1. 2nd Year Practicals November 2008
Dr Jonathan Stirk
Room C44
Office Hours: Wednesdays 10-11am
Demonstrator: Maria Ktori
Contact by
Room: A24
Office hour: Mondays 2pm

2. Selective Attention & the Flanker Compatibility Effect (FCE)
Structure of practical
5 week structure
Week 1
Mini-lecture, example exp’ts, literature search
Week 2
Develop hypothesis, select project
Week 3
Pilot study, collect data
Week 4
Data analysis (Mini-lecture)
Week 5
Presentations
Week 6
Hand in written report (Deadline Monday 8th December, 2008 by 4pm.)

3. Aims of this practical To learn about the flanker compatibility effect
To design an experiment to test a specific hypothesis about flanker effects
To learn to implement a design using E-Prime software
To learn to collect and analyze data using computer software (E-Prime, SPSS)

4. What is attention?
‘Attention is the process of concentrating on specific features of the environment, or on certain thoughts or activities. This focusing on specific features of the environment usually leads to the exclusion of other features of the environment ‘.
Colman (2001)

5. What is selective attention?
2 main types of attentional tasks
Divided attention tasks (dual tasks)
Paying attention equally to more than one thing
E.g. Reading out loud a story , whilst writing down dictated words (Spelke, Hurst & Neisser, 1976), driving whilst listening for a specific news item on the radio.
Selective attention tasks
Paying attention to one source of information whilst ignoring everything else
E.g. Identifying words presented to the left ear, whilst ignoring words presented to the right ear in a dichotic listening task (Cherry, 1953)

6. Models of selective attention
Where within the flow of information does specific information become selected and other information dismissed? i.e. When does selection take place?
Does selection occur early in processing or later on?
Sensory Store
Further processing
Response
STIMULI
Sensory Store
Further processing
Response
STIMULI

7. Early versus late models of selective attention
Early-selection models assume that selection occurs early-on in processing [after analysis of physical characteristics/features e.g. Broadbent (1958)]. From this point on unattended information receives little or no further processing.
So NO semantic (identification) processing of the ignored/unattended information.

8. Early versus late models of selective attention
Late-selection models propose that ALL stimuli are analysed up to the point of identification (to a semantic level) and selection occurs after this point, i.e. later on in the processing stream.
So to-be-ignored stimuli receive considerable processing and selection occurs much closer to the response end.

9. Early and late selection
All messages in
Physical characteristics
Meaning
Selected message
Selected message

10. BIG questions! Some questions in attentional research are:
“To what extent are irrelevant stimuli processed in selective visual attention tasks?”
“How can we explain what is and isn’t selected?”

11. How can we examine the extent to which irrelevant information is processed?
Priming studies
Do to-be-ignored stimuli prime future performance on a cognitive task?
Flanker tasks
Do surrounding irrelevant stimuli affect performance on target stimuli?
Eriksen & Eriksen (1974): classic flanker effect
A response competition paradigm (similar to Stroop!)
This is a selective visual attention task
It can also be used to examine ‘automatic’ processing of stimuli (processing without attention)
Or… Capture of attention by irrelevant stimuli

12. Eriksen & Hoffman (1973)
Original exp’t used circular displays of letters and S’s had to identify the presence of a target (out of 4 possible targets) flanked by distracters H U M H U A A M M U U H

13. The flanker compatibility effect
Flankers are stimuli which are presented spatially close to target stimuli and which should be ignored
Despite the irrelevance of flankers to the target task they are often shown to interfere with target responses
The original task involved being presented with 5 letter strings and determining the identity of the middle letter by moving a lever to the left or right
More modern versions involve left and right hands pressing specific buttons/keys to identify a target

14. Eriksen et al (1974): linear display task
LEFT HAND RESPONSE
RIGHT HAND RESPONSE
REVERSE MAPPINGS CAN BE USED TOO!
Target: H K S C
flankers
flankers
H H K H H
E.g.
Respond left
S S C S S
E.g.
Respond right
target

15. Compatibility of responses
However, the compatibility of the target and flanker responses is important
RT to target:
Incompatible trials > Compatible trials
Stimuli
Compatibility
Response hand
Target
Flanker
HHKHH
Compatible L K H
KKHKK
SSCSS R C S
CCSCC
SSKSS
Incompatible
CCKCC
CCHCC
SSHSS
HHCHH
KKCKK
HHSHH
KKSKK

