Continuous visual control of interception.
ABSTRACT People generally try to keep their eyes on a moving target that they intend to catch or hit. In the present study we first examined how important it is to do so. We did this by designing two interception tasks that promote different eye movements. In both tasks it was important to be accurate relative to both the moving target and the static environment. We found that performance was more variable in relation to the structure that was not fixated. This suggests that the resolution of visual information that is gathered during the movement is important for continuously improving predictions about critical aspects of the task, such as anticipating where the target will be at some time in the future. If so, variability in performance should increase if the target briefly disappears from view just before being hit, even if the target moves completely predictably. We demonstrate that it does, indicating that new visual information is used to improve precision throughout the movement.
- SourceAvailable from: Robert J van Beers[Show abstract] [Hide abstract]
ABSTRACT: Do people perform a given motor task differently if it is easy than if it is difficult? To find out, we asked subjects to intercept moving virtual targets by tapping on them with their fingers. We examined how their behaviour depended on the required precision. Everything about the task was the same on all trials except the extent to which the fingertip and target had to overlap for the target to be considered hit. The target disappeared with a sound if it was hit and deflected away from the fingertip if it was missed. In separate sessions, the required precision was varied from being quite lenient about the required overlap to being very demanding. Requiring a higher precision obviously decreased the number of targets that were hit, but it did not reduce the variability in where the subjects tapped with respect to the target. Requiring a higher precision reduced the systematic deviations from landing at the target centre and the lag-one autocorrelation in such deviations, presumably because subjects received information about smaller deviations from hitting the target centre. We found no evidence for lasting effects of training with a certain required precision. All the results can be reproduced with a model in which the precision of individual movements is independent of the required precision, and in which feedback associated with missing the target is used to reduce systematic errors. We conclude that people do not approach this motor task differently when it is easy than when it is difficult.Experimental Brain Research 07/2013; · 2.22 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: An experiment was conducted in a driving simulator to test how eye-movement patterns evolve over time according to the decision-making processes involved in a driving task. Participants had to drive up to a crossroads and decide to stop or not. The decision-making task was considered as the succession of two phases associated with cognitive processes: Differentiation (leading to a prior decision) and Consolidation (leading to a final decision). Road signs (Stop, Priority and GiveWay) varied across situations, and the stopping behavior (Go and NoGo) was recorded. Saccade amplitudes and fixation durations were analyzed. Specific patterns were found for each condition in accordance with the associated processes: high visual exploration (larger saccade amplitudes and shorter fixation durations) for the Differentiation phase, and lower visual exploration (smaller saccades and longer fixations) for the Consolidation phase. These results support that eye-movements can provide good indexes of underlying processes occurring during a decision-making task in an everyday context.Journal of Eye Movement Research. 08/2014; 7(4):1-14.
- [Show abstract] [Hide abstract]
ABSTRACT: People are extremely good at hitting falling balls with a baseball bat. Despite the ball's constant acceleration, they have been reported to time hits with a standard deviation of only about 7 ms. To examine how people achieve such precision, we compared performance when there were no added restrictions, with performance when looking with one eye, when vision was blurred, and when various parts of the ball's trajectory were hidden from view. We also examined how the size of the ball and varying the height from which it was dropped influenced temporal precision. Temporal precision did not become worse when vision was blurred, when the ball was smaller, or when balls falling from different heights were randomly interleaved. The disadvantage of closing one eye did not exceed expectations from removing one of two independent estimates. Precision was higher for slower balls, but only if the ball being slower meant that one saw it longer before the hit. It was particularly important to see the ball while swinging the bat. Together, these findings suggest that people time their hits so precisely by using the changing elevation throughout the swing to adjust the bat's movement to that of the ball.Frontiers in Human Neuroscience 01/2014; 8:342. · 2.91 Impact Factor
Continuous visual control of interception
Eli Brenner⇑, Jeroen B.J. Smeets
Research Institute MOVE, Faculty of Human Movement Sciences, VU University Amsterdam, Van der Boechorststraat 9,
NL – 1081 BT Amsterdam, The Netherlands
a r t i c l ei n f o
Available online 25 February 2011
a b s t r a c t
People generally try to keep their eyes on a moving target that they
intend to catch or hit. In the present study we first examined how
important it is to do so. We did this by designing two interception
tasks that promote different eye movements. In both tasks it was
important to be accurate relative to both the moving target and
the static environment. We found that performance was more var-
iable in relation to the structure that was not fixated. This suggests
that the resolution of visual information that is gathered during the
movement is important for continuously improving predictions
about critical aspects of the task, such as anticipating where the
target will be at some time in the future. If so, variability in perfor-
mance should increase if the target briefly disappears from view
just before being hit, even if the target moves completely predict-
ably. We demonstrate that it does, indicating that new visual infor-
mation is used to improve precision throughout the movement.
