Attention II Flashcards
What is the binding problem?
(feature binding)
= different features of visual stimuli are processed by different brain circuits
-> nonetheless unified object perceived
-> attention for binding of features
What is Treisman’s Feature Integration Theory (1986)?
(feature binding)
- preattentive stage
-> object is analyzed into features
-> features are independent of each other
(free-floating) - Focused attention stage
-> features are combined (bound) into one object using attention
Object –> preattentive -> focused –> perception
stage attention
stage
What are illusory conjunctions?
(feature binding)
= combinations of features from different objects
illusory conjunctions occur because…
-> in the preattentive stage, features are not yet
associated with a specific object
(free-floating)
-> in the focussed attention stage, features are
correctly reported
(when attention is directed to objects)
What cognitive process avoids illusory conjunctions?
(feature binding)
Top-Down knowledge
-> we know usual feature of most objects
-> aid to feature analysis
How can the Feature Integration Theory be neurally proven?
(feature binding)
Patient
-> Inabillity to focus attention on individual objects
(parietal lobe damage)
-> showed strong illusionary conjunctions, even when viewing stimuli for up to 10 seconds
What is the definition of visual search?
(visual search)
= looking for a target in a display containing distracting elements
Target
-> The goal of a visual search
Distractor
-> in visual search, any stimulus other than the target
set size
-> the number of items in a visual search display
What makes visual searches easy or hard?
(visual search)
Feature search
-> Target is defined by a single feature
-> If target feature is sufficiently salient, it pops out
Parallel search
-> can process several items at once using covert
attention
-> reaction time does not depend on set size
(even with more same colored items the different
colored one sticks out)
Conjunction search
-> Target is defined by two or more features
(requires feature binding)
-> often inefficient (no pop out)
serial search
-> usually requires overt attention
(because we have to look at each item)
-> reaction time depends on set size
(because of serial processing)
What is the efficiency of search measured by?
(visual search)
Search efficiency
-> average reaction time as a function of set size
Measured as search slope (ms/item)
-> small slope: search is efficient
-> larger slope: search is less efficient
How does Guided Search Theory by Jeremy Wolfe work?
(visual search)
- Input stimulus: Find stimuli
- Calculate local salience and store it in feature maps
(bottom-up guidance, stimulus driven) - Feature maps indicate bright white spots where
activation is strongest
(strong local contrast in these image parts) - Top-down weighting of different features
(user-driven) - Priority map created
(sum of feature maps weighted by top-down
importance) - Priority map guides attentional selection of objects
What is Bottom-up guidance about?
(visual search)
= identical elements can have different salience in different parts of the image
(eg. rotation of lines not matching to rest)
=> local differences
What is the deciding factor in how a priority map looks?
(visual search)
Priority map changes as weight of bottom-up and top-down guidance changes.
No specific top-down goals
-> strong weight of local salience
-> will attract attention bottom-up
No bottom-up salience (local contrast)
-> strong weight or top down goals
Can real objects be viewed as complex conjunctions?
(visual search)
In the real world, we rely on more than just feature / conjunction search
->
How can we study visual search in real-world scenes?
(visual search)
By tracking eye movements using…
-> an Eye Tracker
-> Virtual Reality for eye tracking
-> Mobile glasses for eye tracking
-> Augmented Reality
->
How do you study eye movements using an eye tracker?
(visual search)
Eye movement: overt attention
-> fixation: keeping eyes still to gather information from a point of interest
-> Saccades: moving eyes from one point to another
(rapid)
What are bottom-up determinants of eye movements?
(visual search)
Itti & Koch (2001): Salience model
-> Input image is decomposed into basic features
-> Feature maps of local contrast
-> Feature maps are combined into salience map
-> Attention selects point of strongest salience
(winner-takes-all)
-> from then always to next most salient point
(inhibition of return)
-> modulation of all steps by top-down attentional bias and training
