Action Video Game Modifies Visual Selective Attention

letters to nature
..............................................................
Action video game modiﬁes visual
selective attention
C. Shawn Green & Daphne Bavelier
Department of Brain and Cognitive Sciences, Center for Visual Science, University
of Rochester, Rochester, New York 14627, USA
.............................................................................................................................................................................
As video-game playing has become a ubiquitous activity in
today’s society, it is worth considering its potential consequences
on perceptual and motor skills. It is well known that exposing an
organism to an altered visual environment often results in
modiﬁcation of the visual system of the organism. The ﬁeld of
perceptual learning provides many examples of training-induced
increases in performance. But perceptual learning, when it
occurs, tends to be speciﬁc to the trained task; that is, generali-
zation to new tasks is rarely found1–10. Here we show, by contrast,
that action-video-game playing is capable of altering a range of
visual skills. Four experiments establish changes in different
aspects of visual attention in habitual video-game players as
compared with non-video-game players. In a ﬁfth experiment,
non-players trained on an action video game show marked
improvement from their pre-training abilities, thereby establish-
ing the role of playing in this effect.
We ﬁrst used the ﬂanker compatibility effect, a standard exper-
imental paradigm in attentional studies, to determine whether
video-game playing produces an overall increase in attentional
capacity11. This task measures the effect of a to-be-ignored dis-
tractor on a target task. The size of this distractor effect has been
proposed to provide an index of ‘left-over’ attentional resources. In
Figure 1 Measure of attentional resources. a, Flanker compatibility task. Participants
particular, it has been reported12 that the distractor effect is large
were asked to decide whether a square or a diamond appeared within one of the six rings
when the target task is easy, but small when the target task is made
(target task), while ignoring a distractor shape presented outside the rings. The distractor
difﬁcult by increasing the number of ﬁller shapes in the ring (Fig.
shape could be either compatible (the same as the target shape) or incompatible (the
1a). The currently accepted basis for this ﬁnding is that when the
alternative target shape). Of interest was the difference in target processing speed
target task is easy, spare attentional resources ‘spill over’ to the
between compatible and incompatible trials. This difference, also called the ‘compatibility
distractor, processing it to some extent and thereby inﬂuencing
effect’, measures the attentional resources available to the participant. By inserting no,
target processing. As the target task becomes more difﬁcult, fewer
one, three or ﬁve extra shapes into the rings, the difﬁculty of the target task was
attentional resources remain to process the extraneous distractor.
manipulated to test the possibility that VGPs show compatibility effects at task difﬁculties
Irrelevant processing is only truly prevented once the target task
for which attentional resources are usually exhausted in NVGPs. b, Flanker compatibility
becomes so difﬁcult that all available resources are devoted to the
effect. As previously reported12, the compatibility effect (and spare attentional resources)
target task, leaving none to the distractor.
decreased with task difﬁculty in NVGPs (task difﬁculty £ compatibility, P , 0.001). In
This task therefore affords an opportunity to test the hypothesis
VGPs, by contrast, the compatibility effect remained as the target task was made more
that video-game playing increases the capacity of the visual atten-
difﬁcult (population £ task difﬁculty £ compatibility, P , 0.003). As a result, VGPs
tional system. If video-game players (VGPs) do indeed have a
showed a signiﬁcantly greater compatibility effect than NVGPs when the target task was
greater attentional capacity, they should exhaust their visual atten-
difﬁcult, indicating greater attentional resources in VGPs (**P , 0.01). Error bars denote
tion resources more slowly than non-video-game players (NVGPs)
s.e.m.
as the target task becomes more difﬁcult. As predicted by the
hypothesis of a greater capacity in their population, VGPs showed
video-game playing enhances the number of visual items that
an overall distractor effect that remained even when the target task
can be unerringly apprehended.
was difﬁcult (Fig. 1b). Thus, at difﬁculty levels where NVGPs have
These two experiments indicated that video-game playing
long depleted their attentional resources, VGPs possess sufﬁcient
enhances attentional capacity in an area well within the ‘training’
resources to perform the target task, with resources still remaining
(video-game playing) zone (0–58 from ﬁxation); however, it was
to spill over to the distractors. These results indicate that VGPs
unclear whether video-game playing also facilitates processing
possess enhanced attentional capacity.