? 2011 Elsevier B.V. All rights reserved.
In general, people look at objects when they interact with them or intend to interact with them
(Horstmann & Hoffmann, 2005; Johansson, Westling, Backstrom, & Flanagan, 2001; Land & Hayhoe,
2001; Mennie, Hayhoe, & Sullivan, 2007; Pelz, Hayhoe, & Loeber, 2001; Rothkopf, Ballard, & Hayhoe,
2007). This is also true when intercepting moving objects (Bahill & LaRitz, 1984; Brenner & Smeets,
2007, 2009; Mrotek & Soechting, 2007; Soechting & Flanders, 2008). However, the extent to which
pursuing a target is essential for catching or hitting is not yet clear (Brenner & Smeets, 2010; Dessing,
Oostwoud Wijdenes, Peper, & Beek, 2009; Sharp & Whiting, 1974, 1975). There are several reasons
why it may be advantageous to keep one’s eyes on the target (Wilmut, Wann, & Brown, 2006). The
0167-9457/$ - see front matter ? 2011 Elsevier B.V. All rights reserved.
E-mail address: firstname.lastname@example.org (E. Brenner).
Human Movement Science 30 (2011) 475–494
Contents lists available at ScienceDirect
Human Movement Science
journal homepage: www.elsevier.com/locate/humov
most obvious one is that doing so ensures that the resolution with which the visual information is ac-
quired is maximal. We here evaluate whether this is the main reason for doing so.
Keeping the fovea directed at the ball ensures that one has access to the highest possible spatial
resolution when localising the ball and judging its trajectory. This is probably particularly important
if the ball’s trajectory is not completely predictable, so that one must constantly consider whether one
needs to adjust one’s movement. If it is likely that the movement will have to be adjusted at a certain
moment, such as occurs when one can anticipate that the ball will bounce off an uneven surface, peo-
ple make sure to have their eyes on the ball at that moment (Land & McLeod, 2000). If the trajectory is
predictable, it is probably less important to have the highest possible spatial resolution throughout the
movement. Indeed, for reasonably predictable trajectories of a ball, it is not even necessary to see the
entire trajectory in order to catch the ball (e.g., López-Moliner, Brenner, Louw, & Smeets, 2010;
Whiting & Sharp, 1974). The extent to which vision at various moments is essential for successful
interception is widely debated (e.g., Bootsma & van Wieringen, 1990; Dubrowski, Lam, & Carnahan,
2000; Marinovic, Plooy, & Tresilian, 2009; Müller & Abernethy, 2006; Sharp & Whiting, 1974; Teixeira,
Chua, Nagelkerke, & Franks, 2006; van Soest et al., 2010; Young & Zelaznik, 1992). If there are
moments at which visual information is not very important, it is also unlikely to be necessary to
pursue the target at such moments.
We recently proposed that even for completely predictable target motion it is advantageous to
keep one’s eyes on the target throughout the movement (rather than only at the start), because if
one maintains a high visual resolution, the accuracy with which one can predict where the target will
be when one reaches it will keep increasing as the movement progresses (Brenner & Smeets, 2009).
Directing one’s gaze towards the target early in the movement helps ensure that the movement starts
off more or less correctly, so that only modest adjustments are later needed, and keeping one’s eyes on
the target ensures that such modest adjustments are based on increasingly accurate estimates as the
duration of the prediction decreases because the hand approaches the target. The first experiment of
this study was designed to directly examine to what extent pursuing the target with one’s eyes until
one hits it is beneficial when intercepting targets that move in a completely predictable manner.
