Spatial compatibility and incidental sequence learning in the serial reaction time task

Memory for order information is studied by the presentation of a string of items, after which the subject is asked to recall the information in order. The serial learning task examines how repetition of a sequence leads to incidental sequence learning. Sequenced responses are faster and more accurate than random responses, reflecting advanced response-selection processes. Stimulus-response (S-R) compatibility also reflects response selection processes. Therefore, it is of theoretical interest to examine the interaction between the two. The Salient Features model indicates that task demands define which features will be attended in a set of stimuli or responses, which in turn determines how they will be coded. The Feature Model of memory indicates that the degree to which a cue uniquely specifies an item determines whether the correct item will be retrieved. Therefore, the task demands provide both the information that determines how the item is coded and the cues that specify the response. In serial learning this means that the task demands determine whether the subject encodes a series of responses, stimuli or stimulus-response pairs. The current experiments investigated how spatial compatibility affected sequence learning, and whether sequence learning affected the magnitude of the S-R compatibility effect, which would be expected because both reflect response-selection processes. Four experiments demonstrated that increasing the difficulty of the spatial translation decreased sequence learning. Although sequence knowledge did not eliminate the spatial compatibility effect, they did interact. The compatibility effect decreased faster in the sequenced conditions than in the random response conditions. S-R compatibility and response hand were manipulated to determine what is learned: A sequence of motor responses, stimulus positions or S-R paired associates. Subjects appeared to learn a sequence of S-R pairs and a sequence of stimulus-to-effectors relations. Under dual-task conditions, reliable evidence of sequence learning was found for the compatible S-R mapping, but not for the incompatible mapping. This supports the hypothesis that dual-task conditions suppress the expression of sequence knowledge, but do not prevent sequence learning. Time between stimulus presentations was a factor in sequence acquisition