that blink duration is not taken into account
when judging the duration of a stimulus would support this hypothesis.
Experiment 1
Experiment 1 used a variant of the procedure that was used by Yarrow, Whiteley,
Haggard, and Rothwell (2006) in their studies of temporal antedating during saccades to
investigate whether temporal antedating occurs for voluntary eye blinks as well.
Method
Participants. Sixteen students from the University of Illinois community
participated in Experiment 1. The number of participants was based on Yarrow et al.
(2006), who used 18 participants to find significant temporal antedating effects of 30 ms
during saccades; the temporal antedating hypothesis predicts effects greater than 100 ms
in the case of blinks (because the duration of blinks is much longer than that of saccades)
so we assumed that 16 participants would be sufficient to find an effect (this was
confirmed by power analyses reported below). All participants reported normal or
corrected to normal vision and were naïve as to the purpose of the experiment. Each
received payment for participating in a single 50-minute session. The research was
conducted in accord with APA standards for the ethical treatment of subjects and with the
approval of the University of Illinois Institutional Review Board.
Apparatus. The stimuli were presented on a 21-inch CRT monitor (ViewSonic
G810) with resolution of 800 x 600 pixels and a refresh rate of 85 Hz. Eye movements
were recorded with an Eyelink II video-based eyetracker (SR Research Ltd., Mississauga,

Continuity Across Eye Blinks
!
7!
Ontario, Canada) with temporal resolution of 500 Hz, spatial resolution of 0.1º, and
pupil-size resolution of 0.1% of pupil diameter. The output of the eyetracker was
analyzed online to detect eye blinks. Each data sample from the eyetracker contained a
timestamp in milliseconds, the velocity and the position of the eye, and the area of the
pupil. An eye blink was defined as a period of missing pupil for at least 6 consecutive
ms. Blink onset and blink offset were defined to correspond to the beginning and ending
of the period of missing pupil. Custom C code was written to display stimuli and collect
responses. The participants’ heads were stabilized with a chin-rest, fixed at 49 cm from
the computer monitor. The height of the chair that participants sat in was adjusted for
each individual so that their eyes were centered with respect to the display monitor. The
display background was light gray (luminance = 86.3 cd/m2). Participants made manual
responses by pressing buttons on a Microsoft Sidewinder digital game controller
interfaced with the eyetracking computer.
Procedure
Each participant completed 12 blocks of trials, 6 blocks during which they
blinked and 6 blocks during which they did not blink. The blocks alternated between no-
blink blocks and blink blocks. Odd numbered subjects started with a blink block while
even numbered subjects started with a no-blink block. An instruction appeared on the
display before each block to remind the subject whether it was a blink or no-blink block.
Each block of trials began with a 5-position calibration procedure in which the
edges and center of the screen were fixated. Participants began each trial by pressing a
button on the game controller while fixating a drift correction dot that subtended 0.6º of
visual angle (see Figure 1). After the drift correction dot disappeared, a blank white

Continuity Across Eye Blinks
!
8!
screen was presented for 600 ms; during the last 130 ms of this presentation a tone
sounded as well. What happened next depended on the block type. During no-blink
blocks a 250 ms delay ensued and then a blue letter resembling a block A (the target
stimulus) was presented in the center of the screen for a variable period of time and then
erased. The blue letter had chromaticity of x = .241, y = .186 and luminance = 25.1
cd/m2, as measured by a Minolta CS-100 Chroma Meter (Minolta Camera Company,
Japan). Following a 1000 ms blank screen a second blue block A (the comparison
stimulus) was presented for a constant duration of 600 ms. In contrast, during voluntary
blink blocks subjects were instructed to blink shortly after they heard the tone, and when
the blink was detected, the blue block A target stimulus was presented while the eyes
were closed, so that it was visible on the screen when the eyes reopened. The target
stimulus was presented as soon as blink onset was detected (i.e., when the computer
program detected that the pupil was missing for at least 6 consecutive ms). The target
stimulus was presented for a variable period of time and then erased, and then following a
1000 ms blank screen the blue block A comparison stimulus was presented for a constant
duration of 600 ms. In both conditions subjects were instructed to report whether the
letter they saw first or the letter they saw second was seen for a longer period of time, and
they indicated their response by pressing the left (if the first letter seemed to have a
longer duration) or right (if the second letter seemed to have a longer duration) trigger on
the game controller. No feedback was given. The duration of the target stimulus was
then adjusted on the next trial based on this response, as determined by the modified
binary search (MOBS) procedure (low boundary 200 ms, high boundary 1600 ms, initial
presentation time 900 ms, five reversals to terminate), eventually reaching a value that

