This paper, sadly my last with my talented PhD student Fredrik, is another example of how rubbish my intuition is! As you'll see from Fredrik's other papers, we'd been thinking a lot about how the neuronal encoding of disparity places limits on our spatial resolution, drawing inspiration from the pair of 2004 papers from my former colleagues Bruce and Hendrikje, and from Marty Banks, Mike Landy and Serge Gepshtein (whom Fredrik visited at RIKEN). It was now looking pretty convincing that human spatial resolution for disparity is set by the receptive fields of V1 neurons. Fredrik and I wondered if we could detect the psychophysical signature of other cortical areas in the same way. Area MT contains many neurons which can detect conjunctions between disparity and motion, but these neurons have much larger receptive fields than those in V1. We devised a task which required humans to do just that. It's a harder task than just detecting disparity, but we equalised the difficulty by varying the coherence of the stimulus. I was convinced that people would have even worse spatial resolution for this task than for disparity, reflecting the larger receptive fields in MT. In fact, resolution was only slightly worse than for disparity, nowhere near as coarse as it would have been if MT receptive field size were limiting resolution. I found this surprising and interesting. Yes, the initial disparity/motion information is encoded by V1, but we would expect that information to be "read out" in some other brain area, and at the moment it's not clear where that is. In fact, if I'd thought about things more, I might not have been so surprised. The spatial resolution for motion (on its own, not in conjunction with disparity) is already too good to be explained by MT receptive fields. fMRI suggests some candidate areas, but as far as I'm aware, no one has yet recorded from neurons with properties that seem able to explain our good resolution for motion, disparity and the conjunctions between them.