Top-down Processing is an important perceptual theory in cognitive psychology. The theory establishes the paradigm that sensory information processing in human cognition, such as perception, recognition, memory, and comprehension, are organized and shaped by our previous experience, expectations, as well as meaningful context (Solso, 1998). For a more detailed definition of top-down processing, check out The University of Alberta's Cognitive Science Dictionary.
An Example of Top-Down Processing:
You may be able to identify the phrase in the background of this web-page without much difficulty at first glance. However, on a second look, you may be surprised to see that the "H" used in "THE" looks the same as the "A" used in "MAN". How was this conversion of letters we are at first quite unaware of possible? The top-down theory explains this by noting that we are assimilating the fonts used according to the context of the phrase. Also, based on previous experiences of the English language, we know that a three-letter word beginning with "M" and ending with "N" is most likely, "MAN". It seems that upon perceiving the stimulus, we automatically eliminate the possibility of accepting the stimulus as merely a nonsense phrase containing the words "THE MHN RHN". This is another testament to the top-down theory, as we inherently strive to understand and make sense of the world around us comprehensively instead of taking for granted the sum of the parts.
Some Historical Experiments Relevant to Top-Down Processing
In 1935, John Ridley Stroop created a task that demonstrated top-down processing at work in word recognition. The classic experiment consisted of two conditions. In the first condition, two lists of words were shown to the participants. One list contained color-words printed in black, and the other list contained color-words printed in incongruent colors ("RED," for instance). The participants were asked to read out loud the word itself independent of the print-color. The result? There was no significant difference in reading reaction time between the black-print words and the incongruent color words.
In the second condition, the participants were asked to read out loud the actual color of the incongruent word stimuli (So if shown "RED," the participant is supposed to say "green"), and the actual color of solid-color squares. The reaction time for naming the colors of the incongruent words took, on average, 74% longer than naming the color for solid-color squares. Even after training the participants to read the color and not the word, the reaction time still remains significantly higher with naming colors in incongruent color-words.
It is believed that this phenomenon occurs because humans are much more trained in word recognition than in color recognition. Thus, we are directly tempted to read the word out loud rather than the color. Top-down processing helps explain the Stroop Effect because this automatic-word-recognition mechanism in cognition is most plausibly explained by our extensive familiarity with reading based on previous experiences.
To do the Stroop experiment and see the effect for yourself, go to:
http://epsych.msstate.edu/deliberate/Stroop/ for a demonstration.
Object-Search in Real-World Scenes
Dr. Irvin Biederman
In 1973, Irvin Biederman (currently Director of the Image Understanding Lab at the University of Southern California), A. L Glass, and E. W. Stacy, found conclusive evidence that the contextual, semantic relationship and the categorical organization of everyday objects directly affect recognition and short-term memory. In this visual-search experiment, participants were briefly presented with a picture of an object to look for in the search task. Then, either a coherent or jumbled version of a scene was presented. In the jumbled version, the target object was located in a contextually unusual place. For example, a fire hydrant might be floating in the air instead of positioned appropriately on the sidewalk.
A Coherent Picture and Its Jumbled Version
Upon presentation of the scene, the participant pressed either a "yes" or "no" key as soon as he judged whether the object was or was not in that scene. Biederman found that if a picture is coherent, the participants performed much faster in reaction time in finding the cued object than in the jumbled version. The presence of contextual relationship in the scenes had a direct effect on reaction time as well as accuracy.
Contextual cues, or the top-down driven schema in everyday scenes, affects processing in cognition. The internal model of expectancy based on experience (to look for a cup on a kitchen countertop, for example) has a profound impact on how we perceive and recognize the world around us.
Stephen Palmer, Eleanor Rosch, and Paul Chase found that there is a privileged, or canonical, perspective for perceiving objects. When we identify an object, there is a preferred canonical perspective that contains the most salient characteristics for recognizing that particular object. This "prototypical" perspective implies the existence of an internal category in memory for mapping the object, or the "necessary precondition for definition of the object and its attributes" (Palmer, Rosch, & Chase, 1981).
In the experiment, participants were shown photos of different objects taken from different perspectives.
Twelve perceptive views of a horse used in Palmer, Rosch, and Chase experiment (1981)
Participants were then asked to identify the objects in the picture. As the pictures increased in canonicalness, the reaction time for identification decreased. Again, top-down processing is involved in organizing memory and directing recognition based on prior knowledge of the object's best conceptual representation.
Here are some optical illusions that illustrate the power of top-down processing.
A Few Illusions in Perception that are Related to Top-Down Processing
Other Top-Down Processing Links on the Web
Top-Down Processes-- A PsyQuest
Top-Down Processing in Auditory Perception
Some Top-Down Negations and the Opposing Theory of Bottom-Up Processing
Artificial Intelligence and What Best Models the Human Brain-- Top-Down or Bottom-Up?
Biederman, I., Glass, A. L., & Stacy E. W. (1973). Searching for objects in real world scenes. Journal of Experimental Psychology, 97, 22-27.
Palmer, S. E., Rosch, E., & Chase, P. (1981).
Canonical perspective and the perception of objects. J. Long & A. Baddely
(Eds.), Attention and performance IX (pp. 135-151). Hillsdale, NJ: L. Erlbaum
Associates.
Solso, Robert L. (1998). Cognitive psychology (5th ed.). Needham Heights, MA: Allyn and Bacon.
Stroop, J. R. (1935). Studies of interference
in serial verbal reactions. Journal of Experimental Psychology, 28, 643-662.