Observational learning of a spatial discrimination task by rats: learning from the mistakes of others?

Learning by observing others has been acknowledged as a powerful learning strategy. Whereas in several species observation of fear conditioning or other operational procedures can improve subsequent performance during actual learning, much less attention has been paid to observational learning of spatial discrimination tasks. To this end, we developed a set of procedures in which the spatial memory of adult rats, Rattus norvegicus, was tested in an eight-arm radial maze. Moreover, in view of controversial information concerning the incidence of mistakes made by demonstrators on the effectiveness of observational learning, our observer rats watched experienced or nontrained demonstrators. Food-deprived observers and demonstrators were initially habituated to the maze with all arms baited. Then observers were placed in a mesh cage positioned above the maze while a demonstrator rat was locating the spatial position of three baited arms. Rats observing conspecifics progressively learning the spatial discrimination improved subsequent performance compared to a control group watching an empty maze, but only if the configuration of baited arms presented during demonstration and testing matched. Therefore, rats integrated relevant spatial information during observation and used it efficiently when their spatial discrimination was tested in the maze. However, when the information was provided by trained demonstrators, making no mistakes and visiting only baited arms, observer rats failed to exhibit improved performance. Nevertheless, when given an initial habituation without food rewards, rats were subsequently able to benefit from observation of trained demonstrators thus showing that watching mistakes was not necessary for successful observational learning. Together, these findings indicate that rats can acquire spatial information via observation enabling more pertinent search strategies during testing and that for observation to be beneficial, what is observed must be sufficiently relevant or novel to complement existing knowledge (here initial habituation with or without rewards).