Ten trials (305 participants) were included in the review: five were RCTs, three were non-randomised controlled trials, and two were uncontrolled clinical trials. The average quality rating for the RCTs was 13.5 (range 9 to 18).
All outcomes: The effect sizes for the various domains in the trials ranged from -0.82 to 1.21 for cognitive exercise, and from -1.18 to 0.88 for memory strategy training. Most of the results were not statistically significant. Large effects (over 0.5) were reported for half the memory outcomes in trials of cognitive exercise, and 37% of outcomes in trials of memory strategies.
Depression: Two RCTs reported improvements in mood following cognitive exercises. One of these reported a significant and clinically relevant reduction in psychiatric and depressive symptoms, with exercises compared with cholinesterase inhibitors (ES 0.82, 95% CI 0.14 to 1.50). One non-randomised controlled trial and one uncontrolled trial reported small reductions in depressive symptoms after cognitive exercise.
Number of sessions: One RCT reported a greater effect on memory outcomes with a higher volume of cognitive exercise training; 60 sessions had an effect size of 0.82 or 0.99, while 12 sessions had an effect size of 0.10. This effect was not evident for memory strategy training (one non-randomised controlled trial).
Long term: Four trials reported long-term outcomes. Compared with control, memory strategy training decreased objective memory performance at three (one RCT) and six months (one RCT) after the intervention. With cognitive exercise, improvements in function were reported at three months (one RCT), but not at five months (one uncontrolled trial).