BACKGROUND:Memory problems are a common cognitive complaint following stroke. Memory rehabilitation programmes either attempt to retrain lost or poor memory functions, or teach patients strategies to cope with them. OBJECTIVES:To determine the effectiveness of cognitive rehabilitation for memory problems following stroke. SEARCH STRATEGY:We searched the Cochrane Stroke Group Trials Register (last searched September 2006). In addition, we searched the following electronic databases; the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library Issue 2, 2005), MEDLINE (1966 to June 2005), EMBASE (1980 to June 2005), CINAHL (1982 to June 2005), PsycINFO (1980 to July 2006), AMED (1985 to June 2005), British Nursing Index (1985 to June 2005), CAB Abstracts (1973 to May 2005) and the National Research Register (June 2006). We handsearched relevant journals and searched reference lists. SELECTION CRITERIA:We selected controlled trials of memory retraining in stroke. We excluded studies with mixed aetiology groups unless 75% or more of the participants had a stroke or separate data were available for the stroke patients. DATA COLLECTION AND ANALYSIS:Two review authors selected trials for inclusion, assessed quality, and extracted data. MAIN RESULTS:Two trials, involving 18 participants, were included. One study compared the effectiveness of a mnemonic strategy treatment group with a 'drill and practice' control, while the other compared the effectiveness of an imagery mnemonics programme with a 'pragmatic' memory rehabilitation control programme. Formal meta-analyses could not be performed due to a paucity of studies and lack of commonly-employed outcome measures. The results do not show any significant effect of memory rehabilitation on performance of objective memory tests, and no significant effects of treatment on subjective and observer-rated measures of memory. AUTHORS' CONCLUSIONS:There was no evidence to support or refute the effectiveness of memory rehabilitation on functional outcomes, and objective, subjective, and observer-rated memory measures. There is a need for more robust, well-designed and better-reported trials of memory rehabilitation using common standardised outcome measures.