. 2017 Dec 27;27(4):440–484. doi: 10.1007/s11065-017-9363-3

Table 1.

Summary of findings: cognitive intervention impact on participants with mild cognitive impairment (MCI)

	Author	Study design	Sample size & age means	Intervention and Duration	Outcome measures	Results	General conclusions	NIH quality (standard score)
1.	Balietti et al. (2016)	RCT, inactive control	n = 70 (MCI) Tx n = 37 Cntl n = 33 Age: $\bar{x}$ = 75.74 s = 0.716	• 10 sessions, 60 min-1×/ week (10 weeks) • Group based, multi-component cognitive training using cognitive exercises, effective aids, and mnemonic strategies (errorless learning, spaced retrieval, organization, categorization, clustering, method of loci, visual imagery, and face-name associations)	DS-FWD; CSST; AMT, phonemic fluency, semantic fluency, prose recall, word pair learning	• Cognitive training group demonstrated statistically significant increased improvements in attentive matrices (p = 0.010), phonemic verbal fluency (p = 0.018), and recall of prose passages (immediate p = 0.037, delayed p = 0.038, and total recall p = 0.006) • MCI controls evidenced no change from baseline scores in follow-up evaluation	• While this study was primarily focused on effect of cognitive training on platelet phospholipases A₂ activity (tPLA₂A), a benefit in cognition was observed in several areas of cognition • Increases were noted in AMT, phonemic fluency, and recall of prose passages • There appeared to be no benefit in attention, CSST, semantic fluency, or word pair learning	−0.713
2.	Barban et al. (2016)	RCT, single-blind, inactive control	n = 106 (MCI) Tx n = 46 Cntl n = 60 Age: $\bar{x}$ = 74.4 s = 5.7	• 24 sessions, 60 min session (2×/ week for 3 months) • Group based, computer administered with software designed in multicomponent cognitive exercises/training in memory, logical reasoning, orientation, language, constructional praxis, questions concerning autobiographical data (SOCIABLE)	RAVLT, ROCT, TMT A&B, Phonemic Fluency, MMSE	• Significant differences observed post-training in RAVLT scores (p = 0.009), Phonemic fluency (p < 0.001), and MMSE (p = 0.01) in the MCI group • No differences found on RCFT or TMT measures • Improvement in RAVLT memory scores maintained after a three month ‘rest’/ period of no training or booster	• There was a medium positive effect of computer-based training on verbal memory, phonemic fluency, and mental status • Training benefit was sustained on verbal memory after three months of ‘rest’ (no training) • Computer-based multi-component training demonstrated positive effects on cognition sustained at three months; verbal memory in particular • Unclear if scores are significantly different from individuals in the passive, control arm	−0.026
3.	Barnes et al. (2009)	RCT, single-blind, active controls	n = 47 (MCI) Tx n = 17 Cntl n = 19 Age: $\bar{x}$ = 74 s = not reported Range = 54–91	• 100 min/ day, 5 days/ week • Computer-based, at-home independent exercises using restorative training strategies targeting auditory processing speed and working memory	1° - RBANS 2° - CVLT-II, COWAT, BNT, California TMT, Design Fluency tests, Spatial Span test, GDS	• Most group differences not statistically significant (i.e. RBANS total scores ↑ 0.36 SD in the intervention group (p = 0.097) compared to 0.03 SD in the control group (p = 0.88) for a non-significant difference); however, effect sizes for measures of learning/memory favored intervention group (range: 0.16 to 0.53 SD) • Largest effect size observed for Spatial Span, scores ↑ significantly in the intervention group (p = 0.04) and ↓ significantly in the control group (p = 0.02) for a significant effect size (p = 0.003)	• Intensive computer-based mental activity training is feasible in elders with MCI • Training appeared to benefit spatial span ability • Trend toward benefit of training on measures of learning/ memory • Training did not appear to benefit language, visuospatial skills	−0.026