Continuity Across Eye Blinks
!
9!
was subjectively equal to the fixed duration (600 ms) of the comparison stimulus (Tyrrell
& Owens, 1988). The MOBS procedure yields efficient threshold estimates by
combining binary search and bracketing techniques; Monte Carlo simulations show that it
provides more precise measures with fewer stimulus presentations than conventional
staircase techniques (see Tyrrell & Owens, 1988, for further details). Blocks finished
when the MOBS criteria were satisfied, which typically took 6-25 trials. Six estimates of
the subjective duration of the target stimulus were collected per condition (blink vs. no-
blink) for each subject, one per block.
Results
Mean blink latency (i.e., when the eyelids started to move) from tone offset was
403 ms (sd = 248 ms). Mean blink duration (i.e., total eyelid movement time) was 251
ms (sd = 121 ms). On average the pupil was covered beginning 432 ms (sd = 249 ms)
after tone offset and it remained covered for an average of 130 ms (sd = 73 ms).
Recall that the target stimulus was presented as soon as the pupil was covered but
it did not become visible to the subject until after the pupil became uncovered an average
of 130 ms later. To determine whether perception of the target stimulus was temporally
antedated to the beginning of the blink, for each subject a mean subjective duration
estimate (PSE) for the target stimulus was calculated by taking the average of the MOBS
termination values for the six blocks in each condition. In the blink condition this value
was corrected post hoc to correct for the amount of time that the stimulus had been
presented while the eyes were closed (i.e., pupil was covered), similar to the experiments
of Yarrow and colleagues. The temporal antedating hypothesis predicts that the
subjective duration estimate (PSE) of the target stimulus should be shorter under blink

Continuity Across Eye Blinks
!
10!
than under no-blink conditions because under blink conditions participants antedate their
perception of the target stimulus (which is visible only after the blink has ended) to the
beginning of the blink. If antedating is done to the time of blink initiation the PSE for
voluntary blinks should be 251 ms shorter than its corresponding no-blink control (since
this was the duration of the voluntary blink), whereas it should be 130 ms shorter if it is
merely antedated to the time that the pupil is first covered by the eyelids. A post-hoc
power analysis using the G*Power analysis program (Faul, Erdfelder, Lang, & Buchner,
2007) indicated that we had power of 0.99 to detect an effect of 130 ms.
Contrary to the predictions of the temporal antedating hypothesis, the PSE for
voluntary blink trials (mean = 676 ms, se = 39 ms) was longer (not shorter) than the PSE
for no-blink trials (mean = 650 ms, se = 20 ms) but the difference was not significant,
t(15) = 0.89, sd = 120, p > 0.38. The effect size (d) was 0.19, based on the procedure
described by Dunlap, Cortina, Vaslow, and Burke (1996) for paired-sample t-tests. The
scaled (r = 1) JZS Bayes Factor in support of the null hypothesis was 3.65 (Rouder,
Speckman, Sun, Morey, & Iverson, 2009).
Discussion
If people antedate their perception of a stimulus that is presented during a blink to
the beginning of the blink, then the duration of such a stimulus should be overestimated
by 130 - 250 ms relative to the duration of a stimulus that is presented while the eyes
remain open. We found instead only a small and non-significant difference (in the wrong
direction) in the perceived duration of a stimulus presented during a blink compared to a
no-blink control. These results are inconsistent with the hypothesis that temporal

