During the 1970s and 80s, gustatory theory outran data. There was a lively debate over the fundamental structure of the taste system: Were there basic tastes? Gustatory neuron types? Was the afferent signal for a taste confined to one coding channel (labeled-line) or extended across the activity of all neurons (across-fiber pattern)?
That debate fractured and eventually wearied the field, which thankfully embraced a burst of empirical data from molecular biologists, electrophysiologists, and brain imaging scientists who turned their tools to taste. The path for the past 15 years has been one of discovery: How sapid stimuli are recognized, neural circuitry and neurochemistry organized, which areas process taste and its hedonic sequelae in primates, and which are driven by taste in humans.
Cheered by this progress, attention has shifted from theoretical debates to a practical appreciation of function. It was clear that some neurons are more important in representing a particular taste quality, but also clear that the signal could hardly be restricted to one channel. That compromise was enough of a theoretical platform to satisfy the empiricists, and the field thrived.
Yet, it is useful occasionally to return to our origins and ask whether a supposition accepted decades ago has limited us to discover only what was ordained by that supposition, in this case, that the system is composed of a small number of discrete basic tastes.
No one is better qualified to remind us of this than Robert Erickson, an early and consistent skeptic of basic tastes. Here, Erickson represents 45 years of theory and data in opposition to the notion and the degree to which its acceptance has shaped our understanding of taste.
Notwithstanding the clear value of this article, I offer four cautions. First, Erickson mixes three independent concepts: basic tastes, gustatory neuron types, and a labeled-line coding strategy. It is seductive yet facile to conflate them. Basic tastes provide the labels, gustatory neuron types offer the lines, and the labeled-line strategy appears to emerge intuitively. Yet, the existence of basic tastes and neuron types supports neither labeled-line nor patterning theories. Vision has basic (primary) colors, yet uses an across-fiber strategy.
The existence of basic tastes as the targets to be detected, and neuron types as the lines to be labeled, are necessary but not sufficient to support labeled-line theory. Patterning, however, is silent on the issue of neuron types. It makes no distinction between a gustatory dimension spanned by 1,000 neurons, each with a unique sensitivity profile, or by only five neuron types, each with 200 members. Patterning only requires that the information within neuron types be shared. Meaningful interpretations can only derive from a comparison of activity across neurons with different profiles, whether members of discrete types or not.
Second, Erickson dismisses the concept of basic tastes by imposing a definition, derived from other senses, that is recognized as too stringent for taste. Advocates rarely promoted the requirement that basic tastes in proper combination be capable of composing all other tastes, and certainly not that they be the only qualities that can be tasted. Erickson's point that arguments for basic tastes are contaminated with cultural and linguistic biases is well taken. But support extends beyond mere familiarity or the availability of descriptors to include distinct transduction mechanisms and the likely existence of gustatory neuron types (discussed later). Thus, qualities that have survived to be termed basic have an underlying neural machinery to support their status.
Erickson writes that we do not know what basic tastes are (true), yet, we appear unconcerned with that ignorance. Far from unconcerned, we regret that the evidence is so flimsy as only to support the notion that basic tastes are a convenient rubric for organizing studies, yet we are satisfied that recent advances imply that this effort is not misguided, even if it is not resting on the clear definition Erickson and others crave. Perhaps the field is beguiled by using basic tastes as a guide, but the concept has permitted an unprecedented rate of discovery.
Third, Erickson does not review data that have informed these issues with greater sophistication than is presented here. The existence of neuron types has been addressed in studies that transcend the attempt to classify response profiles. Taste cells are statistically separable into clusters, suggesting neuron types. Yet, those within each cluster are not functionally identical, permitting the argument that clusters reflect the scientist's eagerness to impose order on a system void of them. A resolution of whether neuron types exist could not be made on a taste system at rest. It had to be set into motion.
Taste neurons change responsiveness when the animal is subjected to alterations in physiological state (Hajnal et al. Reference Hajnal, Takanouchi and Norgren1999; Jacobs et al. Reference Jacobs, Mark and Scott1988; McCaughey & Scott Reference McCaughey and Scott2000), conditioning paradigms (Chang & Scott Reference Chang and Scott1984), and taste modifiers (Scott & Giza Reference Scott and Giza1990). Taking advantage of this convenient discovery, the question of the existence of taste neuron types could then be recast as follows: As the system changes to accommodate these manipulations, do the neurons within one putative cluster change as a group, in a manner different from, even opposite to, the accommodation made by cells in other clusters? Such common purpose would imply a functionally distinct group, and so would argue for gustatory neuron types. In each case, the answer was affirmative. Gustatory neuron types are likely.
Finally, Erickson overreaches in proposing that all sensory and motor messages are encoded in patterns. There is inevitably spread across receptors in any interaction with the environment. When that distribution is as narrow as the receptor physiology allows, and when mechanisms such as lateral inhibition are employed to tighten it further, it may fairly be represented as a labeled line. Erickson, on Thomas Young's shoulders, offered a profound insight with the dichotomy between topographic and nontopographic modalities. He might continue to embrace that dichotomy here and recognize that certain processes are best served by specificity.
Such, however, is not the case for taste, the breadth of whose neurons virtually requires that the encoded quality be read across them. The discipline of taste was, and remains, largely sympathetic to Erickson's position and grateful for his contributions.
