Hi, I have 68 samples of neurons where I am plotting their length distributed across distance from the neuron origin, where distance is binned into tenths. Looking at the shape of these plots, many of the neurons look bimodal or trimodal to me. I want to see if anyone has some feedback for the following method for quantifying multimodality: For each sample neuron, if I connect the length value for each bin to the next bin with a line I get a spatial profile I'll call the 'value line'. So I can measure the area above and beneath the value line and take the ratio between those areas. Beneath the line is obvious. By above the line, I mean wherever there are 2 local peaks with a trough between them (i.e. whenever the slope goes from positive to negative), I would measure between a line connecting those peaks and the lower boundary provided by the value line (i.e. the area of trough). The area would be measured here as a polygon where the number of sides depends on the number of bins between the peaks. Then I'd sum all of these local areas to get the total 'above the value line' area and divide by the area underneath the value line to get a continuous value for multi-modality that I can compare across neurons, Here, larger values mean more multi-modal. I like this idea because it doesn't assume normality. Does this method make sense? Is there a term use for it? Does anyone know of any references that use this method (I don't care if the subject matter relates)? Thanks, kbrownk