Continuity Across Eye Blinks
!
11!
antedating may occur for stimuli presented during eye blinks, even though such
antedating has been found for saccadic eye movements.
One perhaps non-optimal aspect of the procedure that we used in Experiment 1
was that the target stimulus was presented as soon as pupil occlusion was detected; as a
consequence, the stimulus was unseen for the period of time that the pupil was covered
and this time then had to be subtracted from the perceived duration estimate calculated by
the MOBS algorithm. It seemed possible that this might yield an inaccurate estimate of
the apparent duration of the target stimulus, so in Experiment 2 we replicated the
procedure of Experiment 1 with the exception that the target stimulus was presented as
soon as the pupil became uncovered during the eye blink (i.e., at the end of the pupil
occlusion period rather than at the beginning). Of interest was whether the duration of
the stimulus would be antedated to the beginning of the blink, as predicted by the
temporal antedating hypothesis.
Experiment 2
Method
Participants. Sixteen students from the University of Illinois community
participated in Experiment 2. All participants reported normal or corrected to normal
vision and were naïve as to the purpose of the experiment. Each received payment for
participating in a single 50-minute session. None had participated in the first experiment.
Apparatus and Procedure. The apparatus and the procedure were the same as in
Experiment 1, except that the target stimulus was presented when the pupil became
uncovered at the end of the blink (more specifically, given our refresh rate of 85 Hz,
within 12 ms from detection of blink offset) instead of at the beginning of pupil

Continuity Across Eye Blinks
!
12!
occlusion. Because of visual blink suppression, perceptually it appeared as though the
stimulus had been presented while the eyes were still closed.
Results
Mean blink latency from tone offset was 522 ms (sd = 462 ms). Mean blink
duration was 255 ms (sd = 146 ms). On average the pupil was covered beginning 537 ms
(sd = 435 ms) after tone offset and it remained covered for an average of 172 ms (sd =
140 ms).
To determine whether perception of the target stimulus was temporally antedated
to the beginning of the blink, for each subject a mean subjective duration estimate (PSE)
for the target stimulus was calculated by taking the average of the MOBS termination
values for the six blocks in each condition. Recall that the temporal antedating hypothesis
predicts that the subjective duration estimate (PSE) of the target stimulus should be
shorter under blink than under no-blink conditions because under blink conditions
participants antedate their perception of the target stimulus to the beginning of the blink.
If antedating is done to the time of blink initiation (i.e., when the eyelids start to move)
the PSE for blinks should be 255 ms shorter than in the no-blink control condition (since
this was the duration of the blink), whereas it should be 172 ms shorter if it is merely
antedated to the time that the pupil is first covered by the eyelids (our power to detect a
172 ms effect was greater than 0.99). In fact the PSE for blink trials (mean = 651 ms, se =
49 ms) was longer than the PSE for no-blink trials (mean = 602 ms, se = 44 ms) but the
difference was not significant, t(15) = 1.31, sd = 149, p > 0.21, d = 0.26. The scaled (r =
1) JZS Bayes Factor in support of the null hypothesis was 2.43.

