An action video game modifies visual processing

[Guest guest]

The following may be of some interest:

-------

An action video game modifies visual processing

Maximilian Riesenhuber

Trends in Neurosciences 10.1016/S0166-2236(03)00380-1

Taken from

http://gateways.bmn.com/magazine/article?pii=S0166223603003801

In a recent paper, Green and Daphne Bavelier show that playing

an action video game markedly improved subject performance on a range

of visual skills related to detecting objects in briefly flashed

displays. This is noteworthy as previous studies on perceptual

learning, which have commonly focused on well-controlled and rather

abstract tasks, found little transfer of learning to novel stimuli,

let alone to different tasks. The data suggest that video game

playing modifies visual processing on different levels: some effects

are compatible with increased attentional resources, whereas others

point to changes in preattentive processing.

Will hours of playing 'Where's Waldo?' make striped sweaters jump out

at you on your next trip to the department store? Would it help a

baggage screener to better pick out suspicious objects from cluttered

suitcases? To what extent training on one visual task transfers to

other tasks is the key question in perceptual learning. In fact,

although a host of experiments have shown that subjects improve with

practice on a number of tasks, these same experiments often find that

subtle changes of the experimental paradigm between training and

testing – such as changing the shape, location or orientation of the

stimuli – can have a profound effect on performance [1] (but see Ref.

[2]). Such extreme specificity of learning is not of much use in the

real world, where generalization and transfer from the training

examples to novel scenarios, or even to different tasks, are key.

In an elegantly simple and surprising paper, Green and Bavelier [3]

now have provided evidence that habitual video game players (VGP)

exhibit superior performance relative to non video game players

(NVGP) on a set of benchmark visual tasks that tested the ability to

process cluttered visual scenes and rapid stimulus sequences – skills

likely to be trained by action games, which commonly require players

to identify and track opponents quickly in cluttered displays and to

switch rapidly between different targets. Importantly, Green and

Bavelier demonstrated that this advantage is not a result of self-

selection (i.e. not because subjects with superior visual abilities

tend to prefer playing video games). Subjects with little or no video

gaming experience showed significant improvement on the benchmark

tasks after playing just ten hours of a first-person-shooter video

game, Medal of Honor.

Improved object detection in clutter

What differences between NVGPs and VGPs did Green and Bavelier find,

and how can those differences be interpreted? In one task, subjects

had to detect a briefly flashed and masked target object (a triangle

in a circle) along one of eight radial spokes made up of distractor

objects (squares) emanating from the fixation point. Subjects had to

report the spoke the target stimulus appeared on. VGPs showed large

performance advantages over NVGPs across all distances from the

fixation point that were tested (up to 30° eccentricity). Green and

Bavelier interpret this difference as an enhanced allocation of

spatial attention over the visual field. Previously, Ball et al. [4],

using the same task, argued for a central role of preattentive

mechanisms because target detection was found to be independent of

the number of distractors, suggesting a parallel process.

Interestingly, comparing subject performance with and without

distractors, Ball et al. also found that introducing distractors

decreased the diameter of the central area over which the target

could be reliably detected. This is compatible with observations by

Green and Bavelier in another target detection task, in which both

VGPs and NVGPs appeared to process probe objects in the periphery

better when there were few simultaneously presented distractors (low

clutter) than when there were many (high clutter). This effect might

be related to recent physiological data regarding the behavior of

neurons in monkey inferotemporal cortex (IT), a brain area crucial

for object recognition in the primate [5]. Neurons in IT have big

receptive fields and show tuning to complex stimuli such as hands or

faces. A recent study [6] showed that, in the presence of

simultaneously presented clutter objects, receptive fields of IT

neurons appear to shrink around an object presented at fixation. This

provides a possible mechanism to increase robustness of object

recognition in cluttered scenes by decreasing the region of the

visual field in which distractors can interfere with the

representation of an object at fixation (introducing a second object

into the receptive field of an IT neuron commonly interferes with the

response to the first stimulus [7]). It is interesting to note that

the physiological effect also occurred if the central object was not

the target for action (i.e. presumably was not attended). Although

this physiological mechanism could underlie the effect of clutter on

object processing in the periphery [3], it cannot explain the

additional observation that VGPs in the high-clutter case showed a

superior ability relative to NVGPs to process probe objects presented

close to the fixation point (D. Bavelier, personal communication),

which would require that the IT neurons of VGPs are also less

susceptible to interference caused by multiple objects within their

receptive fields. Clearly, a better understanding of how the visual

system performs object recognition in cluttered scenes is needed to

link these behavioral effects to underlying physiological mechanisms.

Higher 'subitizing' capacity

Green and Bavelier further tested their subjects on an enumeration

task, in which subjects viewed a flashed display containing a varying

number of squares and then had to state how many items the display

contained. What is commonly found is that subjects can reliably and

quickly enumerate up to around four items [8] but that performance

drops off precipitously for five and more items. The former process

is called 'subitizing' and the latter is called 'counting'. Green and

Bavelier found that VGPs were able to subitize more items than NVGPs

(with averages of 4.9 and 3.3 items, respectively), and they

interpreted this result as another sign of an increased attentional

capacity of VGPs. By contrast, current theories posit that subitizing

is a preattentional process (e.g. evidenced by the fact that subjects

were able to subitize whenever attention was not required to identify

the targets to be enumerated [9]) and would explain the advantage of

VGPs with an ability to individuate preattentively a greater number

of objects than NVGPs.

A preattentive account of subitizing is also supported by recent

positron emission tomography (PET) studies [10] that found subitizing

activated foci in the occipital extrastriate cortex (consistent with

a preattentive visual process), whereas counting involved a greater

network of brain regions, including some implicated in spatial

attention.Improved ability to switch between targetsFinally, Green

and Bavelier tested subjects on a variant of an 'attentional blink'

task [11], in which subjects had first to detect a target (a white

letter in a stream of black letters, briefly flashed individually on

a screen), and then to detect whether another target (an 'X')

appeared in the following displays. Normal subjects show an

impairment in X-detection performance for short lags between the

first (black letter) and second (X) targets – the attentional blink —

that slowly disappears with an increasing number of intervening items

between the two targets. VGPs showed the same qualitative effect, but

a significantly lesser impairment than NVGPs that also disappeared

more quickly, suggesting that VGPs showed less interference between

the two tasks than NVGPs did. Such interference might be due to an

interaction of target-related top-down modulations and bottom-up

input in visual processing. For instance, recent physiology

experiments [12,13] have found that neuronal activation in area V4

(an intermediate visual area providing input to IT) can be modulated

by a target cue presented elsewhere in the visual field, and that

there is a 150–300 ms lag between a change of the target cue and a

corresponding change of V4 neuron modulation. Another set of

experiments [14,15] has shown that neurons in V4 and IT can show

response modulations 150–180 ms after stimulus onset, depending on

whether or not the stimulus in the receptive field of a neuron is a

target. If such target-dependent activity modulations are also

triggered by the detection of the first target in the attentional

blink paradigm, they might contribute to the observed impairment in

detection of secondary targets that follows within a certain

interval, because the top-down modulations would not be compatible

with the changed bottom-up input. The observed advantage of VGPs

would then imply a shortened time course of these modulations as a

result of training, suggesting interesting predictions for

physiological experiments....

---------

Carruthers

Wakefield, UK