Repetition-spacing and item-overlap effects in the Hebb repetition task

In four experiments using a variation of the Hebb repetition task, we investigated the effects on learning rate, of repetition spacing and of the overlap in experimental items between repeating and nonrepeating lists. In the first two experiments it was shown that when repeating and nonrepeating lists were all permutations of the same items, learning was slower than when they shared no items. Under no-item-overlap conditions in a third experiment, the learning rate for a repeating sequence was shown to be substantial and essentially equivalent for repetitions spaced at every 6th, 9th and 12th trial. Concurrent learning of several different sequences was also demonstrated. When participants were retested after several months on lists that they had previously learned, there was evidence that the learned representations were long-term and order-specific. The results are discussed in relation to two recent models of the Hebb effect.