Forty years after C.P. Snow's infamous Two Cultures lecture — about the inability of artists and scientists to speak one another's languages — rapprochement is in the air. In the UK, science is sexy in the arts. Media-friendly scientists like Richard Dawkins are on every chat show. Chaos theory and the Copenhagen Interpretation currently enthral London theatre audiences. Last year, Edinburgh gallery-goers flocked to an exhibit on the art and science of faces. Geneticist Steve Jones features in an advertising campaign for hi-tech cars. And at the Mirrors in Mind exhibition at the National Gallery last year, hard-nosed scientists and art historians had a joint study day.The Wellcome Trust, the world's largest medical research charity, has recently been using some of its money to encourage collaboration of scientists and artists, through the Sci∼Art competition. It's a laudable aim but the cynical, including this runner-up, suspect some winning entries have more to do with artists representing science than with serious science. I could well imagine a winning entry that had a grandmother, mother and daughter team of knitters collaborating with the neighbourhood physicist and video artist to explore quantum chromodynamics in wool. The risk is that science-with-art does not acknowledge that most real science is about difficult ideas, hard thinking, messy data and the occasional lucky break.My expectations were therefore not high when I went to The Painter's Eye, a Sci∼Art sponsored exhibition at the National Portrait Gallery in London. I was wrong. The exhibition is fascinating, and scientifically and artistically sophisticated. Humphrey Ocean, a portrait painter, was filmed drawing portraits while wearing an eye-movement tracker, using a pencil whose three-dimensional position was continually mapped, and while inside an fMRI scanner. Meticulous editing has generated a beautiful video of the emerging portrait overlaid with the artist's foveal fixations on the subject. Every pencil movement is recorded, including many in which the pencil is a few millimetres above the surface of the paper while the artist, who has a very deliberate and controlled technique, practices the mark he will eventually make (see FigureFigure).Figure 1Humphrey Ocean's Luke 2 drawing (left) and the hand movements that created it (right). All movements of the pencil within one centimetre of the paper are shown in black. The hand was making many more movements near the paper than were required just to draw the lines.View Large Image | View Hi-Res Image | Download PowerPoint SlideAnd, with a pleasing circularity, the hand movement data have been used to create a sculpture that is a piece of art in its own right. Martin Moore's wire sculpture traces out the three-dimensional movements of Ocean's pencil tip, such that the flat portrait itself seems to hang eerily in space.Perhaps least interesting are the fMRI data which, unsurprisingly, show increased frontal activity in the professional artist compared with inexperienced controls, where posterior, visual cortical activity predominates. The reason is clear from the artist's own insights: “I'm sure of what I am seeing, I'm not sure what I'm going to do about it. So I make a decision. The final result is made up of a great many decisions.” As with most experts, sensory, motor and cognitive skills are mostly automatized, leaving high-level executive decisions to be made by the frontal lobes of the brain.The exhibition is summed up in the four display captions: “The eye captures; the brain processes; the hand implements; the eye evaluates.” The first three are beautifully elucidated. Of the latter, sadly, all is silence. What happens when the artist tries out good and bad versions of a line? Does a picture composed of chosen lines differ aesthetically from one made of rejected lines?Sadly, the exhibition is too restricted to neuroscience and to impressive but overly technical plots and coloured scans that would probably baffle most visitors. Cognitive psychology and experimental aesthetics could have helped make use of the wonderful data for understanding the crucial question of how an artist makes a truly good portrait.