A neural coding model for sensory intensity discrimination, to be applied to gustation

Abstract

This paper presents a model of the neural coding and discrimination of sensory intensity. The model consists of five stages: (1) the coding of stimulus intensity in peripheral receptors or neurons by a ‘rate’ code. The relevance of comparing different analysis intervals for the response is pointed out; (2) neural processing, according to either ‘labeled-line’ or ‘across-fiber pattern’ theory. In addition, two possible non-linearities in the processing are considered: a threshold mechanism, and ‘contrast enhancement’ by reciprocal inhibition; (3) a neural discriminator, based on signal-detection theory; (4) a memory stage; (5) an effector organ providing a behavioral output. Emphasis is put on stages 2 and 3. The model produces predictions of the differential threshold, which should be directly testable in a behavioral two-alternative forced-choice paradigm. The model will be applied to gustatory intensity discrimination in rat in a subsequent study (Maes and Erickson 1984). The Discussion pays attention to the relative contributions of peripheral and central “noise” sources. It also compares the present model with Beidler's (1958) approach through just noticeable differences (JND's). The model presented here seems more adequate in providing an understanding of sensory information processing.