outside the training range. To test this possibility, we compared
We conﬁrmed this increase in capacity more directly through
the distribution of visual attention in VGPs and NVGPs at three
the use of an enumeration task. When asked to report how many
locations, one well within the training range (108), one at its
squares are presented in a brieﬂy ﬂashed display, normal subject
boundary (208) and one well outside the training range (308; all
performance seems to be best captured by two distinct processes:
VGPs reported playing in a ﬁeld that extends no more than 188 from
an ‘automatic’ operation for small numbers of targets in which
ﬁxation). We adapted the ‘useful ﬁeld of view’ task (refs 17–19 and
performance is rapid, accurate and independent of number
Fig. 3a), which measures participants’ ability to locate a target
(subitizing); and a slower, serial counting process for larger
amongst distractors. This measure does not correlate well with
numbers of items13. The number of items that can be subitized
standard tests of visual acuity but rather provides a measure of
gives an estimate of the number of items that can be attended at
attentional resources and their spatial distribution17. VGPs far
once14–16. VGPs could subitize more items than NVGPs (4.9
outperformed NVGPs at all eccentricities (Fig. 3b), distractor levels
versus 3.3 items; population effect, P , 0.01; Fig. 2). Thus,
and distractor–eccentricity pairs. These results indicate an
534
© 2003 Nature Publishing Group
NATURE | VOL 423 | 29 MAY 2003 | www.nature.com/nature

letters to nature
the ability to process items over time, particularly the ability to
avoid ‘bottlenecks’ of attention that often occur in temporal
processing. The experimental paradigm that we used to test the
temporal aspects of visual attention is the attentional blink task (refs
20–22 and Fig. 4a). Attentional blink refers to the phenomena
wherein subjects have difﬁculties reporting a second target when it
appears a few hundreds of milliseconds after the onset of a ﬁrst
target. In our case, we used a variant of the attentional blink, in
which participants were asked to identify a ﬁrst target and then
detect a second target. This task includes two distinct attentional
bottlenecks. First is the attentional blink per se, which is the
difﬁculty of processing a second target that comes 200–500 ms
after the onset of the ﬁrst one. Second is the cost of switching
tasks between the ﬁrst and second target (from identiﬁcation to
detection); unlike the attentional blink per se, this effect is most
Figure 2 Enumeration performance. When asked to report the number of squares brieﬂy
pronounced when the two targets are temporally adjacent and then
ﬂashed, VGPs were able to apprehend more items at once than were NVGPs (4.9 versus
decreases slowly as the time between the two targets increases. This
3.3). Overall, VGPs were signiﬁcantly more accurate than NVGPs (78% versus 65%,
attentional bottleneck is not speciﬁc to vision but rather appears
P , 0.003) and, as expected, this difference only emerged for numbers above the
amodal21,23,24. Thus, by using an identiﬁcation/detection attentional
subitizing range of NVGPs. Error bars denote s.e.m. (*P , 0.05, **P , 0.01).
blink task, we could test whether the enhanced capacities after
video-game training not only applied to a purely visual bottleneck
enhanced allocation of spatial attention over the visual ﬁeld, even at
but also generalized to an amodal one.
untrained locations, in VGPs.
The three experiments described so far indicated that video-game
playing enhances the capacity of visual attention and its spatial
distribution. We next examined the temporal characteristics of
visual attention and asked whether the pressure to act rapidly on
several visual items, which is inherent to most action games, alters
Figure 4 Measure of attention over time. a, Attentional blink task. Black letters were
rapidly presented at ﬁxation in a standard rapid, serial, visual presentation manner. At a

random time in the stream, a white letter was presented (ﬁrst target). After this ﬁrst target,
an ‘X’ (second target) was presented among the following letters 50% of the time. After
Figure 3 Measure of attention over space. a, Sequence of displays in the useful ﬁeld of
the trial, the subject gave the identity of the ﬁrst target and then indicated whether the ‘X’
view task. Participants indicated the spoke on which the small target (triangle within circle)
was presented. Of interest is the performance of subjects on the ‘X’ detection task, given
appeared. By presenting the target at different eccentricities, this task affords an
that they have correctly identiﬁed the white letter. b, Attentional blink performance. At
opportunity to test the spatial distribution of visual attention. b, Localization accuracy.
early lags, VGPs performed better (less blink) than NVGPs; as lag increased, the effect of
VGPs showed large enhancements in localization ability at all eccentricities. The
the attentional bottlenecks decreased and, as expected, the two populations became
superiority of VGPs at 308 indicates that the enhancement of spatial attention observed in
comparable (lag by population P , 0.02). Error bars denote s.e.m. Points without error
this population is not limited to trained locations. Error bars denote s.e.m. (**P , 0.01).
bars indicate that the s.e.m. was smaller than the size of the square (**P , 0.01).