16. Defining the flanker compatibility effect
The FCE is the difference in RT between the two types of compatibility trials
FCE = Incompatible trials – compatible trials
E.g. 500 ms-420 ms FCE of 80ms
Sometimes the effect is measured with respect to a base-line condition
One in which flankers are Neutral with respect to target responses
E.g XXSXX (where the X flanker does not belong to the target set)
RT differences can then be framed as “costs” or “benefits”
i.e. we can examine facilitation and interference

17. What factors moderate the FCE?
Research has shown that the FCE is quite robust
However, a number of factors have been shown to moderate the effect
Early research suggested that flanker-target distance was important
Eriksen & Eriksen (1974) showed that larger spatial separation (eccentricity) reduced the FCE
Distracters within 1° of visual angle could not be ignored
Possible evidence for a ‘fixed-width spotlight’ of selective attention (Posner, 1980)

18. Fixed-width spotlight metaphor
Flankers cannot be ignored as they are within the space selected for attention
S C S
< 1 deg
Fixed width (2 deg)

19. Fixed-width spotlight metaphor
Flankers may now receive less processing
> 1 deg
S C S
Fixed width (2 deg)

20. Explanations of separation effects
The spotlight metaphor helps to explain the effects of target-flanker separation on the FCE.
However, other explanations are viable
Visual acuity decreases the further objects are from the point of fixation
So perhaps increasing the size of flankers/targets is important in controlling for acuity problems
Distance is confounded by Gestalt grouping
The law of proximity suggests that closeness effects grouping of stimuli

21. Law of proximity
Grouped by column
Grouped by row

22. So does perceptual grouping affect the FCE?
What if attention is to objects rather than space?
If attention is object-based then principles of grouping may affect what is selected for further processing
Driver & Baylis (1989) used ‘common motion’ to compete the ‘distance’ vs ‘grouping’ hypotheses

23. Driver & Baylis (1989) H X T T
The results showed that moving distant distracters (e.g. the H’s above) produced more interference than the static closer distracters (e.g. the T’s above).
So, perceptual grouping seems important in the allocation of attention and in the FCE

24. Further effects of grouping
Harms & Bundesen (1983)
Used colour segregation of targets/distracters
E.g. (1) F T F versus (2) F T F
This encouraged colour segregation of targets/distracters in condition 2
Smaller flanker compatibility effects in condition 2

25. Further factors moderating FCE
Miller (1991) manipulated five factors to try and eliminate the FCE and determine any boundary conditions
Poor spatial resolution
Inability to hold attentional focus on a fixed location
Inability to focus completely on an empty display location
Inability to filter out stimuli which onset at the same time as the target during the task
Inability to prevent analysis of all stimuli when there is insufficient demand by the attended items

26. Consistent & varied mapping
Miller hypothesised that we are unable to maintain attention on a fixed location and this may be why attention leaks to the irrelevant distracters
In the linear task the target is always in the same spatial location
So, he varied the locations of targets/distracters and used a __ (bar) pre-cue to direct attention to the location
The FCE was NOT diminished when varied mapping was used

27. +

28. H X

29. Miller’s Boundary Conditions
Perhaps it is not the constancy but rather the emptiness of the attended location which prevents early selection from fully excluding other locations from further processing
Necessary object hypothesis
Miller used an RSVP version of the flanker task to test this

30. RSVP task
200ms
F = flanker T = target
200ms 5
200ms 4
200ms 3 2 1
The necessary object hypothesis predicts an FCE only when the target appears in frame 1 (as there is no previous object in the target location)
However, results showed that the FCE was present in later frames refuting the hypothesis

31. Miller’s Boundary Conditions
Maybe we can’t filter out flankers because they onset at the same time as the target
Yantis & Jonides (1984) had shown that abrupt onsets attract attention in a visual task
Miller varied onset/offset transients of flankers/targets
Used ‘figure 8’ concept.

32. Yantis & Jonides’ ‘figure 8’
Results showed that transients had no effect on the FCE
Transients therefore do not seem to be responsible for the partial leakage of unattended stimuli through an early selection mechanism.

