Rate of information processing in visual perception: Some results and methodological considerations.

Abstract

Inadequacies involved in the methodologies for studying rate of information processing in visual perception were discussed in terms of current models and knowledge of visual perception. A different methodology was introduced based on a 10-channel tachistoscope which permitted the presentation of letter stimuli at varying rates. The task was visual search with .9 required to detect whether an A (target) had occurred in a sequence of letters presented on a trial. Each letter was presented for approximately 2 msec, with the rate between letters varying from 5 msec, to approximately 3 sec. In addition to rate the number of letters presented in a trial (sequence length) and the location of the target in the sequence were studied. The only significant effect was the number of letters in a sequence. Detection sensitivity decreased with increasing sequence length. Control experiments indicated that this change in sensitivity was not attributable to an overloading of the encoding mechanism but rather was attributable to a change in the judgmental task. An increase in the number of noise letters disproportionately increased the number of false alarms relative to the hits. A simple probability model was able to describe the obtained results with fair accuracy. Presentation rate was immaterial. Detection was the same with only 5 msec, between letters in the sequence as with 3 sec. Failure to find a rate effect was interpreted in terms of filtering processes that operate in visual perception prior to the encoding process.The rate at which a human can process procity has been found to hold for temporal information in the visual channel would ap-durations of as much as 100 msec, or more pear at the first level of analysis to depend for perceived brightness, and recent work on the number of discrete percepts that he (Kahneman, 1966; Kahneman & Norman, can experience in a unit of time and the 1964) has suggested critical durations of 300 amount of information that can be processed msec, or longer for form perception. If from a single percept. Limits on the discrete luminance and contour information are rate of visual perceptions are set by the capa-summed over temporal intervals of this durability of the system for temporal resolution, tion, these summation intervals would set Bloch's law dealing with time-intensity reci-limits on the number of noninteracting per-iThis investigation was supported by United ce P ts that could be experienced in a unit of States Public Health Service Research Grant MH-time unless other mechanisms in the visual 1206 and a United States Public Health Service system served to terminate the integration Research Career Program Award K6-MH-22,014 esg with a ch j n stimu l a tion (Schur-2 Requests for reprints should be sent to Charles r _ .

Attention capacity and task difficulty in visual search

“…The method used here for testing capacity limitations was first introduced by Eriksen and Spencer (1969) and Shiffrin and Gardner (1972). It involves assessing the accuracy of judgments about a briefly presented display when its different parts are presented successively (SUCC), as compared to when the entire display is presented simultaneously (SIM).…”

The role of attentional resources in automatic detection

“…As Eriksen and Spencer (1969) and Kinchla (1974) have pointed out, increasing the number of inputs increases the "noise" in the decision process and lowers detection even if all inputs are processed in parallel and independently. A parallel analysis of signals does predict, however, that detection should be independent of the rate at which inputs are presented, assuming that masking, acuity, and other peripheral factors are held constant.…”

Visual discrimination, categorical identification, and categorical rating in brief displays of curved lines: Implications for discrete encoding processes.

“…A poststimulus mask was introduced to further limit or at least attenuate subsequent processing of the fading visual image (Eriksen & Spencer, 1969;Estes & Taylor, 1964Lupker & Massaro, 1979;Morotomi, 1981). The use of a poststimulus mask was not, however, critical to obtaining the type of performance reported below (Foster,Note 1).…”