Continuity Across Eye Blinks
!
13!
Omnibus analysis. Because Experiment 2 was a replication of Experiment 1
(except for a minor change in procedure), we conducted an omnibus analysis combining
the data from the two experiments. The PSE for blink trials (mean = 664 ms, se = 31 ms)
was longer than the PSE for no-blink trials (mean = 626 ms, se = 24 ms) but the
difference was not significant, t(31) = 1.6, sd = 134, p = 0.12, d = 0.24. The scaled (r =
1) JZS Bayes Factor in support of the null hypothesis was 2.20. Note that the difference
was also in the direction opposite to that predicted by the temporal antedating hypothesis.
Discussion
The results of Experiment 2 replicated those of Experiment 1. The temporal
antedating hypothesis predicts that the duration of a stimulus presented during a
voluntary eye blink should be overestimated by 172 - 250 ms relative to the duration of a
stimulus that is presented while the eyes remain open. We found instead only a small and
non-significant difference in the perceived duration of a stimulus presented during a blink
compared to a no-blink control. These results are inconsistent with the hypothesis that
the perception of a stimulus presented during an eye blink is antedated to the beginning
of the eye blink, even though such antedating has been found for saccadic eye
movements.
One weakness of the first two experiments is that the conclusion that temporal
antedating does not occur for eye blinks relies on accepting the null hypothesis. So, in
Experiment 3 we examined the temporal antedating hypothesis in a new way, by having
subjects blink while a stimulus was already present on the display (as opposed to
presenting the stimulus during the blink) and examining whether blink duration was taken
into account when evaluating the duration of the stimulus. This would be expected under

Continuity Across Eye Blinks
!
14!
the temporal antedating hypothesis, and it is also predicted by the alternative hypothesis
that perceptual information is maintained across an eye blink to produce uninterrupted
visual experience (e.g., Bristow, Frith, & Rees, 2005).
Experiment 3
In Experiment 3 the MOBS procedure was used to calculate the point of
subjective equality in stimulus duration between a constant-duration stimulus whose
viewing was interrupted by an eye blink (in the blink condition) or not interrupted by an
eye blink (in the no-blink condition) and a comparison stimulus that varied in duration.
A letter resembling a blue block A was presented for 1000 ms and participants either
blinked or did not blink during its presentation. They then compared their perception of
how long the letter had been presented with a comparison stimulus that varied in
duration. Both the temporal antedating and perceptual maintenance hypotheses predict
that the points of subjective equality should be equal under blink and no-blink conditions,
either because subjects antedate their perceptual experience of the stimulus to the
beginning of the blink (in the case of the temporal antedating hypothesis) or because
stimulus information is maintained in memory during the blink (in the case of the
perceptual maintenance hypothesis).
Method
Participants. Sixteen students from the University of Illinois community
participated in Experiment 3. All participants reported normal or corrected to normal
vision and were naïve as to the purpose of the experiment. Each received payment for
participating in a single 50-minute session. None had participated in either of the first
two experiments.

Continuity Across Eye Blinks
!
15!
Apparatus and Procedure. The apparatus was the same as in Experiments 1 and
2. Each participant completed 6 blocks during which they blinked and 6 blocks during
which they did not blink. Order (blink condition first or no-blink condition first) was
counterbalanced across subjects. An instruction appeared on the display before each
block to remind the subject whether it was a blink or no-blink block.
Each block of trials began with a 5-position calibration procedure in which the
edges and center of the screen were fixated. Participants began each trial by pressing a
button on the game controller while fixating a drift correction dot that subtended 0.6º of
visual angle (see Figure 2). After the drift correction dot disappeared, a blank white
screen was presented for 529 ms. Then, in both blink and no-blink conditions, a blue
letter resembling a block A (the target stimulus) was presented in the center of the screen
for 1000 ms. During no-blink blocks subjects were instructed to keep their eyes open.
On voluntary blink blocks subjects were instructed to blink as soon as the initial letter
was presented. Following the 1000 ms presentation of the target stimulus, in both
conditions a 529 ms blank screen was then presented, followed by the presentation of a
second blue block A (the comparison stimulus) whose duration was determined by the
modified binary search (MOBS) procedure (low boundary 400 ms, high boundary 1600
ms, initial presentation time chosen randomly between 600 and 1400 ms, five reversals to
terminate), eventually reaching a value that was subjectively equal to the fixed duration
(1000 ms) of the target stimulus. Six estimates of the subjective duration of the target
stimulus were collected per condition (blink vs. no-blink) for each subject, one per block.
In both conditions subjects were instructed to report whether the letter they saw
first or the letter they saw second was seen for a longer period of time, and they indicated