33. Miller’s Boundary Conditions
What if processing of the irrelevant flankers is because attentional ‘capacity’ is underloaded leaving room for processing of the flankers?
Perceptual underload hypothesis
So Miller varied the amount of relevant information and examined the FCE

34. Perceptual Underload Stimuli
TARGET (attended)REGION
Number of letters varied
Flankers

35. Perceptual Underload Stimuli
Results showed that the FCE was eliminated for the larger set sizes
Finally a boundary condition for FCE?
NO as there was a confound of timing
Further experiments did NOT support the underload hypothesis

36. So what are you going to do?
Get into small groups (3) and design an experiment to investigate a factor which may effect the FCE
Design needs to be at least a 2 x 2 factorial design
E.g. 2 IV’s!
1. Compatibility of flankers (compatible vs. incompatible)
2. Other variable of your own!

37. Examples of factors to manipulate
Any grouping factor e.g. Colour segregation
Harms & Bundesen (1983)
Number of flankers?
Nature of flankers?
Pictures vs. words?
Target-flanker separation
E.t.c.

38. So for example… manipulate distance S,C left H,K right (response pairings)
Compatible
Incompatible
Near
S C S
H C H
Far
S C S
H C H
Leads to 4 conditions (cells) in the design, tested within-subjects

39. How are you going to do this?
Using E-Prime to control stimulus display
Create stimuli materials in E-Prime or maybe using Paint or other graphics program (PowerPoint plus Paint)
DEMO OF TEMPLATE (using letter stimuli and manipulating target-flanker DISTANCE)

41. So minimum number of trials is 32
Samples need to be weighted to balance out compatible/incompatible trials

42. What to do - recap
So choose a further IV that you can manipulate at 2 levels
E.g. you may manipulate a grouping factor at 2 levels
You might look at what type of information (e.g. semantic?) can influence target response
Create stimuli for your experiment
Program E-Prime
Run design

43. Types of flanker tasks you can use
Classic Letter flanker task S C S
Colour flanker task * * * (Left- red/white, Right- Blue,green) respond to target colour
Letter-number task 2 A 2 (classify target as either a letter or a number)
Spatial/Arrows flanker task < < < vs. < > <
Semantic classification flanker task
Classify names as male/female
E.g. John Samantha John (incompat) vs.
June Samantha June (compat)
Classify target as large/small etc.
Remember this is essentially a response competition paradigm. If target responses are slowed then it must be because of some flanker processing.

44. Some References
Bindemann, M., Burton, A., & Jenkins, R. (2005). Capacity limits for face processing. Cognition, 98(2),
Diedrichsen, J., Ivry, R.B., Cohen, A. & Danziger, S. (2000). Asymmetries in a unilateral flanker task depend on the direction of the response: The role of attentional shift and perceptual grouping. Journal of Experimental Psychology: Human Perception and Performance, 26,
Driver, J. & Baylis, G.C. (1989). Movement and visual attention: the spotlight metaphor breaks down. Journal of Experimental Psychology: Human Perception & Performance, 15(3),
Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of a target letter in a nonsearch task. Perception & Psychophysics, 16,
Eriksen, C.W. (1995). The flankers task and response competition: a useful tool for investigating a variety of cognitive problems. Visual Cognition, 2, (available as a .pdf from me)
Harms, L & Bundesen, C. (1983). Color segregation and selective attention in a nonsearch task. Perception & Psychophysics, 33,

45. Some References
Miller, J. (1991). The flanker compatibility effect as a function of visual angle, attentional focus, visual transients, and perceptual load: a search for boundary conditions. Perception & Psychophysics, 49 (3),
Shomstein, S. & Yantis, S. (2002). Object-based attention: sensory modulation or priority setting? Perception & Psychophysics, 64(1),
Styles, E. (1997). The psychology of attention. UK: Psychology Press [Chapter 3]
Jenkins, R., Lavie, N. & Driver, J. (2003). Ignoring famous faces: category-specific dilution of distractor interference. Perception and Psychophysics, 65(2),
Lachter, J., Forster, K. I., & Ruthruff, E. (2004). Forty-five years after Broadbent (1958): Still no identification without attention. Psychological Review, 111(4),