week 4 - visual search Flashcards
role of attention in visual search
searching a cluttered environment for task relevant ifo is difficult
when the target shares one or more features with other things in the image (distractors) it is more difficult
- we know that the colour and shape of vision are processed in different parts of the brain
THIS IS THE BINDING PROBLEM
visual search paradigms
feature search
target has a unique feature that is not shared by other items in the display
therefore it pops out
visual search paradigms
conjunction search
target has no unique feature that is not shared by other items in the display making visual search more difficult
feature integration theory
anne treisman 1988, 1992
illusory conjunctions of features can occur
when focused attention is absent people can experience shapes and objects that are combinations of other shapes and objects they did not pay attention to
feature integration theory
can occur when:
focused attention is absent
relevant stored knowledge is absent
spatial attention is diverted
display is presented in peripheral vision
feature integration theory
not just guessing:
can occur with high confidence
do not occur under all circumstances that decreases performance
illusory conjunctions happen across space not time
feature integration theory
evidence against
FIT argues that an object is only an object if it is attended to
however negative priming tasks (tipper 1985) show semantic (meaning) processing of unattended stimuli
feature integration theory
evidence against
tipper 1985 study
Present someone with an image or word or sound
Call this a ‘prime’
Then present them with another stimulus
Call this a ‘target’ or ‘probe’
You can manipulate what the prime is and how its related to the probe to change how fast people can make decisions about the probe
This is priming conceptually - you can change the way people behave on the basis of something that you showed them or presented to them briefly beforehand
In this task –> negative priming task
We have the probe that the participant looks at
Red outline of a dog blue outline of a spanner
Just before that (few 100ms) participants are presented with one of the 5 primes
Attended repetition - a stimulus which is exactly the same
The participants job is to pay attention to the red thing (tell them to do this) and then when the probe appears you ask them to name the red thing
In condition where the thing you have to pay attention to is repeated you are at your fastest to name that thing
Attended sematic - still faster then control
Dogs and cats on same level (co-occur)
The issue for FIT comes on the other side of the scale
If you look at the ignored repeated trials
Have the slowest reaction times
Slower to identify that a dog is a dog in the probe condition if just before hand you have ben presented with a dog or a cat that you are not supposed to pay attention to (because in the prime conditions they are blue and have been told to pay attention to the red object)
Under FIT should not process that the primes are a cat or dog because you shouldn’t be paying attention to them But because the conditions of a blue cat/dog were slower than a control trial must mean that it Is possible for the cat/dog to have been processed on some level
This tasks shows that you can process unattended objects (combinations of features - colour and shape) without paying attention to them. And they can subsequently influence your behaviour
This doesn’t work with feature integration theory
feature integration theory
strengths
an important contribution to explainng what happens within the attentional spotlight
influenced thinking on a variety of topics from early sensory encoding to later attentional control
feature integration theory
weaknesses
doesnt explain why the similarity of distractors is influenctial
neglect/extinction patients have problems with both conjunction and single feature targets
- doesnt explain how you can process thats to the level of meaning when not paying attention to them
guided search/dual path model
wolfe 1998
in real world search people usually have expectations of where to find certain things
prior knowledge can make things more efficient
ehinger 2009, found that observers fixate at relevant parts of a scene very early on (when down images of places)
guided search/dual path model
how does processing work?
unlike FIT where processing moves from parallel (pre-attentive) to serial (attentive)
wolfe assumes a simultanteous mix of serial and parallel strategies for visual search
early pre attentive processes produce an activation map where each item in the display has its own level of activation
- object with the greatest activation receives attention 1st
combines topdown and bottom-up processing for efficient search