Continuity Across Eye Blinks
!
16!
their response by pressing the left (if the first letter seemed to have a longer duration) or
right (if the second letter seemed to have a longer duration) trigger on the game
controller. No feedback was given.
Results
Mean blink latency from target stimulus onset was 433 ms (sd = 228 ms). Mean
blink duration was 219 ms (sd = 50 ms). On average the pupil was covered beginning
469 ms (sd = 227 ms) after target onset and it remained covered for an average of 114
ms (sd = 50 ms). Thus, the target stimulus was visible for an average of 417 ms after
blink offset (as measured from the end of pupil occlusion).
As in the first two experiments, for each subject a mean subjective duration
estimate (PSE) for the target stimulus was calculated by taking the average of the MOBS
termination values for the six blocks in each condition (blink vs. no-blink). Both the
temporal antedating hypothesis and the perceptual maintenance hypothesis predict that
the subjective duration estimate (PSE) of the target stimulus should be equal under blink
and no-blink conditions because under blink conditions participants take the duration of
the blink into account, either by antedating their perception of the target stimulus to the
beginning of the blink or by maintaining a memory representation of the stimulus during
the blink. Instead we found that the PSE for blink trials (mean = 851 ms, se = 34 ms) was
significantly shorter than the PSE for no-blink trials (mean = 968 ms, se = 35 ms), t(15) =
3.52, sd = 132, p < .005, d = 0.88, power > .90. The scaled (r = 1) JZS Bayes Factor in
support of the alternative hypothesis (i.e., blink mean is different from no-blink mean)
was 14.2. Note that the PSE under blink conditions was 117 ms shorter than under no-
blink conditions, which is almost identical to the duration that the pupil was covered

Continuity Across Eye Blinks
!
17!
during voluntary blinks (114 ms). This provides strong support for the hypothesis that
the time that the stimulus was occluded by the closed eyelids was ignored in subjects’
estimation of how long the stimulus was presented.
Discussion
In this experiment the MOBS procedure was used to calculate the point of
subjective equality in stimulus duration between a constant-duration stimulus whose
viewing was interrupted by an eye blink or not and a comparison stimulus that varied in
duration. The temporal antedating hypothesis predicts that the points of subjective
equality should be equal under blink and no-blink conditions because subjects antedate
their perceptual experience of the stimulus to the beginning of the blink, thereby filling in
the interval during the blink. The perceptual maintenance hypothesis makes the same
prediction because it holds that stimulus information is maintained in memory during the
blink. The results of Experiment 3 were inconsistent with both hypotheses; instead, it
appears that participants did not take blink duration into account when assessing the
duration of the stimulus. Participants judged a stimulus that was interrupted by a blink as
being 117 ms shorter than it actually was, relative to a no-blink condition. The duration
of the target stimulus under no-blink conditions was also underestimated to some extent,
due to the well-known time-order error in stimulus comparison (e.g., Needham, 1934;
Woodrow, 1935), which is probably due to a fading mental representation of the first
stimulus during the interstimulus interval prior to its comparison with the comparison
stimulus (e.g., Schab & Crowder, 1988).
Although subjects underestimated the duration of the stimulus interrupted by an
eye blink, it appears that they did perceive the stimulus as being continuous. This is