2. Experiment 1: eye movements
Virtual targets moved from left to right at a constant velocity across a surface. They were to be hit
with a stylus. The stylus was initially at a starting point near the subject’s body. When intercepting
such targets, subjects tend to pursue the target with their eyes for most of the time (Brenner & Smeets,
2007, 2009). Even if subjects are explicitly instructed to fixate a static point, they cannot avoid follow-
ing the moving target with their eyes just before hitting it (Brenner & Smeets, 2010). Moreover, even if
we could train subjects not to pursue the target, adding such a second task could influence subjects’
precision (Wilmut et al., 2006). We therefore wanted to influence the eye movements without any ex-
plicit instructions or constraints. To do so we compared interception in two slightly different tasks
that required a similar spatial and temporal accuracy, but were designed to give rise to different
eye movements: hitting a small target into a gap and hitting a target through a small gap.
We reasoned that subjects would want to direct their gaze towards the smallest relevant structure,
which would be the small target when the task was to hit the moving target into the larger gap, but
would be the small gap when the task was to hit through the static gap just as the larger target passes
behind the gap. In these tasks the smallest structure was also the first one that the subject’s hand
encounters, which is also likely to encourage them to direct their gaze towards it. We expect this to
have consequences for the precision of their hand movements, which we expect to be highest in rela-
tion to the structure that they are looking at.
The setup and tasks are shown schematically in Fig. 1. Images were projected at 85 Hz and a res-
olution of 1024 by 768 pixels onto a back-projection screen that was 20 cm above a half-silvered
E. Brenner, J.B.J. Smeets/Human Movement Science 30 (2011) 475–494
mirror. There was a large (WACOM A2) drawing tablet 20 cm below the mirror, positioned so that it
coincided precisely with the apparent position of the screen as seen through the mirror. Subjects inter-
cepted the virtual targets by moving a stylus across the drawing tablet. The tablet determined the sty-
lus’ position at 200 Hz. Lamps between the half-silvered mirror and the drawing tablet (not shown)
ensured that subjects could clearly see the stylus and their hand as well as the target. The setup
was calibrated by having the experimenter align the tip of the stylus with small disks presented on
the screen, allowing us to later present images of any desired dimensions at any desired position
on the surface of the drawing tablet.
Movements of the subjects’ eyes were recorded at 250 Hz using an Eyelink II (SR Research Ltd.,
Canada). Eye orientation was calibrated by having subjects follow jumping discs with their eyes before
the session. We related distances between the horizontal and vertical positions on the screen to
changes in the pupil positions reported by the Eyelink to later be able to determine the velocity of
the eye. Head movements were not accounted for when calculating pursuit gain. Whenever we report
where subjects were looking (rather than the eye’s velocity) we account for the initial orientation of
the head and for any drift in the Eyelink data on the basis of the subject’s fixation of the starting point
just before the trial (see Section 2.1.5).
Delays within our setup were accounted for, both when analysing the data and when providing
feedback to the subjects. We considered the actual time at which the images were presented (consid-
ering rendering delays and delays in the projector) and the actual time at which the stylus was at a
given position (considering the time it takes the tablet to measure the position and convey it to the
computer) for everything except during the short interval just after the stylus hit the target. Since
the interception was only registered 62 ms after it had occurred, the target moved on for 5 frames be-
fore the appropriate feedback could be presented. Subjects did not notice this, probably at least partly
because the hand occluded the target at that moment.
2.1.2. Tasks and conditions
The two tasks were to hit a target as it passed behind a gap in a grey bar, and to hit a target into a
gap in a bar (Fig. 1). The white target and the grey bar both always extended 1 cm in depth (the irrel-
evant dimension). The target always moved to the right at 20 cm/s. The path of the center of the target
was 20 cm further from the subject than the center of the 5 mm diameter circular starting position.
The bar was either closer than the target, so that the target moved just behind the bar, or beyond
the target, so that the target moved just in front of the bar. Although the target and the bar both
task: hit the target
into the gap
task: hit the target
through the gap
target size: 1 cmgap size: 1 cm
Fig. 1. Subjects had to intercept a moving target by sliding a stylus across the surface of a drawing tablet. The starting point, the
target and a grey bar with a gap in it, were projected on a screen above a half-silvered mirror (A). The distance between the
screen and the mirror was identical to that between the tablet and the mirror, so that the images appeared to be on the tablet.
