Is Change Blindness attenuated by domainspecific expertise An expertnovices comparison of change det

1
Is Change Blindness attenuated by
domain-specific expertise? An expert-novices
comparison of change detection in football
images

• Steffen Werner Bjørn Thies

2

• Change blindness typically occurs when either a
  saccade or a blink disrupts the intake of visual
  information, or full mask or mud splashes

3

• Three related factors that appear to be involved
  in successful change detection
• Changes to objects at the centre of interest are
  usually better detected than changes of objects
  of marginal interest
• The observer must attend to the altered aspect of
  the image or display before and after the change
  occurs
• Attributes of the object that are changed must be
  effortfully encoded and perhaps even verbalized

4

• Representation of a view should be a function of
  both the characteristics of the image and the
  individual observer
• Individual characteristics of the observer may
  play a role in determining the centres of
  interest within an image, the selection of
  meaningful elements within an image, which parts
  to attend to, etc.

5

• Domain-specific semantic changes of a scene (i.e.
  changes that alter the meaning or interpretation
  of the scene) would be more easily detected by
  subjects with experience in a particular domain
  (i.e. expert) than by those without experience
  (i.e. novices)
• Main hypothesis football experts would be more
  sensitive to changes in football images that
  alter an aspect relevant to the game situation
  (semantic change) than to unrelated changes
  (non-semantic change)

6
Method

• Participants
• 24 experts24 novices
• Stimulus material
• Action scenes
• Birds eye views of playing formations
• Traffic-related scenes
• Figure1
• Two type of changes
• Semantic changes
• Non-semantic changes

7

• Location of the change was controlled by dividing
  the images into five sections
• Centre, upper-left, lower-left, upper right,
  lower-right
• Experts agreed on 82.4of the cases that images
  designed to include a semantic change in football
  image altered a relevant aspect of the scene
  whereas they rated only1.7 of images with
  non-semantic changes to include such a change
• Novices 61.7 of semantic change as relevant
  and additionally thought 28.0 of non-semantic
  changes to be relevant to the scene
• Traffic pictures both groups showed similar
  agreement

8

• Procedure
• Original image (500mses)?mask (500mses) ?altered
  image (500mses)
• presented for a maximum of 40 sec
• Search time unsuccessful searches

9
Results

• Overall, successfully detected within 40 sec
  78.9of cases unsuccessfully detected 21.1
• Figure2
• Figure3

10
Discussion

• The results for detection times show that experts
  have a distinct performance advantage over
  novices for detecting semantic changes in images
  from their domain of expertise in addition to
  being faster on domain-related images in general
• Analysis of unsuccessful searches, imply that
  experts mainly have a general advantage to
  process visual information from their domain of
  expertise
• Experts encode semantically relevant information
  quicker or more efficiently than novices. This
  allows them to notice semantic changes faster and
  with a higher degree of success than novices

11

• Has implications for broader application of the
  change blindness paradigm to other cognitive
  processes involve the selective acquisition of
  information from visual stimuli
• Used as a diagnostic tool provide a new,
  non-verbal test of attentional biases

12
Visual span in expert chess players evidence
from eye movements

• Eyal M. Reingold, Neil Charness, Marc Pomplun,
  and Dave M. Stampe

13

• Chase and Simon (1973a, 1973b)
• ????????????????,??????????????????,??????????????
  ????????????????????random broad
  configurations?,????????????????????????
• ??????????????,???????????????????????chunks?
• Chase and Simon ?????????????????????????????????

14

• ???????????????????????????? --- perceptual
  advantage is a fundamental component of chess
  skill
• ????chess-related visual pattern? ???visual span,
  ???fixation???fixation??????????
• ??gaze-contingent window technique???visual span
  --- the smallest possible window that does not
  significantly interfere with the participants
  task performance

15
Method

• Visual Span in Flicker Paradigm
• Materials
• Figure1
• 20???
• Chess configuration
• Random configuration maintained the spatial
  configuration of the chess position from which it
  was derived, but destroyed the chess relation
  information
• 1000ms original ? 100ms blank ? 1000ms modified

16

• Procedure
• 16 baseline trials for each condition
• 24 blocks with eight RT measurements in each
  block
• ???blocks??chess?random configurations??conditions
  ??????,????blocks????????condition
• The first block in each condition, window size
  was set to 80 in diameter
• ?????block?RT???gt102 normative?RT??window size
  
• ?????block?RT???lt98 normative?RT?? window size

17

• ???????????1.280,???????????9
• Visual span????????????window size????

18

• Check Detection
• Materials
• Figure2
• Each display contained a Black King in the top
  left or right square and one or two potentially
  checking pieces
• Symbol notation Letter notation
• Procedure
• 384 trials (192in each of the notation
  conditions)
• A trial was terminated as soon as the participant
  made a yes/no response regarding the check status
  by pressing one of the two response buttons

19
Results

• Visual span in the flicker paradigm
• Figure 3
• Consistent with chase and Simons hypothesis, the
  increases in visual span and speed of responding
  that characterize expert performance on trials
  with chess, but not random, configurations
  clearly indicate an encoding advantage
  attributable to chess experience, rather than to
  a general perceptual or memory superiority.
• The results are especially impressive considering
  that the spatial layout was identical for chess
  and random configurations, with only the identity
  of pieces being different

20

• Check detection
• Error rate
• Experts(1.3)ltintermediate(2.6)ltnovices(3.7)
• Symbol(1.9)ltletter(3.7)
• RT
• Experts(861ms)ltintermediate(1087ms)ltnovices(1207ms
  )
• Symbol(1036ms)ltletter (1145ms)
• Figure4

21

• This center-of-gravity effect reflects a large
  disparity between skill groups in the percentage
  of trials without an eye movement
• Percentage of trials without an eye movement
• Expert(15.9)gtintermediate(2.6)gtnovices(1.6)
• Expert tended to make shorter saccades than
  less-skilled players
• Figure5
• Experts made proportionately fewer fixations on
  pieces than did intermediate players and novices

22

• Consistent with Chase and Simons chunking
  hypothesis
• The magnitude of these effects was stronger for
  more familiar symbol notation than for the letter
  notation, demonstrating that the experts
  encoding advantage is related at least in part to
  their chess experience, rather than to a general
  perceptual superiority

23
Discussion

• Experts have advantage in extracting perceptual
  information in an individual fixation
• For check detection, expert extracts the
  necessary interpiece relations from both foveal
  and parafoveal regions. The larger visual span of
  experts in this task results in fewer fixations
  per trial, and a greater proportion of fixations
  between, rather than on, individual pieces.

24

• For flicker paradigm, advanced chess skill
  attenuates change blindness by improving target
  detection in meaningful
• In the flicker paradigm provided powerful
  demonstrations of the effects of familiarity on
  perception

25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)