This continuing renewal grant for three years is on percep- tion in both hearing and vision. The ability of human listeners to discriminate temporal changes in sounds is important in under- standing speech and other complex auditory signals. Humans often can identify a particular sound (including speech sounds) by its "quality" independent of its intensity. Similarly, in vision, human observers usually show some degree of visual constancy, which is the invariance of certain aspects of object perception despite changes in viewing conditions. Both the auditory and visual tasks are discrimination functions that can be studied by testing how well subjects can detect a particular constant feature of the stimulus (particular sound profile or particular visual size) presented under a variety of conditions (intensities or distances, for example). To clarify the discrimination processes involved, a quantitative model will be developed for the recognition of constant features. This model then will be tested by seeing how well experimental data fit the predicted curves for perception thresholds. The subjects will be tested using a discrimination procedure, rather than the more usual "matching" procedure. The major objective will be to measure the ability of listeners to discriminate temporal variation in sound, and compare the results of these experiments with predictions of the model. The secondary objective will be to develop a discrimination procedure that would be useful in measuring visual constancy. This work is important and novel because it may demonstrate significant parallels between certain aspects of auditory and visual perception. The results of these studies could be useful in comparing adult human perception with that of infants and of animals.