NATURE | VOL 423 | 29 MAY 2003 | www.nature.com/nature
© 2003 Nature Publishing Group
535

letters to nature
VGPs outperformed NVGPs on second-target correct detections
the VGP population) for 1 h per day for 10 consecutive days. A
from lag 1 to lag 5, indicating less attentional blink (Fig. 4b). The
control group was trained, over the same time span, on the game
ﬁnding of a population difference as early as lag 1 indicates that
Tetris. This game contains a challenging visuo-motor component
video-game training enhances task-switching abilities as well as
but, whereas action games require that attention is distributed
decreasing the attentional blink. Thus, both the visual and amodal
and/or switched around the ﬁeld, Tetris demands focus on one
bottlenecks identiﬁed during temporal processing of visual infor-
object at a time. Tetris, therefore, would not be expected to change
mation are reduced in VGPs. Clearly, these individuals have an
the aspects of visual attention described above and thus affords an
increased ability to process information over time; however,
excellent control for monitoring improvement due to enhanced
whether this is due to faster target processing, such as faster
visuo-manual expertise and test–retest improvements.
selection and stabilization of information in memory, or to an
Before and after training we tested each group on the enumera-
increased ability to maintain several attentional windows in parallel,
tion, useful-ﬁeld-of-view and attentional-blink experiments. The
cannot be determined from our current data.
training was successful as all participants improved their scores on
We also carried out a training experiment to alleviate concerns
the video game on which they were trained. Notably, action-game
over the source of the VGP and NVGP population differences. By
training led to greater performance improvement than did the
selecting VGPs, we may have selected individuals with inherently
control game on all three experimental tasks. Enumeration
better attentional skills than NVGPs. The consequence of this
increased by 1.7 items in individuals trained on action games,
potential split would be that VGPs, with their greater natural ability,
whereas the control group showed no improvement (population
did well at video games and so played often, whereas NVGPs, whose
by improvement, P , 0.05). Training on action video games
inherent abilities limited their success, avoided playing games as a
enhanced the useful ﬁeld of view and lead to faster recovery from
result. Another possible confound was that, because of relatively
the attentional blink (Fig. 5). In addition, we found that the
superior visuo-motor coordination, VGPs were better able to per-
improvement in these attentional measures was marginally corre-
form the motor aspect of the tasks and thus had more spare
lated with improvement in video-game playing (adjusted r 2 ¼ 0.43,
resources to devote to the visually demanding part of the tasks.
P ¼ 0.13; see Supplementary Information). Thus, 10 days of train-
To rule out these potential confounds, a group of NVGPs under-
ing on an action game is sufﬁcient to increase the capacity of visual
went action-video-game training, in which they were asked to play
attention, its spatial distribution and its temporal resolution.
Medal of Honor (a game similar to those reported being played by
By forcing players to simultaneously juggle a number of
varied tasks (detect new enemies, track existing enemies and
avoid getting hurt, among others), action-video-game playing
pushes the limits of three rather different aspects of visual
attention. It leads to detectable effects on new tasks and at
untrained locations after only 10 days of training. Therefore,
although video-game playing may seem to be rather mindless,
it is capable of radically altering visual attentional processing.
There are several ways by which video-game training could
lead to such enhancements. Changes in known attentional
bottlenecks is certainly a possibility; however, speeded percep-
tual processes and/or better management of several tasks at the
central executive level are also likely to contribute. It will be
for future studies of the effect of video-game practice to
determine the relative contribution of these different factors
to skill learning.
