Abstract: Objective  To evaluate the effect of computerized cognitive training on elderly with mild cognitive impairment.Methods  A randomized controlled trial on the effect of computerized cognitive training on elderly with mild cognitive impairment was searched in Chinese and English databases, and metaanalysis was performed using RevMan 5.2 software．Results  A total of 12 articles were included,  involving 488 patients.Metaanalysis showed that the improvement of MMSE and MoCA in the intervention group was better than that in the control group after the intervention (MD=1.78, 95%CI 0.41，3.15，P<0.05;  MD=4.02, 95%CI 1.33，6.70, P<0.05),  the improvement of DSF and DSB in the intervention group was better than that in the control group after the intervention(MD=1.01,  95%CI 0.39，1.64, P<0.05; MD=0.67, 95%CI 0.16，1.17, P<0.05),  the improvement of delay recall in the intervention group was better than that in the control group after the intervention(MD=1.20, 95%CI 1.00，1.39 , P<0.05),   But, it had no obvious effect on SF, GDS and ADL. Conclusion  Computerized cognitive training can improve the global cognitive function of the elderly with mild cognitive impairment,  especially attention and shortterm memory,  as well as delayed memory. But, it has no obvious effect on language abilities, depressive and everyday activities. Largesample,  highquality randomized controlled study still needs more verification.

Key words: mild cognitive impairment, Alzheimer's disease;computerized cognitive training