Continuity Across Eye Blinks
!
18!
shown by the finding that the perceived duration of the stimulus was considerably longer
than its post-blink presentation time. The mean post-blink duration of the stimulus
interrupted by a voluntary blink was 417 ms, but the perceived duration of this stimulus
was 851 ms. Thus, it appears that stimulus continuity was perceived even though the
duration of the blink itself was ignored in subjects’ estimation of its duration.
General Discussion
We investigated two hypotheses that have been proposed to explain why vision
appears continuous across the temporal interruptions caused by eye blinks. The first, the
temporal antedating hypothesis, claims that people antedate their perception of a stimulus
presented during a blink to the time of blink onset, producing the perception of a
continuously present stimulus. Two experiments used a variant of procedure used by
Yarrow and colleagues (e.g., Yarrow, Haggard, Heal, Brown, & Rothwell, 2001; Yarrow,
Johnson, Haggard, & Rothwell, 2004; Yarrow, Whiteley, Haggard, & Rothwell, 2006) to
study temporal antedating across saccades to investigate whether temporal antedating
occurs across eye blinks as well. Participants judged the duration of a stimulus that was
presented at the beginning (Experiment 1) or end (Experiment 2) of an eye blink against
that of a constant duration stimulus presented during fixation. If people antedate their
perception of a stimulus that is presented during a blink to the beginning of the blink,
then the duration of such a stimulus should be overestimated by approximately 130 – 250
ms (i.e., the duration of the blink) relative to the duration of a stimulus that is presented
while the eyes remain open. We found instead only small and non-significant differences
in the perceived duration of a stimulus presented during a voluntary eye blink compared
to a no-blink control. These results are inconsistent with the hypothesis that temporal

Continuity Across Eye Blinks
!
19!
antedating occurs for stimuli presented during eye blinks, even though Yarrow and
colleagues found that such antedating occurs for saccadic eye movements.
A third experiment examined the temporal antedating hypothesis in a different
way, by having subjects blink while a stimulus was already present on the display (as
opposed to presenting the stimulus during the blink) and examining whether blink
duration was taken into account when evaluating the duration of the stimulus. This
would be expected under the temporal antedating hypothesis, and it is also predicted by
the second hypothesis that we considered, the perceptual maintenance hypothesis, which
posits that perceptual information is maintained across an eye blink to produce
uninterrupted visual experience (e.g., Bristow, Frith, & Rees, 2005). This experiment
calculated the point of subjective equality in stimulus duration between a constant-
duration stimulus whose viewing was interrupted by an eye blink or not and a comparison
stimulus that varied in duration. The results of this experiment were inconsistent with
both the temporal antedating and perceptual maintenance hypotheses; rather, participants
judged a stimulus that was interrupted by a blink as being 117 ms shorter than it actually
was, relative to a no-blink condition. The average blink duration was 114 ms; thus,
instead of stimulus information being antedated or maintained in memory across an eye
blink, it appears that it is simply ignored instead. Despite this, it appears that stimulus
continuity was preserved, because participants perceived the duration of the blink-
interrupted stimulus as being considerably longer than its post-blink presentation time.
Similar results using a somewhat different procedure were reported in abstract form by
Duyck, Collins, and Wexler (2015).

Continuity Across Eye Blinks
!
20!
In all of our experiments we assessed the subjective duration of visual experience
across an eye blink against a no-blink control in which the visual stimulus was
uninterrupted. Because visual experience across an eye blink appears to be continuous, it
seemed to us that using a continuous visual stimulus was the appropriate control
condition to use. One could conceivably explore other conditions, however, such as
mimicking the retinal effects of an eye blink by blanking the display for durations
equivalent to those of eye blinks. Although this may mimic the retinal effect of an eye
blink in some ways, perceptually such interruptions are very salient and lead to a stimulus
appearing to be discontinuous rather than continuous. As noted earlier, Volkmann,
Riggs, and Moore (1980) showed that people rarely notice the blank periods produced by
an eye blink, but dimming the lights in a room for the same duration is very noticeable.
Because we were interested in investigating why the visual world appears continuous (as
opposed to discontinuous) across eye blinks, a continuous visual control condition rather
than a discontinuous one seemed appropriate to us. Furthermore, others have shown that
visual stimuli that are discontinuous in time are actually perceived to be longer than
visual stimuli that are continuous, which is the opposite of what we found (e.g., Kanai,
Paffen, Hogendoorn, & Verstraten, 2006; Yuasa & Yotsumoto, 2015). Thus, we are
confident that our results are due to stimulus duration being ignored during eye blinks as
opposed to being caused by retinal differences between the blink and no-blink control
conditions.
Having ruled out the temporal antedating and perceptual maintenance hypotheses,
the question still remains: Why does the visual world appear continuous across eye
blinks? It is important to note that although the temporal antedating and perceptual

