Wei Ji Ma

I started out as a theoretical physicist, before realizing that I wasn’t very good at that and that I found the brain more interesting. Like many physicists, I switched to computational neuroscience. Unlike most, I was quickly drawn to the beauty of constructing mathematical models of human behavior as well. I did postdocs with Christof Koch at Caltech and with Alex Pouget at the University of Rochester. My lab works on decision-making at several levels, with an emphasis on the role of uncertainty. We use psychophysics to test models in which observers take uncertainty into account (perform “probabilistic computation”) against alternative models in domains such as visual search, categorization, multisensory perception, and change detection (a working memory task). We study how models of probabilistic computation can be implemented in neural populations. Most recently, we have started a line of work on sequential decision-making in two-player combinatorial games (such as chess and go).