An Essay on Attention in Video Games

Voila- my masterpiece:

Video Games: A Portal to Attention

Video games are becoming more and more popular as they become more advanced. They advance in many respects, primarily in graphics quality and game-play, but another pivotal point, one that is often overlooked is how they help with advancement, especially with attentional growth. Video games allow growth in several vital sections. Video game players (VGPs) are more capable at focusing on multiple objects, attending peripheral objects and concentrating on a central object while consciously attending multiple objects, both peripheral and central. Additionally, they are capable of having increased ability in switching tasks under certain circumstances.

In 2006 Green and Bavelier ran a series of three different experiments that indicated how playing video games and not playing them changed the way the participants attended to various stimuli. The first test was a perceptual load experiment, one which tested the natural attention to peripheral objects. The purpose of the test was to determine the differences in attentional load between the VGPs and the non video game players (NVGPs).For the test Green and Bavlier had a display of six potential locations where the target stimulus would appear. For low load tests they would display only the target stimulus and a distractor either centrally located or peripherally. The participants were specifically told not to look at the distractor stimulus. They then had to determine which of two possible target stimuli had appeared. In some cases the distractor stimulus would be the same as the target stimulus and in other cases it would be the other potential target stimulus. In the high load tests all six potential locations for the target stimulus to appear were filled with stimuli, however, these stimuli did not match either of the two potential target stimuli. Again, the distractor stimulus could either match the target stimulus or not as well as being either centrally located or peripherally placed. Green and Bavelier had sixteen participants, all men with normal or corrected vision. They were placed into the two categories, VGP and NVGP based on a questionnaire. VGPs had to play action video games at least several times a week, preferably daily for the past six months. NVGPs had played little, preferably no action games in the past six months. The results showed that with both the centrally located and peripherally located distractor stimuli, VGPs had a higher difference in RT between compatible and incompatible stimuli (compatibility effect) as compared to the NVGPs. Additionally the results showed that the VGPs had faster reaction times in all scenarios than the NVGPs. NVGPs presented a decrease in in the size of the compatibility effect for both centrally located and peripheral distractors, whereas VGPs showed a decrease in compatibility effect only for the peripheral distractors. This shows that the VGPs have an increased natural ability to unconsciously attend to objects other than the target. This conclusion was reached because the VGPs were more affected by the presence of both a compatible and incompatible distractor stimulus than the NVGPs. It could be surmised that the VGPs had a greater spatial distribution of attention thus causing this difference, however, it was shown that both the NVGPs and the VGPs had similar spatial distributions of attention.

Green and Bavelier followed up with another experiment with another set of sixteen participants, again equally distributed between VGPs and NVGPs. The test, a Useful Field of View (UFOV) looked at attention to twenty-four locations, described by the intersection of eight radial lines extending from the center of the visual field and three concentric circles with radii of 10°, 20° and 30°(neither were shown). The VGPs participating in the test reported that most video games are within 10°, the outer edges of attention required are at 20° and that 30° was way out of normal conscious attention when playing a video game. The focal point of the field of view was a 4°x4° outline of a square. Stimuli were rapidly presented (a filled in triangle within a outline of circle in an area of 3°x3°) at one of twenty-four possible locations (mentioned above). After the stimulus had been presented (6.7 ms at 10° and 13.4 ms at 20° and 30°) a mask was shown for 750 ms. The mask completely covered all possible locations for the target stimuli as well as being randomly generated each time so no localizations could be made between target locations and the mask. After the mask the potential target locations were numbered and shown with the spokes and circles visualized. The participants then had as much time as they need to respond to which spoke they believed the stimulus to be on. A previous test, Ball et al. (1988) had shown that when a participant knew the spoke where the stimulus appeared, there was a 90% chance that they would know the eccentricity as well. Therefore, for this test participants were not required to enter the eccentricity of the stimulus. Each target location was tested in the first round followed by a round with twenty-three distractors. This round had a distractor stimulus (4°x4° outline of a square) at each of the other twenty-three potential target locations. There was a third round in which there were distractors at all of the other potential target locations as well as halfway in between every possible target location. This resulted in forty-seven distractors. This was followed by a center-shape discrimination task in conjunction with the three tests before. In the center of the visual field, either a isosceles triangle or a diamond would be shown at the same time as the target stimulus. The participants then had to identify the shown shape as well as the spoke on which the target stimulus appeared. These tests were also done with twenty-three and forty-seven distractors. The difference between twenty-three and forty-seven distractors both in this experiment and others has been shown to be negligible so the data from the two were collected into a 'distractor present' set of data. The results again support the superior attentional capabilities of VGPs. With no distractor, VGPs averaged about 80% accuracy, doing better with targets located 20° out. NVGPs averaged closer to 33%, also doing better when the target was in the middle circle. With distractors present the VGPs maintained about 80% accuracy except for 30° when their accuracy fell to 60%. NVGPs remained at their lower status of about 33% accuracy. As predicted by the previous experiment by Green and Bavelier the central task inhibited the NVGPs significantly without disrupting VGPs accuracy. NVGPs went from 33% accuracy to 25% accuracy. VGPs went from an overall average of 77% accuracy to 76%. This experiment conclusively supports the fact that VGPs have more attentional resources available than NVGPs.

