Three Views of Association in Concurrent Detection Ratings
Thomas D. Wickens Lynn A. Olzak University of California, Los Angeles, CA 90024
The ultimate goal of psychophysical research is to explain the behavioral responses of observers to multidimensional, real-world stimuli. This straightforward objective is exceedingly difficult to achieve. A full analysis must not only provide a physiologically based model of sensory information processing along each relevant dimension, but also a model of how multiple information sources within and across dimensions are combined or otherwise interact. Finally, the model must separate and account for the higher-level cognitive processes involved in attentional and decisional behaviors.
As we currently understand human perceptual processes, sensory information is processed simultaneously by many systems and subsystems. Along various dimensions, sensory information is first analyzed into a series of fundamental components. In later stages of processing these components are combined within and across senses to create a coherent representation of the world. Our understanding of the initial analysis process is based largely on experiments designed to test whether two stimuli are processed by independent subsystems. Traditional approaches to this question use experiments in which the effects of a secondary stimulus on the perception of a test stimulus are assessed. They include masking, adaptation and subthreshold summation paradigms. A more recent modification of the traditional approach involves comparisons of simple detection and identification performance when the two tasks are performed simultaneously (see chap. 10).
From traditional experiments, we know that at the grossest level, separate sensory systems respond in parallel to stimulation in different modes--air pressure, light, chemicals, and so forth. Within each system, a more refined analysis may take place, with subsystems again acting in parallel to extract different