Feature integration theory
The feature integration theory ( in the original feature integration theory ) by Anne Treisman (1963 ) is a theory that explains human object recognition using visual attention.
The starting point for Treis 's views, the filter theory of attention by Donald Broadbent is from 1958, which states that there is a so-called " filters " are in the brain, which selects very early stimuli due to physical characteristics (early selection, in contrast, see for a better understanding also: theory of the late selection of German & German, 1980) Since both ( Treisman - German & German controversy ) could give no clear answer, followed in 1995 by an alternative of Lavie: not the Selektionsort ( early / late) but the request to the target situation determines the interference.
- 2.1 Visual Search
- 2.2 Texture area separation
- 2.3 Illusory conjunctions
- 2.4 Identification and localization
Division into levels
Preattentive processing
The characteristics of the not yet detected object are first released and just next to each other (float free ). This process is referred to as preattentive stage of processing up and running quickly, automatically and unconsciously. Similar features (eg blue ) form dimensions (eg, color). The dimensions are dependent on each of detectors which are sensitive to the corresponding features. Similar detectors are in turn organized into feature maps. For each dimension, there is therefore a separate feature map. A certain place on these maps corresponds to a specific location in the visual field (retina ) responsive to stimuli that fall on this area of the retina.
Attentive processing
This matching locations of various cards are assigned by directed attention in the next stage of attentiven processing each other, because here the features are assembled into an object, which represents an overall bottom-up processing (see top-down and bottom-up ). A central location map the outputs of the detectors of all feature maps are available at each of the one place where is located at the time the focus of attention. Characterized the characteristics are interrelated. Due to the required directed attention, this process takes longer than the preattentive processing. We focus our attention Treisman compares with a kind of cone of light ( spotlight ): the object is located within the " light ", it may be perceived as a unified whole.
Four paradigms
Treisman used four experimental paradigms used to justify the theory of stages. So can give processing, such as different reaction times references to either preattentive ( short reaction time) or attentive ( long reaction time).
The paradigms are:
- Visual search
- Illusory conjunctions (links)
- Texture area separation
- Identification and localization
The visual search here is the most important.
Visual search
The visual search for a possible target stimulus ( target) takes place in a search display, which contains a variable number of deflectors ( distractors ). The total number of the presented target stimuli and deflector is referred to as display size. From the experimental subject is to decide whether a target stimulus is on display beside the deflectors. The features of the deflector can be in only one dimension, such as color, shape, motion, etc. different from the characteristics of the target stimulus ( single feature search) or in a combination (conjunction ) of several feature dimensions ( feature Conjunction search).
The parallel search is präattentiv (without attention), as the target stimulus in its feature of which stands the deflector and immediately strikes the eye ( pop-out effect). This is only possible if there is only a varying feature ( single feature search). The display size has no affect on the reaction time. The pop-out effect can be superimposed by evolutionionären processes: An angry face in a lot of happy faces pop out; not a happy face in a crowd of angry faces, however.
If the display is searched for a target stimulus, composed of various features of the deflector ( feature Conjunction search), one speaks of a serial search. Directed attention is required because the patient must cover the entire display " scanning " to discover said target stimulus. The search rate is more than 10 ms per item, whereby the response time of the display size is a function (the larger the display, the longer the reaction time).
Depending on whether a target stimulus is present in the display or not, a distinction is made in the serial search between the self - terminating search (self -terminating search) and the exhaustive search ( exhaustive search). In the self - terminating search, the display is only as long searched until the target stimulus was found, which is of course only possible if he is present at all. This happens on average after half of the items has been searched. In the exhaustive search all display items are scanned before it can be determined that no target stimulus is present. Consequently, the ratio between the duration of an exhaustive and the average duration of a self - terminating search 2:1.
Texture area separation
The texture area separation is a parallel process, which requires no attention. For only targets that differ in one feature suitable, ie only features and no conjunctions.
Illusory conjunctions
Are on the second level of feature integration theory as mentioned above, the characteristics of the first stage using directed attention combined to form a coherent object. However, if the attention is not on the specific place where the object is located, focused, there is a danger that the features are incorrectly linked and an illusory conjunction occurs.
Identification and localization
Varies the target in only one feature of the distractors, one can identify this, without knowing exactly at what point it is on the display. That of the characterization work condition identification and localization independently. When the join condition, the target must, however, using directed attention ( ie, on the second level) are located. Only then his identification.
Feature condition: concentration on one feature (color yellow) to find the target stimulus.