The third experiment was to show causality. It is possible that VGPs started playing video games because they were good at it- they had naturally better attentional resources as compared to NVGPs. Additionally another possible point of causality for the previous results could be the fact that the VGPs are used to using visuomotor skills. Thus it has to be shown that those who do not play an action video game but still play one are worse than those who do play action video games. For the experiment, Green and Bavelier found thirty-two NVGPs approximately half male and half female. They were randomly divided into two groups, control and experimental. The first step in this test was to establish a base line using the procedures in their second experiment. They made small changes based on participant response from the first time they ran it. They simplified the mask to a white noise mask, gave an equal amount of time for the targets to be shown at all three eccentricities and made the central task harder. Over the course of the next thirty days the control group played Tetris everyday (for a maximum of two hours everyday, minimum of five hours every week, maximum of eight hours every week and a sum of thirty hours after the month). The block preview option was turned off so the participants only had to attend to the falling block while still using visuomotor skills.. The experimental group played Unreal Tournament 2004 (under the same time restrictions as the control group) a first person shooter. The game had a simple user interface but still required attending to multiple objects including several peripheral objects. As the players improved in the action game the difficulty was raised. Finally at the end of the experimental period the difficulty was returned to starting difficulty to show a quantitative skill difference in the participants. At the end both groups showed signs of about equal progress in their respective games. After the end of their 'training' both groups were tested using the modified second experiment again. For the tests, both with and without a central discriminatory task, without distractors, the experimental group's pre-test and the control group's pre- and post-test scores were all similar- approximately 85%, 80% and 75% for 10°, 20° and 30° of eccentricity respectively. The experimental group's post-test scores increased to 95%, 95% and 85% on the same scale. With distractors present, again the experimental group's pre-test and the control groups pre- and post-test scores were very close to each other- 60%, 40% and 30%, again relative to the same scale of eccentricity. The post-test for the experimental group however showed a significant increase with the distractors present to have a score series of about 80%, 70%, and 60%. The results clearly show and increase in visual attention and attentional resources available to the experimental group. Additionally, it has been shown that their increased attentional capacity has expanded to outside the normal range of video games (30°) as well as the fact that their central attentional capacity was not diminished as shown by their steady results even with a central discrimination task present.