Continuity Across Eye Blinks
!
21!
maintenance hypotheses attempt to explain perceptual continuity via mechanisms that
“fill in” the blank period during the blink (so that the perceived duration of a stimulus
interrupted by an eye blink reflects its actual physical duration), perceived continuity
might instead result from the perceptual system simply ignoring the visual interruption
caused by the eye blink. For example, the motor signal (efferent signal) that accompanies
an eye blink may signal to the perceptual system that any disruption in phenomenal
experience that accompanies an eye blink is to be ignored because it is caused by the
blink and not by a change in the external world (Deubel, Bridgeman, & Schneider, 2004).
In support of this hypothesis, Deubel, Bridgeman, and Schneider (2004) found that a
blink operated differently than a blank interval when it comes to the detection of stimulus
displacements across saccades. Detecting that a stimulus has been displaced is difficult if
the displacement occurs during a saccade (such that the displaced stimulus is visible
immediately after the saccade ends), but Deubel and Schneider (1994), Deubel,
Schneider, and Bridgeman (1996), and Deubel, Bridgeman, and Schneider (1998) found
that displacement detection improved dramatically if a blank interval separated saccade
offset and the presentation of the displaced stimulus. Deubel, Bridgeman, and Schneider
(2004) found that replacing the blank interval with an eye blink of similar duration had
no beneficial effect, however; subjects were no better at detecting stimulus displacements
when vision was interrupted by a blink than they were when no blank interval separated
saccade offset and stimulus onset. Deubel et al. (2004) proposed that the extraretinal
signal that accompanies an eye blink allows the perceptual system to distinguish between
internally generated and externally generated sources of temporary object disappearances,
so that blinks are interpreted differently than blanks. In other words, Deubel et al. (2004)

Continuity Across Eye Blinks
!
22!
found that blink duration was ignored by the perceptual system, just as we found in our
third experiment.
In conclusion, visual experience remains continuous across eye blinks despite the
loss of visual input that they produce. Our results show that perceptual experience is
neither maintained nor antedated across eye blinks, but rather is ignored, presumably in
response to the extraretinal signal that accompanies the eye blink.

Continuity Across Eye Blinks
!
23!
References
Bodis-Wollner, I., Bucher, S. F., & Seelos, K. C. (1999). Cortical activation patterns
during voluntary blinks and voluntary saccades. Neurology, 53, 1800-1805.
Bristow, D., Frith, C., & Rees, G. (2005). Two distinct neural effects of blinking on
human visual processing. NeuroImage, 27, 136-145.
Deubel, H., Bridgeman, & Schneider, W. X. (1998). Immediate postsaccadic
information mediates space constancy. Vision Research, 38, 3147-3159.
Deubel, H., Bridgeman, & Schneider, W. X. (2004). Different effects of eyelid blinks
and target blanking on saccadic suppression of displacement. Perception &
Psychophysics, 66, 772-778.
Deubel, H., & Schneider, W. X. (1994). Perceptual stability and postsaccadic visual
information: Can man bridge a gap? Behavioral and Brain Sciences, 17, 259-
260.
Deubel, H., Schneider, W. X., & Bridgeman, B. (1996). Postsaccadic target blanking
prevents saccadic suppression of image displacement. Vision Research, 36, 985-
996.
Dunlap, W. P., Cortina, J. M., Vaslow, J. B., & Burke, M. J. (1996). Meta-analysis of
experiments with matched groups or repeated measures designs. Psychological
Methods, 1, 170-177.
Duyck, M., Collins, T., & Wexler, M. (2015). Does time stop when we blink? Journal
of Vision, 15(12):370, doi: 10.1167/15.12.370.

