<i>Aims:</i> Different patterns of intact and disturbed working memory function can be observed in schizophrenic patients depending on the type of n-back task. We investigated whether these patterns can be induced in healthy subjects by experimentally preventing a motor encoding strategy. <i>Methods:</i> Thirty-two healthy subjects were asked to solve 2 types of n-back task. In the continuous matching task, the subjects had to compare the present stimulus with the one occurring n stimuli back. In the continuous delayed response task (CDRT), the subjects had to select a response depending on the stimulus n stimuli back. Both types of n-back task are assumed to differ with respect to the encoding strategies that can be used to solve the tasks. The use of a motor strategy was prevented by a random arrangement of the target buttons. <i>Results:</i> When the position of the target buttons was predictable, CDRT was solved faster and with higher accuracy than the continuous matching task. However, CDRT was solved more slowly and less accurately when the arrangement of the target buttons varied between the trials. This resulted in a comparable performance in both types of the n-back task. <i>Conclusions:</i> The behavioural alteration in schizophrenic patients in n-back tasks can be induced in healthy subjects by experimentally preventing the use of a motor encoding strategy.