Search

The identity of sensory stimuli is encoded in the spatio-temporal patterns of responses of the encoding neural population. For stimuli to be discriminated reliably, differences in population responses must be accurately decoded by downstream networks. Several methods to compare patterns of responses have been used by neurophysiologists to characterize the accuracy of the sensory responses studied. Among the most widely used analyses, we note methods based on Euclidean distances or on spike metric distances. Methods based on artificial neural networks and machine learning that recognize and/or classify specific input patterns have also gained popularity. Here, we first compare these three strategies using datasets from three different model systems: the moth olfactory system, the electrosensory system of gymnotids, and leaky-integrate-and-fire (LIF) model responses. We show that the input-weighting procedure inherent to artificial neural networks allows the efficient extraction of information relevant to stimulus discrimination. To combine the convenience of methods such as spike metric distances but leverage the advantages of weighting the inputs, we propose a measure based on geometric distances where each dimension is weighted proportionally to how informative it is. We show that the result of this Weighted Euclidian Distance (WED) analysis performs as well or better than the artificial neural network we tested and outperforms the more traditional spike distance metrics. We applied information theoretic analysis to LIF responses and compared their encoding accuracy with the discrimination accuracy quantified through this WED analysis. We show a high degree of correlation between discrimination accuracy and information content, and that our weighting procedure allowed the efficient use of information present to perform the discrimination task. We argue that our proposed measure provides the flexibility and ease of use sought by neurophysiologists while providing a more powerful way to extract relevant information than more traditional methods. [ABSTRACT FROM AUTHOR]

Most highly derived olfactory systems, such as the insect antennal lobe, discriminate among a wide array of monomolecular odourants and blends of odourants. Given the relatively limited number of neurons used to code these odours, this ability implies that neural representations for odours overlap in a cross-fiber coding scheme. Here we use the generalization of a conditioned feeding response in the sphinx moth, Manduca sexta, to quantify three geometry-based dimensions of odour space in which monomolecular odours may be assessed. In a series of experiments we show that generalization of a conditioned response from one monomolecular odour to another is a function of differences in length and shape of the carbon chain as well as the functional group on the molecule. When moths were conditioned to 2-hexanone or 1-decanol and tested with a number of alcohols and ketones, we found that the generalization of the conditioned response decreased as a function of the chain length and functional group. In contrast, when conditioned to 1-hexanol, moths failed to distinguish alcohols from ketones of the same chain length. In all of these cases, chain length did not interact with functional group, thus indicating the independence of these dimensions. Differential conditioning of alcohols and of alcohols and ketones revealed interaction of excitatory and inhibitory generalization gradients within an odour 'dimension'. When odourants were sufficiently distinct, the peak of the generalization gradient was shifted away from the conditioning odour and in an opposite direction from the unreinforced odour. Altogether, these data substantiate the claim that these molecular characteristics are relevant coding dimensions in the moth olfactory system. These data are consistent with a cross-fiber coding scheme in which odours are coded by spatio-temporally overlapping sets of neurons, both in the periphery and in the antennal lobes.

The proboscis extension response conditioning protocol has been used to explore olfactory-based associative learning in an array of insects. We have monitored a different feeding reflex, which involves activation of the cibarial pump, to demonstrate olfactory learning in the moth Manduca sexta. In the first experiment, four different treatment conditions were used to assess associative (Pavlovian) learning. The results indicate that an excitatory cibarial pump response develops and is retained for at least 24 h only when odor is forward-paired with the presentation of sucrose. Three control treatments, backward pairing, air (no odor) pairing and random pairing, failed to increase the cibarial pump response. However, an excitatory cibarial pump response developed when the backward- and air-paired groups received forward pairing of odor and sucrose on the following day. In contrast, moths experiencing random pairing on day 1 displayed a slower rate of acquisition during forward pairing on day 2, which may indicate inhibition. The second experiment investigated discrimination learning. Two odors were randomly presented, one odor being forward-paired with sucrose (+), the other presented alone (-) in a counterbalanced design. Again, only when odor was forward-paired with sucrose did learning occur. We discuss the implication of these findings for a broader comparative analysis of learning in insects.