Continuity Across Eye Blinks
!
24!
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible
statistical power analysis program for the social, behavioral, and biomedical
sciences. Behavior Research Methods, 39, 175-191.
Gawne, T. J., & Martin, J. M. (2000). Activity of primate V1 cortical neurons during
blinks. Journal of Neurophysiology, 84, 2691-2694.
Hari, R., Salmelin, R., Tissari, S. O., Kajola, M., & Virsu, V. (1994). Visual stability
during eyeblinks. Nature, 367, 121-122.
Kanai, R., Paffen, C. L., Hogendoorn, H., & Verstraten, F. A. (2006). Time dilation in
dynamic visual display. Journal of Vision, 6(12):1421-30, doi: 10.1167/6.12.8.
Manning, K. A., Riggs, L. A., & Komenda, J. K. (1983). Reflex eyeblinks and visual
suppression. Perception & Psychophysics, 34, 250-256.
Needham, J. G. (1934). The time-error in comparison judgments. Psychological
Bulletin, 31, 229-243.
Riggs, L. A., Volkmann, F. C., & Moore, R. K. (1981). Suppression of the blackout due
to blinks. Vision Research, 21, 1075-1079.
Rouder, J. N., Speckman, P. L., Sun, D., Morey, R. D., & Iverson, G. (2009). Bayesian t
tests for accepting and rejecting the null hypothesis. Psychonomic Bulletin &
Review, 16, 225-237.
Schab, F. R., & Crowder, R. G. (1988). The role of succession in temporal cognition: Is
the time-order error a recency effect of memory? Perception & Psychophysics,
44, 233-242.
Stern, J. A., Walrath, L. C., & Goldstein, R. (1984). The endogenous eyeblink.
Psychophysiology, 21, 22-33.

Continuity Across Eye Blinks
!
25!
Tyrrell, R. A., & Owens, D. A. (1988). A rapid technique to assess the resting states of
the eyes and other threshold phenomena: The Modified Binary Search (MOBS).
Behavior Research Methods, Instruments, & Computers, 20, 137-141.
Volkmann, F. C., Riggs, L. A., & Moore, R. K. (1980). Eyeblinks and visual
suppression. Science, 207, 900-902.
Woodrow, H. (1935). The effect of practice on time-order errors in the comparison of
temporal intervals. Psychological Review, 42, 127-152.
Yarrow, K., Haggard, P., Heal, R., Brown, P., & Rothwell, J. C. E. (2001). Illusory
perceptions of space and time preserve cross-saccadic perceptual continuity.
Nature, 414, 302-305.
Yarrow, K., Johnson, H., Haggard, P., & Rothwell, J. C. (2004). Consistent chronostasis
effects across saccade categories imply a subcortical efferent trigger. Journal of
Cognitive Neuroscience, 16, 839-847.
Yarrow, K., Whiteley, L., Haggard, P., & Rothwell, J. C. (2006). Biases in the perceived
timing of perisaccadic perceptual and motor events. Perception & Psychophysics,
68, 1217-1226.
Yuasa, K., & Yotsumoto, Y. (2015). Opposite distortions in interval timing perception
for visual and auditory stimuli with temporal modulations. PLoS One, Aug
20;10(8):e0135646. doi: 10.1371/journal.pone.0135646.

Continuity Across Eye Blinks
!
26!
Figure Captions
Figure 1. Sequence of events for trials in Experiment 1.
Figure 2. Sequence of events for trials in Experiment 3.

Continuity Across Eye Blinks
!
27!

Continuity Across Eye Blinks
!
28!