The serial reaction time task is often used to research (implicit) sequence learning. There is an ongoing debate about what exactly is learned in the task. Is it the (spatial) sequence of the stimuli, the sequence of motor responses or the (spatial) sequence of keys used to answer? Although there are studies showing evidence for each of the learning possibilities as well as their combination (visuomotor learning, Pedraza et al., 2023 or feature codes, Haider et al., 2020), studies using motor responses seem to show stronger expressions of sequence knowledge. Thus, the expression of learned knowledge could benefit from the response format. The present study tries to disentangle the relevance of motor responses and the spatial layout of the keys for the expression of sequence knowledge. To this end four conditions are tested in which a learning phase and a transfer phase are presented with different combinations of finger-key-mappings. In the first two conditions, the specifity of the finger key mapping is varied. Either each finger is assigned to a certain key or the participants are asked to press different keys with the same finger. In the second two conditions, the spatial configuration of the keys is changed. Either the spatial configuration is horizontal or vertical. The expression of sequence knowledge is compared between the learning and the transfer phases in the four conditions.