I'll just mention what I wasn't sure about on the plots, and might be something other people ask. So the vertical axis of "Gabor/D4OG feature" length is on a log scale. So if I see a pixel of the same colour at one height in the graph and the same colour lower down on the graph, what does that mean? Is it that the pixel describes a sample point at some precise value for the feature length, or is it (an approximation to) the total of the responses for all the feature period values that fall within that pixel. Additionally as the feature length increases its energy will generally increase just because there's more points, unless it's normalised in some way?
Basically, whether a given colour means the same power wherever it is in the plot?