Lecture 2: Feature Integration Theory and Guide Search Flashcards
What is automaticity?
A series of actions that follow after one another without any conscious input
Explain Treisman’s Feature Integration Theory. How do the two types of search differ? Criticisms?
● Treisman et al. (1980) put forward FIT to understand visual search. It has 2 stages:
1. Preattentive stage: basic features are processed automatically and in parallel
before the attention stage.
2. Attention stage: follows the feature stage, combining features in a slow, serial
process to form objects. Features incl. colour, orientation, shape + size. This takes longer, because serial searches are more time consuming than parallel. (called conjunction searches). According to wiki, conjugation search is when there are multiple visual properties shared with at least one kind of distractor present
Treisman’s Feature Integration Theory assumptions include:
○ There is an important distinction between the features of objects e.g. colour,
size, and the objects themselves.
○ There is a rapid parallel process in which the visual features of objects in the
environment are processed together, this doesn’t depend on attention.
○ There is then a serial process in which features are combined to form objects.
○ The serial process is slower than the initial parallel process, especially when
the set size is large.
○ Features can be combined by focused attention to the location of the object.
○ Feature combination can be influenced by stored knowledge.
○ In the absence of focused attention or relevant stored knowledge, features
from different objects will be combined randomly, producing “illusory
conjunctions”.
Treisman conducted a study, where it was predicted that focused attention would be
needed to detect conjunctive letter (a green letter T) because it was defined by a combination of features, but would not be required to detect singlefeature targets (just an S). Results were as expected.
○ Conjunction searches are slow
○ Feature search much faster texture segregation makes it easier.
○ Attention combines features together.
Treisman (1993) expanded FIT into a more complex theory, where there are 4 kinds of attentional selection:
- Selection by location: broad or narrow attentional window.
- Selection by features: features include surface defining features (e.g. colour)
and shape defining features (e.g. size). - Selection by objectdefined locations:
- Selection by object file: this occurs late in processing, and determines the
object file controlling the individual’s response. An object file is information
stored about a partially consciously perceived object.
There are some problems with FIT:
○ It asserts that unique target = parallel search, however in feature search,
distractors are generally the same, but conjunction search can have different kinds. Thus not clear if this parallel search is because of the uniqueness of the target or uniformity in distractors.
Explain Wolfe’s guided search model, and how parallel and serial search are used.
Wolfe proposed the theory must be modified, that they are not purely feature
searches. He conducted a metaanalysis and tried to overlay results and discovered
a continuum; a range of responses, never a perfect parallel.
● Thus, Wolfe proposed a Guided Search model:
○ He replaced Treisman’s assumption that initial feature processing is parallel and subsequent processing is serial with the assumption that it is implausible to think that the search tasks can be neatly classified as serial or parallel, as findings are not based on very fast and very slow target detection times, but many findings fall between these two extremes.
○ The assumption here is that initial processing of basic features produces an activation map preattentively and in parallel, with every item in the visual display having its own level of activation, depending on what the target is. This is why some search times are longer when some distractors share similar features with the targets. This activation map theory provides a plausible way in which visual search can be made more efficient by ignoring stimuli not sharing any features with the target.
○ If parallel search still doesn’t find the target feature maps are used in combination with a guided serial search.
Eysenck & Keane:
● Red circle popped out among blue circles:
● Red circle popped out among blue circles:
○ Positive trial was largely independent of items, negative search took roughly linear time
Eysenck & Keane:
● Red circle among red squares
● Red circle among red squares
○ Negative trial took twice as long on average than a positive trial (have to look
through half)
What effect does evaluative pressure have on attention?
● Evaluative pressure reduces distraction by OUTSIDE stimuli (eg, car horn in the distance), but increases distraction from INTERNAL stimuli (eg, distractors in the test)
