Multiple-trial learning is a simple form of learning based on positive outcomes. The learning experience is repeated several time, and each of these experiences is recorded in memory. This form of learning utilizes all of the relevant experiences by looking through memory for all the times in the recent past that a similar situation led to a positive outcome.
The integration (averaging?) of multiple action recordings must result in a single composite action sequence that can be substituted for the individually recorded sequences discussed sofar.
Action sequences with negative outcomes provide yet other challenge to our simple model, as we shall see below.