Neural quantum controversy in sensory psychology.

processes underlying sensory discrimination. Simply stated, the question is whether, under a particular set of conditions, sensation is changed in a continuous or a discrete manner as the corresponding physical stimulus is continuously altered. Evidence on the issue is typically obtained in experiments on differential thresholds. Stevens (1) contends that "some 140" functions have been obtained in vision and hearing which show steplike discontinuities when the percentage of correct responses is plotted against the size of the stimulus difference; purportedly, the functions support the theory of the neural quantum in sensory discrimination (2). Unfortunately, as I will point out, Stevens' conclusion is at best equivocal, and the question remains unsettled on both technical and empirical grounds. Statistical tests. Given the assumptions of the neural quantum (NQ) theory and a two-quantum criterion of judgment, the theory predicts: (i) a linear relationship between the percentage of detections and the size of increment added to a standard stimulus; (ii) a two-to-one ratio between the value of the smallest increment which is always detected and the largest increment which is never detected; and (iii) a slope characterizing the poikilitic (psychometric) function which is inversely proportional to the intercept of the function on the x-axis. Of the three predictions, most experimenters have evaluated the linearity of the psychometric function and the two-to-one ratio. It is surprising, therefore, to find that in Stevens' article "the many NQ functions . . . were all fitted by eye" (1, p. 752) with no attempt at a statistical analysis. This