A
Methods
Participants
Subjects were aged between 18 and 23 years. The VGPs had played action video games on at
least 4 days per week for a minimum of 1 h per day for the previous 6 months. The games
included Grand Theft Auto3, Half-Life, Counter-Strike, Crazy Taxi, Team Fortress Classic,
007, Spider-Man, Halo, Marvel vs Capcom, Roguespeare and Super Mario Cart. The NVGPs
had little, and preferably no, video-game usage in the past 6 months. Experiments 1–4
included only males; in experiment 5, both male and female NVGPs underwent training.
Apparatus, stimuli and procedure
For experiment 1—ﬂanker compatibility (Fig. 1)—the apparatus, stimulus, procedure and
data analysis have been described25.
For experiment 2—enumeration (Fig. 2)—the stimuli consisted of a random number
(between 1 and 12) of white squares (each subtending 0.58 £ 0.58) presented for 50 ms
within an invisible 58 £ 58 square centred on the screen. Subjects were asked to type the
number of squares that they believed were presented; no feedback was provided. We scored
the percentage of trials in which participants correctly reported the number of squares
presented. Only data from trials with 1–10 squares are presented, as trials with 11 and 12
squares were excessively difﬁcult for all populations. To estimate the point at which
subjects switched from subitizing to counting, we modelled the data of each subject with a
bilinear curve. The ﬁrst line was set to have an intercept of (0,0) and was expected to have a
rather shallow slope (subitizing range); the second line was expected to have a steeper slope
(counting). The best bilinear ﬁt was found by minimizing the least-square error. The point
Figure 5 Performance before and after training. a, Training: useful ﬁeld of view. At each
at which the two lines met was then computed and considered to be the point where
eccentricity, the group trained on an action video game improved signiﬁcantly more from
subitizing switches to counting.
For experiment 3—useful ﬁeld of view (Fig. 3)—the stimuli and procedure were similar
their pre-test scores than did the control group trained on a non-action video game
to those used previously17, except that the stimulus size and presentation time were both
(**P , 0.01). b, Training: attentional blink. The group trained on an action video game
decreased to account for the increased ability of comparatively younger subjects (target
recovered faster from the attentional blink than did the control group trained on a non-
stimulus was a ﬁlled triangle within a 38 £ 38 circle outline; distracting squares subtended
action video game (*P , 0.05, **P , 0.01).
48 £ 48, no central task). Separate timings were used to ensure equal difﬁculty at each
536
© 2003 Nature Publishing Group
NATURE | VOL 423 | 29 MAY 2003 | www.nature.com/nature

letters to nature
eccentricity: 6 ms for targets presented at 108; and 12 ms for targets present at either 208 or
search: the visual attentional blink and an amodal task-switch deﬁcit. J. Exp. Psychol. Human Percept.
308.
Perform. 24, 979–992 (1998).
For experiment 4—attentional blink (Fig. 4)—the stimuli and procedure described
25. Proksch, J. & Bavelier, D. Changes in the spatial distribution of visual attention after early deafness.
previously20 were used except that eight lags instead of nine, and eight trials per condition
J. Cogn. Neurosci. 14, 1–5 (2002).
instead of ten, were used. VGPs had almost signiﬁcantly higher-second-target detection
accuracy in the control condition (detecting the X in the absence of a ﬁrst target, 95.6% to
Supplementary Information accompanies the paper on www.nature.com/nature.
87.9%, P ¼ 0.061). Conservatively, all analyses were done on the difference between the
variable of interest and the baseline detection condition to eliminate different baseline
Acknowledgements We thank T. Monacelli, D. McColgin, J. Cohen and K. Schneider for help
abilities as a potential confound in interpreting the depth of the attentional blink between
with subjects and software support; and R. Aslin, A. Pouget and D. Knill for feedback on the
populations.
manuscript. This research was supported by the NIH and the James S. McDonnell-Pew
For experiment 5—training (Fig. 5)—all subjects were tested on the useful-ﬁeld-of-
Foundation.
view, attentional-blink and enumeration tasks (in this order) before training; they then
underwent training for 1 h per day for 10 days with one of two possible kinds of video
game. After training they were retested on the same three tasks (in the same order). The
Competing interests statement The authors declare that they have no competing ﬁnancial
experimental group trained on Medal of Honor: Allied Assault (Electronic Arts, 2002). This
interests.
game simulates Second World War combat situations and was chosen to be similar to
those played by our VGPs. It has a relatively simple interface, uses ﬁrst-person point of
Correspondence and requests for materials should be addressed to D.B.
view and requires effective monitoring of the whole visual ﬁeld (extent from ﬁxation about
(daphne@cvs.rochester.edu).
