Cascades of response vectors of olfactory receptor neurons in Xenopus laevis tadpoles

Olfactory receptor neurons (ORNs) of Xenopus laevis tadpoles respond to water-born stimuli such as amino acids. Their sensitivity spectra with respect to amino acids have recently been shown to become more selective over ontogenetic stages [Manzini & Schild (2004) J. Gen. Physiol., 123, 99-107]. In this paper, we undertake a theoretical analysis of this data set and determine the correlational relationships among odorant responses represented as binary response vectors. We first show that, on the one hand, the number of 204 ORN classes (out of 283 recorded ORNs) cannot be explained by a random expression pattern of olfactory receptors (ORs). On the other hand, this number does not appear to be reconcilable with the idea that individual ORNs express one type of OR each. The covariance matrix of stimulus responses shows that the responses to some stimuli are correlated to those of others. Furthermore, the response vectors show positive as well as negative correlations among each other. While the positive correlations can partly be explained by the differing response frequencies to the odorants used, the negative ones cannot. Finally, we analyse the similarity among responses using the Hamming distance as a distance measure, the result being that most response vectors differ from others by small Hamming distances. Such vectors are shown to form pattern cascades, possibly reflecting a decreasing number of ORs being expressed over ontogenetic stages.