Mousetrap Visualization Questions
I have a slew questions about visualization and mouse-tracking...
Given that these are mostly R questions and not OpenSesame questions, I realize that this might not be exactly the right place to ask it. That said, I did not know exactly where to ask this and I thought that a public forum would be better than a personal email in case anyone else has similar questions .
I suspect that some of my questions will be answered in the following paper (Kieslich et al., 2017. Mouse-and hand-tracking as a window to cognition: A tutorial on implementation, analysis, and visualization), though I don't think it's been published yet. I'm eagerly awaiting the paper.
Question 1: Is there a way to add "confidence intervals" around the plots of MT trajectories? I'm thinking of a shaded band like in pupillometry studies (example is from Mathot, Grainger, Strijkers, 2017).
Question 2: How can we identify and remove individual trajectories? In the bottom right picture below, you can clearly see that there is a single trial than needs to be removed (the one that goes to the left and then all the way around hugging the edge). I know we can export a pdf of the individual trials and then eyeball them but is there a way to do this in R such that they can be "marked." Or do I simply have to match up the trajectory of the pdf with the row of the trial level data and then exclude that row?
Question 3: I tried plotting the individual trajectories for a different experiment but something clearly went wrong. This makes me wonder if the plotted aggregate trajectories are off even though they look normal. I used the exact same sequence of code (shown below) for both the plots above and the plots below, yet it went haywire in this second one. I have no idea why. (Note the two experiments are completely different in design).
MT_DATA <- mt_import_mousetrap(d.CORR, timestamps_label= "timestamps_get_MT", xpos_label = "xpos_get_MT", ypos_label = "ypos_get_MT") MT_DATA <- mt_remap_symmetric(MT_DATA) MT_DATA <- mt_align_start(MT_DATA) MT_DATA <- mt_measures(MT_DATA) MT_DATA <- mt_time_normalize(MT_DATA, save_as="tn_trajectories", nsteps=101) MT_DATA <- mt_sample_entropy(MT_DATA, use="tn_trajectories", save_as="measures", dimension="xpos", m=3) plot_trials <- mt_plot(MT_DATA_FLINT, use="tn_trajectories", points = T, color = "TYPE", facet_col = "SIZE")
Question 4: What is the proper way to analyze/visualize a design with a 4 corner set up like the one below? My first thought in a design like this is to keep the responses in the same location rather than varying them across participants. My second thought is that I wouldn't necessarily want them to be symmetrically re-mapped either. Attraction towards false is (potentially) different than attraction towards uncertain. Then I wonder: will area under the curve or max deviation be accurately calculated if the trajectory initially goes up to "false" but then cuts diagonally and winds up going to "uncertain"? Would these changes be averaged out? Or should I consider a different DV like total_distance?
If I remap the data to be symmetric, it looks like this. (Note the data is only 60 trials of pilot data)
If I do NOT remap the data, it looks like this. And this seems more informative.
So if I choose to plot (and therefore analyze) unsymmetrical data, will the DVs (auc, MAD, etc) be calculated correctly?
Again, I apologize for the long post with numerous questions and I apologize if this thread is not in the correct location. Infinite thanks to any and all who can provide me with guidance on these questions!