What Can Rates of Development Tell Us About Underlying Mechanisms?
Richard Held Massachusetts Institute of Technology
Many investigators of visual development have attributed the appearance and/or changes in sensitivity of discrimination performance with age to neuronal developments in higher visual centers, the cortex in particular ( Atkinson, 1984; Bronson, 1974; Held, 1989; Maurer & Lewis, 1979). The attribution of function to cortex is convincing when based on knowledge of the visual nervous system and its development as studied in animal preparations ( Spillmann & Werner, 1990). Such knowledge provides grounds for asserting that many of the processes that account for properties of visual perception must go on in the cortex. For example, the neuronal equipment for analyzing stereoptic stimuli--single cells responsive to binocular disparities--is not found prior to cortex, but arguments of this kind do not necessarily prove that the changes essential for the development of such properties actually occur at the same loci in the cortex or other higher centers. As Banks and Shannon (this volume) and other investigators have argued, some of these developmental changes may be accounted for at least equally well by lower level (retinal or otherwise) increases in the efficiency of signal processing with resulting increases of the information processing capabilities (resolution and discrimination) of higher level systems that may not have changed. Colleagues and I have suggested just such a possibility in an earlier report on the oblique effect, discussed further on ( Gwiazda, Brill, Mohindra, & Held, 1978). How might decisions be made between these alternatives, given the state of current knowledge? I attempt here to answer this question.
Banks and Bennett ( 1988) applied the logic of sequential ideal-observer analysis, which Geisler spelled out in a recent review ( 1989), to the development