4.	Buschert et al. (2011)	RCT, single-blind, active control	n = 43 aMCI n = 27 mild AD n = 16 Tx n = 10 Cntl n = 12 Age: $\bar{x}$ =71.2 s = 7.0	• 11 session minimum - 20 “units”, 120 min/ week • Group-based, multicomponent cognitive intervention including memory, cognitive stimulation, face-name association, errorless learning, principles of meta-cognition, education	ADAS-cog, MMSE, TMT B, RBANS, story recall, MADRS, QoL-AD	• Significant interaction between treatment and progression for ADAS-Cog (F = 6.2, p = 0.02, η2 = 0.26), MMSE (F = 3.8, p = 0.07, η2 = 0.17), RBANS story memory (F = 3.4, p = 0.08, η2 = 0.16) and TMT = B (F = 3.5, p = 0.08, η2 = 0.16) • Main effects for treatment were found for MMSE (F = 8.5, p < 0.01, η2 = .23), RBANS story memory (F = 12.5, p < 0.01, η2 = 0.41) and RBANS Story recall (F = 9.9, p < 0.01, η2 = 0.36)	• ↑ in global cog status, and specific cog and non-cog functions • aMCI demonstrate significant change in ADAS-Cog scores • aMCI subjects demonstrated tendency toward higher attentional skills demonstrated by TMT-B • No significant effect on memory performance compared to active controls	0.661
5.	Carretti et al. (2013)	RCT, active control	n = 20 (aMCI) Tx n = 10 Cntl n = 10 Age: Tx: $\bar{x}$ = 71.8 s = 2.20 Cntl: $\bar{x}$ = 70.6 s = 2.63	• 5, 90 min sessions • Individual-based, one-on-one sessions using restorative strategies which focused on working memory techniques	NPE, Vocab, CWMS, DS-FWD; DS-BWD, Dot matrix, List recall, Pattern comparison, Cattell test	• Significant effect post training in CWMS (3.8), Dot Matrix (2.3), Cattell test (0.50) vs control CWMS (0.60), Dot Matrix (2.3), Cattell test (−0.40). P-values for these tests where groups improved by 1 SD are <0.01, <0.05, and <0.05, respectively	• Verbal working memory training is a promising approach to sustaining memory function in aMCI • Working memory training showed transfer to some cognitive components of memory part of the core cognitive impairments responsible for MCI → AD	−0.713
6.	Fiatarone Singh et al. (2014)	RCT, double-blind, active control	n = 100 Tx n = 24 Cntl = 27 Age: $\bar{x}$ = 70.1 s = 6.7	• 4–45 min exercises (initial/ group setting); then 45 min sessions, 2 days per week for 24 weeks • CT computer-based multicomponent and multi-domain compensatory training in memory, executive functions, attention, and speed of information processing (COGPACK)	ADAS-Cog; MMSE; GP-Cog; CDR; Matrices, Similarities, TMT A & B, LM I&II; BVMT-R; SDMT; Semantic Fluency, COWAT, MARS-MF	• ADAS-Cog scores: No differences between CT and sham cognitive training • CT group demonstrated modest non-significant changes as compared to controls across cognitive domains • CT group maintained memory ability, no significant improvement or decline	• CT reported to attenuate decline in memory • There was no significant change in cognitive outcomes between CT and sham training in global cognition, executive functions, memory, or speed	2.035
7.	Finn and McDonald (2011)	RCT, single-blind, inactive control	n = 16 (MCI-SD/MD) Tx = 8 Cntl = 8 Age: Tx: $\bar{x}$ = 69.00 s = 7.69 Cntl; $\bar{x}$ = 76.38 s = 6.47	• 30 sessions, 11.43 weeks/ completion • Computer-based restorative training program involving attention, processing speed, visual memory, cognitive control (LUMOSITY)	1° - CANTAB, RVP A, PAL, IED 2° - MFQ, DASS-21	• Significant CANTAB tests: main effect of group - attentional set shifting, visual learning, visual working memory F (2, 14) = 0.35, 1.17, and 3.55, respectively • ↑ in visual sustained attention treatment (pre/ post difference RVPA’ = .03) compared with waitlist ctrl (pre/ post difference RVPA’ = − .06) • No significant data on self-reported memory functioning and perceptions of control over memory	• ↑ in performance on visual sustained attention compared to waitlist controls • No significant changes on other primary outcome measures • MCI patients can ↑ performance significantly when given repeated practice on computerized cognitive exercises but study did not translate to secondary measures	−1.401
