Gustatory responses of cortical neurons in rats. III. Neural and behavioral measures compared.

The responses of 39 cortical neurons to 13 kinds of taste stimuli including the four putative basic taste solutions (sucrose, NaCl, HCl, and quinine HCl) applied to the anterior portion of the tongue were recorded extracellularly in lightly anesthetized rats. The neural responses were analyzed in terms of the four hypotheses of quality coding: across-neuron response pattern, labeled-line, matrix pattern, and across-region response pattern notions. Animals were given a conditioned taste aversion to one of the 11 stimuli by pairing it with a gastrointestinal illness caused by intraperitoneal injection of 0.15 M LiCl. Behavioral taste profiles were constructed for each stimulus from the suppression of rate of drinking, which indicates the extent of generalization of aversion to each of the four basic taste stimuli. Neural taste profiles of each taste stimulus, which indicate the relation of the taste of a stimulus to each taste of the four basic stimuli, differed more or less depending on the kind of quality-coding notions employed. Among the four analyses, across-region correlation coefficients that were derived from an across-region response-pattern theory showed the highest correlation with the behavioral suppression rates. Therefore we conclude that processing of taste information in the cortex involves differences in both response magnitude across neurons and the spatial localization of those neurons. Fluid intake per day of each of the 12 taste solutions was measured by the single-bottle preference method. When the amount of intake was described in terms of an hedonic index (HI), which indicates the hedonic aspect of the taste of each solution, HI's for sucrose, NaCl, HCl, and quinine were 1.17, 0.43, -0.49, and -0.89, respectively. These values represent the degree of deviation of solution intake above (i.e., preferable) or below (aversive) the standard water intake. Then, HI's were calculated for each of the 12 taste stimuli based on the neural taste profiles and actual HI's for each of the four basic taste stimuli. The correlations between the calculated and the actual (or behaviorally obtained) HI's were very high (ranging from 0.832 to 0.941). This result suggests that the hedonic dimension of taste can be matched well by any one of the four proposed hypotheses.