Objective We investigated if a 5-week computerized adaptive working memory training program (Cogmed ® ) of 20 to 25 sessions would be effective in improving the working memory capacity and other neuropsychological functions compared to a non-adaptive working memory training program (active-controlled) in adult patients with mild cognitive impairment (MCI). Methods This randomized double-blinded active control trial included 68 individuals aged 43 to 88 years, 45 men and 23 women, who were diagnosed with MCI at four Memory clinics. The study sample was randomized by block randomization to either adaptive or non-adaptive computerized working memory training. All participants completed the training, and were assessed with a comprehensive neuropsychological test battery before the intervention, and at 1 and 4 months after training. Results Compared to the non-adaptive training group, the adaptive training group did not show significantly greater improvement on the main outcome of working memory performance at 1 and 4 months after training. Conclusion No difference were found between the two types of training on the primary outcome of working memory, or on secondary outcomes of cognitive function domains, in this sample of MCI patients. Hence, the hypothesis that the adaptive training program would lead to greater improvements compared to the non-adaptive training program was not supported. Within group analyses was not performed due to the stringent RCT design.

BackgroundMild cognitive impairment (MCI) is a condition characterized by memory problems that are more severe than the normal cognitive changes due to aging, but less severe than dementia. Reduced working memory (WM) is regarded as one of the core symptoms of an MCI condition. Recent studies have indicated that WM can be improved through computer-based training. The objective of this study is to evaluate if WM training is effective in improving cognitive function in elderly patients with MCI, and if cognitive training induces structural changes in the white and gray matter of the brain, as assessed by structural MRI.Methods/DesignsThe proposed study is a blinded, randomized, controlled trail that will include 90 elderly patients diagnosed with MCI at a hospital-based memory clinic. The participants will be randomized to either a training program or a placebo version of the program. The intervention is computerized WM training performed for 45 minutes of 25 sessions over 5 weeks. The placebo version is identical in duration but is non-adaptive in the difficulty level of the tasks. Neuropsychological assessment and structural MRI will be performed before and 1 month after training, and at a 5-month folllow-up.DiscussionIf computer-based training results in positive changes to memory functions in patients with MCI this may represent a new, cost-effective treatment for MCI. Secondly, evaluation of any training-induced structural changes to gray or white matter will improve the current understanding of the mechanisms behind effective cognitive interventions in patients with MCI.Trial registrationClinicalTrials.gov NCT01991405. November 18, 2013.

In this cross-sectional study, we sought to describe cognitive and neuroimaging profiles of Memory clinic patients with Mild Cognitive Impairment (MCI). 51 MCI patients and 51 controls, matched on age, sex, and socio-economic status (SES), were assessed with an extensive neuropsychological test battery that included a measure of intelligence (General Ability Index, “GAI,” from WAIS-IV), and structural magnetic resonance imaging (MRI). MCI subtypes were determined after inclusion, and z-scores normalized to our control group were generated for each cognitive domain in each MCI participant. MR-images were scored by visual rating scales. MCI patients performed significantly worse than controls on 23 of 31 cognitive measures (Bonferroni corrected p = 0.001), and on 8 of 31 measures after covarying for intelligence (GAI). Compared to nonamnestic MCI patients, amnestic MCI patients had lower test results in 13 of 31 measures, and 5 of 31 measures after co-varying for GAI. Compared to controls, the MCI patients had greater atrophy on Schelten's Medial temporal lobe atrophy score (MTA), especially in those with amnestic MCI. The only structure-function correlation that remained significant after correction for multiple comparisons was the MTA—long delay recall domain. Intelligence operationalized as GAI appears to be an important moderator of the neuropsychological outcomes. Atrophy of the medial temporal lobe, based on MTA scores, may be a sensitive biomarker for the functional episodic memory deficits associated with MCI.

Working memory training (WMT) effects may be modulated by mild cognitive impairment (MCI) subtypes, and variations in APOE-epsilon (APOE-ε) and LMX1A genotypes. Sixty-one individuals (41 men/20 women, mean age 66 years) diagnosed with MCI (31 amnestic/30 non-amnestic) and genotyped for APOE-ε and LMX1A completed 4 weeks/20–25 sessions of WMT. Cognitive functions were assessed before, 4 weeks and 16 weeks after WMT. Except for Processing Speed, the non-amnestic MCI group (naMCI) outperformed the amnestic MCI (aMCI) group in all cognitive domains across all time-points. At 4 weeks, working memory function improved in both groups (p < 0.0001), but at 16 weeks the effects only remained in the naMCI group. Better performance was found after training for the naMCI patients with LMX1A-AA genotype and for the APOE-ε4 carriers. Only the naMCI-APOE-ε4 group showed improved Executive Function at 16 weeks. WMT improved working memory and some non-trained cognitive functions in individuals with MCI. The naMCI group had greater training gain than aMCI group, especially in those with LMX1A-AA genotype and among APOE-ε4-carriers. Further research with larger sample sizes for the subgroups and longer follow-up evaluations is warranted.