Subjects could see their hand through the mirror. The task was either to hit the target as it passed behind the gap, without
hitting the bar (B), or to hit the target when it was in front of the gap (C). In both tasks it took the target 100 ms to cross the gap,
there were two possible gap positions (left or right; 1 cm apart) and the target was midway across the gap at one of two possible
times after it appeared (600 or 650 ms). The target always moved to the right at 20 cm/s.
E. Brenner, J.B.J. Smeets/Human Movement Science 30 (2011) 475–494
extended 1 cm in depth, the analysis only considered the near surface of the target and the center of
the bar. Thus we did not consider a target to be hit if the stylus entered it from the side, or a bar to be
hit if the stylus grazed its corner but was between it and the other bar by the time the stylus was half
way through the gap.
Within each task there were 4 conditions. The gap was either 5 mm to the left or 5 mm to the right
of the center (where the bar was closest to the starting position). The moving target appeared 12 or
13 cm to the left of the gap, so that it reached the gap after either 600 or 650 ms (we will refer to this
time as the ‘urgency’). The different positions of the gap and the different urgencies were introduced
so that visual information about the target’s position and motion had to be used to perform ade-
quately, while simply reproducing a stereotyped movement when the target appeared would lead
to systematic errors.
When the task was to hit the target as it passed behind the gap, the gap was 1 cm wide and the
target extended 2 cm laterally. Subjects were clearly instructed that they should hit the target without
hitting the bar. If they hit the target, it shifted 1 cm away from the bar and stopped moving, and sub-
jects heard a sound. If they missed the target, it continued moving. If they hit the bar, it turned red and
the target continued moving (even if it had been hit). When the task was to hit the target into the gap,
the target was a 1 cm square and the gap was 3 cm wide. Subjects had to hit the target while it was
completely in front of the gap. If they did so successfully, it moved into the gap and stopped there, and
they heard a sound. If they missed the target, it continued moving. If they hit the target while any part
of it was in front of the bar, the bar turned red and the target continued moving along its original path.
In both tasks subjects had about 100 ms to successfully hit the target and had to hit within 1 cm,
but in the first task this 1 cm was static while in the second it was moving. Assuming that subjects
would try to optimize their performance by directing their eyes to where the highest resolution is
needed, we predicted that this difference would give rise to different eye movements. Thus comparing
these two tasks should let us compare performance for different eye movements without us having to
explicitly constrain subjects’ eye movements.
2.1.3. Subjects and procedure
Sixteen subjects took part in the experiment. They could adjust the height and position of the chair
that they sat on as they pleased, to ensure that they could move comfortably, but could not move their
head very far forward because of the mirror. The two tasks were performed in separate blocks of trials
within a single session. Half of the subjects started with the task of hitting through the gap, while the
other half started with the task of hitting into the gap. Each block started with 20 practice trials (5 for
each condition) directly followed by the 100 trials that were later analysed (25 for each condition).
Within each block the conditions were presented in random order. Subjects could rest at any moment
by not placing the stylus at the starting point. As soon as they did place the stylus at the starting point
a new target appeared. The stylus was considered to have been placed at the starting point if its tip
was within the 5 mm diameter of the starting point and moved by less than 1 mm in 250 ms.
Since subjects had to place the stylus exactly on a small disc (the starting point) to start the exper-
iment, they had to direct their gaze towards that position before the trial. However, since the target
and the bar with the gap only appeared after they had kept the stylus there for 250 ms, subjects often
made a vertical saccade towards the region in which the target or gap will appear before they actually
appeared. We were mainly interested in what the eyes where doing while the stylus was moving to-
wards the target. We determined the horizontal eye velocity from the average displacement of the two
eyes. The mean horizontal eye velocity during the last 200 ms before the hit was determined for each
trial. The median of these velocities was determined for each subject and task (using the median elim-
inated the need to consider occasional saccades during that period).
In our analysis of the hand movements we only consider three times: the time the target (and bar
with gap) appeared, the time the hand started to move, and the time of the hit. The hand is considered
to have started moving when it moved 5 mm from where it was when the target appeared. The time at
which the stylus reached the path of the near edge of the target is considered to be the time of the hit,
even if the stylus did not actually hit the target. Linear interpolation was used to achieve a higher
E. Brenner, J.B.J. Smeets/Human Movement Science 30 (2011) 475–494
resolution than provided by the sampling and presentation rates. The reaction time is the interval be-
tween when the target appears and when the hand starts to move. The movement time is the interval
between when the hand starts to move and the time of the hit.