8.	Finn and McDonald (2015)	RCT, inactive control	n = 24 (aMCI) Tx = 12 Cntl = 12 Age: Tx: $\bar{x}$ = 72.83 s = 5.7 Cntl; $\bar{x}$ = 75.08 s = 7.5	• 6 sessions over several weeks (plus one practice session). • Computer-based restorative training program using strategy of repetition-lag training	1° - VPA I & II 2° - WMS-IV – SS; D-KEFS N-S; D-KEFS N-LS; CFQ; DASS-21	• Significant effects of training observed in VPA-II (F = 4.52, p = 0.046) • No other significant effects demonstrated with the exception of practice effects in VPA-I	• Computer-based version of modified repetition-lag training program was associated with improvement in recall of verbal pair associates in aMCI • No evidence of transfer effects on other secondary measures	−0.713
9.	Forster et al. (2011)	RCT, single-blind, active control	n = 24(aMCI), n = 15(mild AD) Tx n = 9 Cntl n = 9 Age: $\bar{x}$ =74.5 s = 8.6	• 26, ~120 min sessions (1×/week for 6 months) • Group-based multi-component cognitive interventions focused on global cognition, mood, and quality of life	ADAS-cog, MMSE, FDG PET scan	• Change in ADAS-Cog scores for aMCI demonstrate interaction treatment and progression (F = 4.7; p = .045; η2 = .22) • Change in MMSE scores for aMCI demonstrate interaction treatment with main effect for treatment (F = 6.8; p = .02; η2 = .29)	• Marginally significant interaction effect for ADAS-Cog and MMSE • MMSE demonstrated main effect for treatment, possibly due to performance decline in controls • ↑ brain energy metabolism in the MCI intervention subgroup.	0.661
10.	Gagnon and Belleville (2012)	RCT, double-blind, active control	n = 24 (MCI-SD/MD) Tx n = 12 Cntl n = 12 Age: Fixed Priority $\bar{x}$ = 67.00 s = 7.80 Variable Priority $\bar{x}$ = 68.42 s = 6.04	• 6, 60 min sessions (3×/ week for 2 weeks) • Computer-based restorative training method to increase attentional control and meta-cognition	1° - (modified) Dual task visual detection/ classical digit span 2° - TEA, TMT-A/B, DAQ, WBS	• Significant: Fixed priority group vs. variable priority group (p < .05) improvement in cost scores for visual detection accuracy was 3.39 to 25.98 respectively. Alpha-arithmetic task, both accuracy and reaction time showed significant main effects of Attention, F (1, 22) = 14.89, (p = .001) and F (1, 22) respectively, and of Intervention, F (1, 22) = 72.80, p = .001, and F (1, 22) = 8.18, p = .01, respectively • Trails A: main effect of Intervention F (1, 22) = 15.91, p = .001, η2 = 0.42 (all participants = faster) • No significant impact on WBS	• Cog intervention may ↑ attentional control in pts. with MCI and an executive deficit.	0.661
11.	Giuli et al. (2016)	RCT, inactive control	n = 97 Tx n = 48 Cntl n = 49 Age: Tx: $\bar{x}$ = 76.0 s = 6.3 Cntl: $\bar{x}$ = 76.5 s = 5.7	• 10, 45 min sessions (1×/ week) • Patient tailored, individual 1:1 sessions. Multi-component strategies in orientation, memory, categorization and clustering, psychological support, psychoeducation, as well as education regarding healthy lifestyles	MMSE, CSST, DS-FWD & DS-BWD, Prose memory, VPA, AMT, Semantic fluency, phonemic fluency, CDR, GDS-30, PSS, ALD, IADL, MAC-Q, Questionnaire of confidence	• Significant improvements in MCI group after training noted in MAC-Q (p < 0.001), Prose memory (p < 0.004), VPA (p < 0.018), CSST (p < 0.040), and AMT (p < 0.001) • No differences observed in DS-FWD, DS-BWD, semantic fluency, or phonemic fluency	• Training in cognitive strategies, psychological support and education regarding healthy lifestyle were associated with improvements in subjective memory complaints as well as an increase in prose recall, word-pairing recall, and sustained attention from baseline scores • Individuals receiving training exceeded performance on outcome measures of passive controls in DS-FWD, DS-BWD, GDS, MAC-Q, IADLs, Prose memory, word pairing, CSST, and AMT.	−0.713