Looking at all three experiments Green and Bavelier have shown a clear link between video game and improved attentional capacity, showing itself in the VGPs' ability to react more accurately to peripheral stimuli while still managing to keep track of a central discrimination task.
More recently Bavelier, along with Dye (2009), examined the development of visual attention in school age children while looking at the value of video gaming. Over the course of five years one hundred fourteen children and forty-seven adults were recruited and tested. They were divided into VGP and NVGP groups using similar qualifiers as Bavelier had used in the 2006 study with Green. Additionally, they were broken up into four categories based on age, 7-10, 11-13, 14-17, 18-22 (elementary, middle and high school age followed by college age). The first training task was a central discrimination task similar to that of the UFOV of the 2006 study. It was then followed up with another training task, a no-distractor UFOV test, continuing to test with the central discrimination task. Finally, they tested all participants with a full UFOV, 23 distractors along with the central discrimination task. For the two training tasks, the data showed that age and amount of video games played had no effect on the training task performance. However, on the main task, the time for which the target was displayed was shortened until the participants performed at a specified threshold- 79.3%. In the 7-10 category, both VGPs and NVGPs had about equal necessary time, however, the VGPs had a slight edge. In the older ages the VGPs had a significant edge, only needing a mean of about 50ms as compared to the NVGPs requirement of approximately 70ms. During the procedures of the study it was shown that the VGPs did not sacrifice central attentional capacity for their increased peripheral attentional capabilities. This study also showed that playing video games affected the attentional resources of children as well as those of adults, validating the belief that video games help improve developmental attention.

The second test used the same pool as the previous one, however seven of the previous participants were not able to partake. The test consisted of a rapid serial visual presentation (RSVP), in which a outline of a shape would be presented in a 10°x10° visual field. There was one target (T1) in the baseline task and two targets (T1 and T2) in the main attentional task. For the baseline, between one and seven shapes would be shown, each displayed for 40ms with a 66ms break between them, followed by T1. T1 could either be a leftward facing red isosceles triangle or a rightward facing red isosceles triangle. After T1, at the same pace between three and six more shapes would be shown. At the end of the test trial, the participant had to determine which direction the isosceles triangle was facing. For the main attentional task, one to seven shapes were shown, then T1, followed by 1, 2, 4, 6, 8, 10 or 12 shapes (the first trial lagged by one, the second by two etc...) the T2 appeared, a blue isosceles triangle point either up or down, and then the trial finished with between three and six more objects. Once all of the objects had been displayed, the participants then determined which way both of the target objects were facing. After seeing T1, there was an 'attentional blink' during which attention could not be focused properly on the shapes. Bavelier and Dye calculated the mean recovery time for each age group of NVGPs and VGPs. The recovery time was determined by finding out the mean time until the participant could determine the direction of T2 80% given accuracy on T1. The younger NVGPs (7-13) need about 450ms while the older NVGPs (14-22) needed between 350ms and 400ms. The VGPs, through all ages only need about 300ms. Clearly, at the younger ages, the attentional resources and attentional blink is considerably lower among VGPs, which would indicate and increase in development of attention.

The final test continued to use the same pool for a multiple object tracking (MOT) test. Sixteen faces (each of 0.4°) were placed on a screen of a 10° circle. At the beginning of the test, one of the faces was a blue sad face, the rest were yellow happy faces. Once the trial was initiated the faces all began to move randomly, bouncing off each other and walls. After two seconds the sad face turned into a yellow happy face. After another five seconds of random movement of all the faces, a question mark appeared over one of the faces. It was programmed so that it would appear over what had been a sad face 50% of the time. For every three correct answers in a row an additional blue sad face was replaced one of the yellow happy faces at the beginning of the next trial to a maximum of eight blue sad faces. If the participant got the answer wrong, one sad face was removed for the next trial. After eight one of the starting faces switched forms eight times (either from blue to yellow or yellow to blue) or seventy-two trials, the test ended. Both groups of 7-10 year-olds had similar scores- with a maximum of 2.9 objects. However, for the three other age groups, VGPs could track between .5 and 1 more objects than the NVGPs of the same ages. The MOT plateaued for the VGPs just under five objects. Additionally there were general improvement for NVGPs as they aged, indicating that MOT is one of the slower functions to develop and that video games do help expedite the rate of growth. However, the plateau may indicate a cap which both VGPs and NVGPs cannot get past. If that is the case, VGPs are fully developed in the area of MOT between the ages of 14 and 17, whereas it would normally take about 25 years (extrapolating from the given graph of NVGPs).
These thee tests help support Bavelier and Green's earlier findings showing that VGPs have better attentional capabilities than NVGPs- at all ages. It is also possible that playing video games accelerates portions of attentional growth to maturity perhaps even 50% faster than in a NVGP.

