Geoffrey R. Loftus - US grants

Research is proposed in six areas. The first, a global area, concerns the application of cognitive equivalence techniques to investigation of learning and forgetting. Of particular interest is evaluation in various situations of the variables that produce equal memory performance do so because they produce equivalent memory representations. The second area of research is to develop and test a mathematical model that integrates acquisition of visual information and visible persistence. The third area is a set of experiments on visual information and visible persistence. The fourth area is a set of experiments on effects of visual stimulus degradation on acquisition of information. The fifth area is a set of experiments on the distinction between perceptual and conceptual processing of visual informtion. Finally, the sixth area is a set of experiments on forgetting from short term memory.

A great deal of everyday human activity is spent acquiring visual information from the environment and using this information to guide behavior. The proposed research is designed to investigate the perceptual and cognitive processes by which visual information is acquired and which result in a long-term memory representation of the acquired information. The research concerns perception of, and memory for, relatively complex visual material (mostly natural scenes). Three lines of empirical investigation are proposed. The first of perceptual processing, that is processing that operates on the visual information present in e environment. The second is of conceptual processing, that is processing that is carried out in short-term memory, and eventuates in the picture's long-term memory representation. The third is of priming, that is, the means by which semantic information, provided just before a picture's appearance, influences visual processing of the picture. In many (although not all) of the experiments, the general paradigm is the following. Pictures are presented for varying durations (e.g., from 30 to 500 milliseconds) and memory performance for the pictures is subsequently measured. Memory performance to exposure duration (called a performance curve) represents the temporal course of processing. Comparisons of performance curves obtained under different levels of some variable (e.g., for dim vs bright pictures) permits an assessment of the variable's effect on picture processing. Associated theoretical investigation is proposed. All experiments are intended to guide and test existing theories of perceptual and conceptual processing that have been proposed by the PI and by others. These theories are all formally and quantitatively specified; hence their predictions with respect to the proposed experiments are clear and unambiguous. Data and theory about normal visual processing provide a foundation for isolating causes of, and guiding solutions to abnormal or suboptimal visual processing. Suppose, for example, that an individual complains that he or she "just can't seem to keep up with what's happening in a complex visual environment." What is causing this problem? Is it, for example, simply a slowing of initial information acquisition? Is it a lack of ability to use appropriate contextual information to guide processing? Is it a lack of ability to manipulate acquired visual information in short- term memory? The proposed research will provide the empirical and theoretical technology that would enable a clinician to differentiate among these (and other) possible reasons for the difficulty.

The proposed research concerns perception of, and memory for visual material (ranging from relatively simple stimuli such as digit arrays to more complex stimuli such as natural scenes). The research is framed largely within a theory, the sensory- response/information-acquisition rate theory that PI and his collaborators have been developing for the past three years. Of central importance to the research is methodology involving equivalence techniques. Equivalence techniques which - essentially entail determination of the rules by which different factors combine to produce equivalent internal states - allow conclusions which, unlike those allowed by evaluation of typical statistical interactions, can be generalized both to theoretical constructs (e.g., "strength") and to other dependent variables (e.g., d') that are nonlinearly related to the dependent variable that happens to be measured in a particular experiment (e.g., proportion correct). Five broad research topics are proposed, designed to investigate, (1) visual degradation (e.g., contrast) effects on visual perception and memory, (2) processing of and memory for visual stimuli decomposed into high and low spatial frequencies, (3) unification of two facets of everyday visual behavior-information acquisition and phenomenological appearance, (4) processing as a path through some multidimensional encoding space, (5) integration of information across a series of brief presentations that mimic eye fixations, and (6) the relation between confidence and accuracy in visual memory. Data and theory about normal visual processing provide a foundation for isolating causes of, and guiding solutions to abnormal or suboptimal visual processing. Suppose, for example, that an individual complains that he or she "just can't seem to keep up with what's happening in a complex visual environment." What is causing this problem? Is it, for example, simply a slowing of initial information acquisition? Is it a lack of ability to use appropriate contextual information to guide processing? Is it a lack of ability to manipulate acquired visual information in short- term memory? The proposed research will provide the empirical and theoretical technology that would enable a clinician to differentiate among these (and other) possible reasons for the difficulty.