We evaluated whether subjects managed to hit the targets, but our main interest was in the var-
iability (and systematic errors) across repetitions of the same kind of trials. When determining these
measures we made no distinction between hits and misses. For each subject, task and condition (gap
on left or right; target aligned with the gap after 600 or 650 ms) we determined the mean and stan-
dard deviation of 3 measures: the position of the target at the time of the hit, the position of the stylus
at the time of the hit, and the position of the stylus relative to the target at the time of the hit. Our
main interest is in possible correlations (across subjects and tasks) between the standard deviations
in these measures and eye velocity. We predict that pursuing the target with one’s eyes will increase
the variability in the stylus’ position (relative to the static surrounding) at the time of the hit, but will
decrease the variability in its position relative to the target. To examine whether this is so we will cor-
relate the subjects’ differences in pursuit gain between the two tasks with the differences in the var-
iability of the position of the stylus on the tablet at the time of the hit and with differences in the
variability of the stylus’ position relative to the target at the time of the hit.
Of the 3200 trials (16 subjects; 2 tasks; 4 conditions; 25 trials each), 19 could not be analysed be-
cause the subject failed to move the stylus or lifted it off the tablet. Fig. 2 shows one subject’s horizon-
tal eye movements and stylus velocities for five consecutive trials of each task. When the task was to
hit the target into the gap, this subject made a leftward saccade towards the target about 250 ms after
the target appeared, and then pursued the target until it was hit (Fig. 2A). Thus the eyes were pursuing
the target throughout the stylus movement (Fig. 2C). When the task was to hit the target through the
gap, she made a smaller saccade, presumably towards the gap, about 250 ms after the target and gap
appeared (Fig. 2B). The stronger tendency to pursue the target with the eyes when hitting into the gap
was quite consistent across subjects, as was the higher velocity of the stylus for that task (see below).
Comparing Figs. 2B and D we see additional velocity peaks when hitting into the gap. Subjects had
periods during which they moved in different ways, but the additional peaks in the velocity profile
are not characteristic for a certain task, because across subjects there were about as many trials with
additional peaks for both the tasks.
Subjects hit more targets when the task was to hit through the gap (Fig. 3A) than when it was to hit
into the gap (Fig. 3B; paired t-test: t(15) = 2.6, p = .02). In particular, subjects missed almost half the
moving, 1 cm wide targets when trying to hit them into the gap (missed moving target and place
and time wrong sectors in Fig. 3B), whereas they only missed the 1 cm wide gap on about a fifth of
the trials when hitting through the gap (missed gap and place and time wrong sectors in Fig. 3A). As
expected, subjects were more inclined to pursue the target with their eyes when hitting a small mov-
ing target into a large gap, than when hitting a large target through a small static gap (Fig. 3C; paired t-
test: t(15) = 5.5, p < .0001). The average pursuit gain during the last 200 ms before the hit was about
0.2 when hitting through the gap and 0.7 when hitting into the gap.
Table 1 and Fig. 4 provide average values of various measures. A repeated measures analysis of var-
iance (with factors task, urgency and gap position) on subjects’ median reaction times revealed that
reaction times were significantly shorter when there was more urgency (p < .001) and when the
gap was on the right (p = .04), and that there was a significant interaction between urgency and gap
position (p = .03). A similar analysis on the median movement times revealed that movement times
were significantly shorter when there was more urgency (p < .001), when the gap was on the right
(p = .006) and when hitting into the gap (p = .03), and a significant interaction between task and ur-
gency (p = .04). Similar repeated measures analyses of variance revealed that the standard deviation
in the position of the target when hit and in the position of the target relative to the stylus at the time
of the hit were smaller when hitting into the gap than when hitting through the gap (p = .04 and
p < .0001, respectively), whereas the standard deviation in the position of the stylus at the time of the
hit was smaller when hitting through the gap (p < .0001). The only other significant effect for the three
E. Brenner, J.B.J. Smeets/Human Movement Science 30 (2011) 475–494