Characterizing the Input-Output Function of the Olfactory-Limbic Pathway in the Guinea Pig

Figure 6

Results of neural activity prediction for the KRLS algorithm with linear (cyan) and RBF (green) kernels. (a, b) # spikes recorded by a representative electrode as the LOT normalized stimulus intensity varied from 0 to 1. Blue circles report the frequency with which a specific number of spikes were recorded for a given stimulus intensity (e.g., larger circles correspond to more frequent # spikes). (a) # spikes after the 1st pulse. (b) # spikes after 2nd pulse. (c, d) Distributions of prediction accuracies reported in box plot format (all recording channels, all experiments). Linear (cyan) and RBF (green) approaches are compared with chance levels (red). denotes significance level with respect to chance levels (Kruskal-Wallis ANOVA for ranks, with Bonferroni correction). (c) Accuracy distribution for # spikes after 1st pulse (left), # spikes after the 2nd pulse (middle), and total # spikes (right) for IPI variation. (d) Accuracy distribution for # spikes after 1st pulse (left) # spikes after the 2nd pulse (middle), and total # spikes (right) for stimulus intensity (stim. int.) variation.

(a)

(b)

(c)

(d)