Here are some graphs I generally plotted by mistake, but that I kept because I thought they were pretty.

This plots object category against the trial that object was presented, with individual participants drawn as blue lines and the red line showing the mean. It's a quick and pretty way of checking your experiment is randomised correctly! I think it looks a bit like Barbara Hepworth: hepworth.

These are some failed ROC curves.

These are some plots of EEG data showing relative power per electrode, recorded using a 128-sensor electrode net. The electrodes you can see 'popping' at 50Hz are picking up mains noise because they are not stuck properly onto the scalp. All the way down to 16Hz you can see corruption by what I think is muscle artifact, in two patches above the eyes and two at the back, close to the neck.

I love my Picasso so I really enjoyed running the data analyses for a recent paper on signal-to-noise ratios in children's TV, where some of the designs are beautiful and definitely influenced by Picasso, and Miro. The paper is about how certain types of visual information, of which Miro paintings are actually a really good example, are easier for babies' brains to perceive because of signal-to-noise ratios. This one is from an analysis looking edge density, which is calculated from the rate of change of low-level visual information content, for a sample frame from In The Night Garden. It was made using scripts from Rosenholz et al.

This is another sample frame from In the Night Garden.

This shows luminance in a sample frame from Charlie and Lola. It was calculated using the Matlab command imagesc, which is a nice one for pretty figures.

This shows the position of infants' heads relative to the screen during a recording session. It reminds me of a de Kooning

This is shows infants' change in pupil size over time. I think it looks like Fraggle hair!

This is another one of pupil size.