138 (height) £ 168 (width)). Subjects played the game straight through for the ﬁrst 8 days.
On days 9 and 10, they returned to the beginning of the game to quantitatively measure
their improvement (comparing performance over mission 1 during their ﬁrst and second
playings). In all available measures of improvement taken (accuracy and deaths to
complete mission), subjects improved their performance after training (accuracy, 17%
..............................................................
improvement; deaths to complete, 42% improvement).
For the control group (n ¼ 8), Tetris (Original Tetris, AcademySoft Elorgwhich, 1987;
Mitochondrial membrane
Tengen, 1988) was displayed to cover the entire extent of the screen. Because Tetris adds
graphics on the side of the screen, however, the effective game area extended 138
remodelling regulated by a
(height) £ 98 (width) from ﬁxation. This game was selected to control for the effect of
improved visuo-motor coordination, while putting few demands on the processing of
several objects at once. Accordingly, the version of Tetris on which subjects were trained
conserved rhomboid protease
had the preview block option turned off. In the two measures of improvement available
(high score, maximum level reached), all subjects improved after training (high score, 71%
G. Angus McQuibban, Saroj Saurya & Matthew Freeman
improvement; mean level improvement, 67% improvement).
Received 4 December 2002; accepted 8 April 2003; doi:10.1038/nature01647.
MRC Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, UK
1. Sagi, D. & Tanne, D. Perceptual learning: learning to see. Curr. Opin. Neurobiol. 4, 195–199
.............................................................................................................................................................................
(1994).
Rhomboid proteins are intramembrane serine proteases that
2. Ball, K. & Sekuler, R. A speciﬁc and enduring improvement in visual motion discrimination. Science
activate epidermal growth factor receptor (EGFR) signalling in
218, 697–698 (1982).
Drosophila1. Rhomboids are conserved throughout evolution2–5,
3. Fiorentini, A. & Berardi, N. Perceptual learning speciﬁc for orientation and spatial frequency. Nature
287, 43–44 (1980).
and even in eukaryotes their existence in species with no EGFRs
4. Karni, A. & Sagi, D. Where practice makes perfect in texture discrimination: evidence for primary
implies that they must have additional roles. Here we report that
visual cortex plasticity. Proc. Natl Acad. Sci. USA 88, 4966–4970 (1991).
Saccharomyces cerevisiae has two rhomboids, which we have
5. Poggio, T., Fahle, M. & Edelman, S. Fast perceptual learning in visual hyperacuity. Science 256,
named Rbd1p and Rbd2p. RBD1 deletion results in a respiratory
1018–1021 (1992).
6. Fahle, M. & Edelman, S. Long-term learning in vernier acuity: effects of stimulus orientation, range,
defect; consistent with this, Rbd1p is localized in the inner
and of feedback. Vision Res. 33, 397–412 (1993).
mitochondrial membrane and mutant cells have disrupted mito-
7. Ramachandran, V. S. & Braddick, O. Orientation-speciﬁc learning in stereopsis. Perception 2, 371–376
chondria. We have identiﬁed two substrates of Rbd1p: cyto-
(1973).
8. Ahissar, M., Laiwand, R., Kozminsky, G. & Hochstein, S. Learning pop-out detection: building
chrome c peroxidase (Ccp1p); and a dynamin-like GTPase
representations for conﬂicting target-distractor relationship. Vision Res. 38, 3095–3107
(Mgm1p), which is involved in mitochondrial membrane
(1998).
fusion6–10. Rbd1p mutants are indistinguishable from Mgm1p
9. Ahissar, M. & Hochstein, S. The spread of attention and learning in feature search: effects of target
mutants, indicating that Mgm1p is a key substrate of Rbd1p and
distribution and task difﬁculty. Vision Res. 40, 1349–1364 (2000).
