A previous group analysis of the effects of a computerized written naming rehabilitation program revealed global improvement with generalization of benefits to untrained items and to untreated oral naming (Deloche et al., 1992). The present multiple single-case analysis of the data indicates a variety of patterns of improvement and of generalization effects among individual patients. Patterns of relationships between written and oral naming behaviors help to explain the type of improvement that was observed.