In a very recent study, Watter and Shedden (2010) examined how proficient VGPs were at switching tasks as compared to NVGPs. Fifty-six people were selected as participants for this test- all were male, with normal or corrected to normal vision and were undergraduates from McMaster University. The test consisted of an A, B, C, 1, 2, or 3 appearing on either a high contrast or low contrast background, either cued or uncued. When it appeared the participant had to identify which had appeared by hitting the appropriate key. The contrast of the stimulus was the value for the difficulty for the stimulus. There was also a difficulty value for the order of the keys, easy was standard left-to-right A then B then C and 1 then 2 then 3. The hard way was B then C then A and 2 then 3 then 1. There were also the variables of whether or not it would be switching from displaying a letter to number or not, and the cue-to-target time interval (CTI) was adjusted either for either 100ms or 1000ms. Finally, the cue would either be red (number), green (letter) or white (neutral). The stimuli were presented in a random fashion, although it was set that no two adjacent stimuli would be the same. Understandably, for all participants, the easy stimuli and easy response pattern led to shorter reactions times than for their harder variants. The VGPs excelled at several tasks especially when long CTIs and informative cues were present. There was additional significant superiority on the part of the VGPs when undergoing a switch from letters to numbers or numbers to letters, but only when the contrast was high (and therefore the task was comparatively easy).

Watter and Shedden had a follow up experiment that also looked at task switching capabilities of VGPs and NVGPs. In this test the stimuli 2, 3, 4, 6, 7, and 9 were displayed and participant had to respond to one of several tasks, which had been cued before the stimulus appeared. They had to identify one of the following- odd vs. even, prime vs. multiple or less vs. more. They responded with index, middle and ring fingers of their left and right hands respectively to each of the potential queries. The participants were informed that no stimulus would match the stimulus previous to it, but the query might be the same as the one before. Finally, there was the possibility of the CTI being either 100ms or 1000ms similar to the previous trial. The results indicate that although the VGPs were faster at responding to all of the stimuli, the proportional differences among their scores when switching tasks was the same as the proportional scores of the NVGPs when they were switching tasks. This disproves any solid link between video games and a constant increase in the ability to more efficiently switch tasks.

Other studies, not intent on the effects of video game playing have noted the value of video games. Rogoff et al. (2005) found in a study on cultural biases of MOT that video game players could attend to more items in a larger visual field, therefore avoiding a attentional bottleneck. Cavangh and Alvarez (2005) also make not of the value of video games. They looked into MOT and found supporting evidence that video games have a strong, positive affect on MOT. Additionally, they mention, real sports also advance the MOT in many people- so the old days of playing sports outside still help with attentional development.

Combining the results from these studies the link between attention and video games is obvious. Playing video games is strongly correlated to having considerably higher reactions, to peripheral objects, multitudinous objects and combinations of the two, without sacrificing much, if any central focus attention. Video games also help with accelerating developmental attention and may be able to advance some aspects of attention to full maturity possibly up to a full decade prior normal maturity. Finally, video games help slightly with the ability to switch between tasks, but only when the differences between the tasks are clear and the tasks are very simple.




Bibliography


CS Green & D Bavelier, (2006) Effect of action video games on the spatial distribution of visuospatial attention. Journal of Experimental Psychology: Human Perception and Performance. 32, pp. 1465-1478.

Matthew W.G. Dye, Daphne Bavelier (2009) Differential development of visual attention skills in school-age children Vision Research doi:10.1016/j.visres.2009.10.010*

Karle, J.W., Watter, S. Shedden, J.M. (2010) Task switching in video game players: Benefits of selective attention but not resistance to proactive interference Acta Physcologica, doi 10.1016/j.actpsy.2009.12.007*

Maricela Correa-Chávez, Barbara Rogoff and Rebeca Mejía Arauz, (2005) Cultural Patterns in Attending to Two Events at Once Wiley InterScience, 76(3), pp. 664-678.

Patrick Cavanagh and George A. Alvarez, (2005) Tracking multiple targets with multifocal attention Trends in Cognitive Sciences , 9 (7), pp. 349-354.


*Article in press. No volume/page numbers available. Found on SCOPUS and available online.