10. Sirenteanu, R. & Regina, R. Perceptual learning in visual search generalizes over task locations and
explaining the rbd1D mitochondrial phenotype. Our data indi-
eyes. Vision Res. 40, 2925–2949 (2000).
cate that mitochondrial membrane remodelling is regulated by
11. Eriksen, B. A. & Eriksen, C. W. Effects of noise letters upon the identiﬁcation of a target letter in
cleavage of Mgm1p and show that intramembrane proteolysis by
nonsearch task. Percept. Psychophys. 16, 143–149 (1974).
rhomboids controls cellular processes other than signalling. In
12. Lavie, N. & Cox, S. On the efﬁciency of visual selective attention: efﬁcient visual search leads to
inefﬁcient distractor rejection. Psychol. Sci. 8, 395–398 (1997).
addition, mitochondrial rhomboids are conserved throughout
13. Kaufman, E., Lord, M., Reese, T. & Volkman, J. The discrimination of visual number. Am. J. Psychol.
eukaryotes and the mammalian homologue, PARL11, rescues the
62, 498–525 (1949).
yeast mutant, suggesting that these proteins represent a func-
14. Trick, L. M. & Pylyshyn, Z. W. What enumeration studies can show us about spatial attention:
tionally conserved subclass of rhomboid proteases.
evidence for limited capacity preattentive processing. J. Exp. Psychol. Human Percept. Perform. 19,
331–351 (1993).
Despite their widespread conservation, the only known function
15. Trick, L. M. & Pylyshyn, Z. W. Why are small and large numbers enumerated differently? A limited-
of eukaryotic rhomboid proteases is the activation of EGFR signal-
capacity preattentive stage in vision. Psychol. Rev. 101, 80–102 (1994).
ling in Drosophila12–15. We therefore examined their function in
16. Tuholski, S. W., Engle, R. W. & Baylis, G. C. Individual differences in working memory capacity and
enumeration. Mem. Cogn. 29, 484–492 (2001).
S. cerevisiae, which has no receptor tyrosine kinases but has two
17. Ball, K. K., Beard, B. L., Roenker, D. L., Miller, R. L. & Griggs, D. S. Age and visual search: expanding
genes encoding rhomboids, which we have named RBD1
the useful ﬁeld of view. J. Optic. Soc. Am. 5, 2210–2219 (1988).
(YGR101w) and RBD2 (YPL246c). Deletion of RBD1 (Drbd1)
18. Myers, R. S., Ball, K. K., Kalina, T. D., Roth, D. L. & Goode, K. T. Relation of useful ﬁeld of view and
caused cells to grow slowly (Fig. 1a), whereas deletion of RBD2
other screening tests to on-road driving performance. Percept. Mot. Skills 91, 279–290 (2000).
19. Ball, K. K., Owsley, C., Sloane, M., Roenker, D. & Bruni, J. Visual attention problems as a predictor of
had no obvious phenotype and did not enhance the phenotype of
vehicle crashes in older drivers. Invest. Ophthalmol. 34, 3110–3123 (1993).
rbd1D cells (data not shown). The rbd1D phenotype was rescued by
20. Raymond, J. E., Shapiro, K. L. & Arnell, K. M. Temporary suppression of visual processing in an RSVP
plasmid-borne expression of RBD1 (Fig. 1a), conﬁrming that the
task: an attentional blink? J. Exp. Psychol. Human Percept. Perform. 18, 849–860 (1992).
slow growth was indeed caused by loss of RBD1.
21. Chun, M. M. & Potter, M. C. A two-stage model for multiple target detection in rapid serial visual
presentation. J. Exp. Psychol. Human Percept. Perform. 21, 109–127 (1995).
Although previously unnoticed, the Rbd1p sequence contains a
22. Broadbent, D. & Broadbent, M. From detection to identiﬁcation: response to multiple targets in rapid
signature motif for mitochondrial localization, as predicted by
serial visual presentation. Percept. Psychophys. 42, 105–113 (1987).
MitoProt16 and other algorithms. To test this prediction, we used
23. Arnell, K. & Jolicoeur, P. The attentional blink across stimulus modalities: evidence for central
homologous recombination to replace the wild-type gene with a
processing limitation. J. Exp. Psychol. Human Percept. Perform. 25, 630–648 (1999).
24. Potter, M. C., Chun, M. M., Banks, B. S. & Muckenhoupt, M. Two attentional deﬁcits in serial target
fully functional gene fused at its carboxy terminus to green
NATURE | VOL 423 | 29 MAY 2003 | www.nature.com/nature
© 2003 Nature Publishing Group
537