12.	Greenaway et al. (2012)	RCT, inactive control	n = 40 (aMCI-SD) Tx n = 20 Cntl n = 20 Age: Tx: $\bar{x}$ = 72.7 s = 6.9 Cntl: $\bar{x}$ = 72.3 s = 7.9	• 12, 60 min sessions (2×/ week for 6 weeks) • Dyad-based (participant and partner) receiving compensatory training memory strategies (Memory Support System)	DRS-II, MMSE, WMS-R/III, CERAD Word List, WMS-R/ III VR, ECog, QoL-AD, CBQ, Chronic Disease Self-Efficacy Scale, Adherence Assessment	• No significant effects of treatment in vs. control in DRS-II or MMSE • Treatment Ecog significant at 8wk [t (17) = 2.4, p < 0.05] but not at 6 months • MCI sense of memory self-efficacy at the end, t (15) = −3.1 (p < 0.01) vs. controls, t (33) = 2.4 (p = 0.02)	• ↑ Functional ability and sense of self-efficacy compared with controls out to 8-week follow-up • ↑ in ADLs and ↑ ECog scores	−1.401
13.	Hampstead et al. (2012a)	RCT, single-blind, inactive control	n = 49 Tx = 11 Cntl = 10 Age: aMCI mnemonic: $\bar{x}$ = 73.5 s = 10.1 aMCI exposure: $\bar{x}$ = 70.5 s = 5.8	• 3, 60–90 min sessions, over 2 weeks • Individual-based, one-on-one compensatory training in object-location built on face-name association and mnemonic strategies	MMSE, RBANS, TMT, GDS, FAQ, ILV, F-N Accuracy, fMRI imaging	Correlation (Spearman’s Rho) results: • Significant RBANS DMI: aMCI mnemonic group = .67 (p = .009), healthy + MCI mnemonic group = .68 (p < .001), and healthy + MCI exposure = .68 (p < .001) •Significant Trails: aMCI mnemonic group = .57 (p = .03) •Significant ILV: healthy + MCI mnemonic group = −.75 (p = <.001) and aMCI mnemonic group = −.81 (p = .001) •Significant amygdala: healthy + MCI mnemonic group = .54 (p = .01)	• Mnemonic strategies ↑ memory for specific content for at least 1 month	−0.0.26
14.	Herrera et al. (2012)	RCT, single-blind, active control	n = 22 (aMCI-MD) Tx n = 11 Cntl n = 11 Age: Tx: $\bar{x}$ = 75.09 s = 1.97 Cntl $\bar{x}$ = 78.18 s = 1.44	• 24, 60 min sessions (2×/ week for 12 weeks) • Computer-based restorative training in memory & attention	DS-FWD, DS-BWD, DMS48, 12-word recall (BEM-144), 16-FR/CR test, MMSE, Doors/ People memory	Significant cognitive outcomes: • Trained group immediately at end of training - Doors A (9.64 ± 0.53), Doors B (6.36 ± 0.66), DMS48 (96.91 ± 0.58), forward digit span (4.91 ± 0.21), BEM-144 (7.28 ± 0.26), 16-FR/CR test (42.91 ± 0.76), and MMSE (2.09 ± 0.22) • Trained group 6 months after training - Doors A (8.55 ± 0.39), forward digit span (4.92 ± 0.23), BEM-144 (6.86 ± 0.52)	• Cog training associated with ↑ episodic recall and recognition post-training which was also sustained at 6 months post-training	−0.026
15.	Jean et al. (2010a)	RCT, single-blind, active controls	n = 22 (aMCI) Tx n = 11 Cntl n = 11 Age: Tx: $\bar{x}$ = 68.55 s = 9.16 Cntl: $\bar{x}$ = 68.55 s = 5.91	• 6, 45 min sessions (2×/week for 3 weeks) • Individual-session restorative focused training in face-name associations using errorless learning and spaced retrieval	1° - Training measure (free recall and cued recall) 2° - CVLT-II, DRS-2, F-N Recall, MMSE, RBMT, MMQ, SES	• Total profile score RBMT improved significantly (t = 7.687, p < .001) while age, MMSE total score, DRS-2 total score, DRS-2 memory subscale score and CVLT-II delayed free recall did not significantly improve model despite the fact that they correlated significantly with the predicted variable	• ↑ explicit residual memory important factor leads to ↑ outcome when using errorless learning or errorful learning to learn face–name associations. • Structured cog training, focusing on memory issues, w/ pt. support, is effective in MCI-A, regardless of the techniques	−0.026
