An approach to visualize the course of solving of a research task in humans

A technique to study the dynamics of solving of a research task is suggested. The research task was based on specially developed software RightWrong Responder (RWR), with the participants having to reveal the response logic of the program. The participants interacted with the program in the form of a semi-binary dialogue, which implies the feedback responses of only two kinds “right” or “wrong”. The technique has been applied to a small pilot group of volunteer participants. Some of them have successfully solved the task (solvers) and some have not (non-solvers). In the beginning of the work, the solvers did more wrong moves than non-solvers, and they did less wrong moves closer to the finish of the work. A phase portrait of the work both in solvers and non-solvers showed definite cycles that may correspond to sequences of partially true hypotheses that may be formulated by the participants during the solving of the task.