Sequencing and interleaving in routine action production

We analysed the sequence of steps enacted during the performance of routine action tasks in the context of the 2 x 3 test. This test requires enactment of three routine tasks, each performed twice. Data from 35 healthy controls and 16 closed head injury patients were analysed to derive 'modal routes' through the 'sandwich', 'toast', and 'present' tasks, along with transition probabilities for the steps in the modal routes. Also investigated were the amount of interleaving and a potential structural constraint on interleaving. The results indicate that serial order production is more flexible than some might predict, and that the patients and controls were more similar than dissimilar on the measures taken. The first result argues against script architectures that are overly rigid. The second implies that the vulnerability to sequence and other errors in this patient group is not due to defects in script memories or serial order programming.