Developmental studies of alertness and encoding effects of stimulus repetition

The findings reviewed here suggest a broadened view of the stimulus-familiarization effect as one of a general class of phenomena resulting from repetition-induced changes in the processes of alterness and encoding. Cantor's original observation of slowed responding to familiarized stimuli can be attributed on this account to experimental conditions under which the magnitude of alterness decrement is greater than encoding facilitation. Exactly the opposite result can be obtained by arranging conditions so as to highlight the encoding facilitation effect relative to the alterness decrement effect. For example, faster responding to familiarized stimuli has been found under conditions in which (1) the imperative stimulus is preceded by a neutral warning signal or, (2) a delay is imposed between the familiarization and the test phases. Earlier, it was noted that the experimental technique of repeated stimulus presentation is widely used in the study of infant perception, with little concern for the mechanisms responsible for what is generically referred to as habituation. Posner and Rothbart (1980) made a similar observation and suggested, on the basis of Kraut's (1976) results, that Habituation in the infant can be viewed as a reduction in the flow of information from the recognition pathway [encoding] into the alerting system. Input is still processed along the recognition pathway, but its failure to activate the alerting system reduces the availability of the central processor and hence of nonhabitual responses. We have developed a model of attention that accounts for our findings and embodies Posner and Rothbart's general idea. It suggests that both the alertness and the encoding functions must reach separate threshold levels for a correct recognition response. The time required to get to each threshold level is changed as a function of stimulus repetition. Repetition slows alertness "rise time," but it speeds encoding "rise time." In comparing the responses to a familiarized color versus those to a novel color, the model can be seen in Fig. 1a. Because any response must await an appropriate level of alertness, the familiarized stimulus is responded to more slowly than the novel one, even though encoding rise time to the familiarized stimulus if fast. Either a delay between familiarization and test or a warning signal acts to mitigate the effects of familiarization on alertness. Alertness rise time for the familiarized stimulus is similar to that for the novel stimulus, but encoding rise time for the familiarized stimulus is faster than for the novel.(ABSTRACT TRUNCATED AT 400 WORDS)