16.	Jeong et al. (2016)	RCT, double-blind, inactive control	n = 147 (aMCI) Tx n = 71 Cntl n = 76 Age: Tx: $\bar{x}$ = 70.8 s = 6.9 Cntl: $\bar{x}$ = 71.6 s = 6.5	• 24, 90 min sessions (2×/week for 12 weeks) • Group-based, 5 per group • Multicomponent cognitive training (memory, attention, executive functions, language, reality orientation, visual-spatial functions), activities to improve ADLs, knowledge for health and daily life, reminiscence therapy and discussion	1° - ADAS-Cog 2° - MMSE, Digit Symbol Coding, Stroop, Animal fluency, COWAT, SRT, DS-FWD, DS-BWD, CDR, GDS-15, NPI, Bayer ADL, PRMQ, AD8, PMT, MMT-Strategy, Composite scores	• Improvements observed in ADAS-Cog (p = 0.03), PMT (p = 0.03), AD8, and NPI • No differences observed in composite scores for logical memory, working memory, executive functions, MMSE, or CDR	• Benefit of cognitive intervention displayed in ADAS-Cog, prospective memory, and informant rating of subject functioning • Comprehensive multi-modal with multiple approaches targeting multiple domains vs. one approach or single domains. • Benefits of training maintained at 6 months post-completion	2.035
17.	Lam et al. (2015)	RCT, double-blind, inactive control	n = 276 (MCI) Tx n = 145 Cntl n = 131 Age: Tx: $\bar{x}$ = 74.4 s = 6.4 Cntl: $\bar{x}$ = 75.4 s = 6.1	• 48, 60 min sessions (3×/ week for 4 months [Time 1], 12 months total) • Group-based, 12–15 per group with homework assignments • Multicomponent or lifestyle activities categorized into cognitive, physical, social, and recreational (33 in total). Cognitive group attended cognitively demanding activities (i.e. reading, discussing newspapers etc.)	1° - CDR-SOB 2° - ADAS-Cog, CMMSE, List Learning (delayed recall), Digit Span, Visual Span, CVFT, C-TMT, MIC, CSDD, CDAD	• No group differences were observed post-training (12 month) • No differences observed in CDR-SOB, CDAD, or CMMSE scores	• No change in general cognition scores (CRD-SOB) over one year may suggest plateau of decline (stabilization), a possible benefit of structured lifestyle activities	2.035
18.	Mowszowski et al. (2014)	RCT, double-blind, inactive control	n = 40 (MCI) Tx n = 25 Cntl n = 15 Age: Tx: $\bar{x}$ = 74.4 s = 6.4 Cntl: $\bar{x}$ = 75.4 s = 6.1	• 14, 120 min sessions (2×/ week, 7 weeks) • Group-based, 10 per group (60 min per group, 2× per week) followed by individually tailored computer-based training using NEAR model • Multicomponent, computer-based cognition training (not-specified) and psychoeducation	1° - EEG 2° - WTAR, MMSE, RAVLT, FAS, Semantic Fluency (Animals), WAIS-III-DS (total), TMT-B	• Cognitive trained group evidenced improvements in phonemic fluency (FAS), vs. controls who declined during the waitlist period • No differences observed in DS, RAVLT, semantic fluency, or TMT-B	• Data from EEG findings suggest enhanced response from frontal and central regions following cognitive training • Cognitive training associated with improvements in phonemic fluency • No increase observed in attention/ working memory, verbal learning & memory, semantic fluency or cognitive flexibility	−0.026
19.	Olchik et al. (2013)	RCT, single-blind, inactive control	n = 30 (MCI) Tx n = 16 Cntl n = 14 Age: Tx: $\bar{x}$ = 70.3 s = 4.3 Cntl: $\bar{x}$ = 70.2 s = 5.7	• 8, 90 min sessions (2×/ week, 4 weeks) • Group-based, 10 per group, comprised of both MCI and normal controls • Group sessions multicomponent training focused on memory with each session beginning with explanatory/ education followed by training in a memorization target task/ strategy (active attention, categorization, association, or visual imagery), and exercises to practice the strategy.	MMSE, Lawton IADL, CRD, Semantic Fluency (Animals), COWAT (FAS), RAVLT, RBMT	• There were no statistically significant effects for memory training across groups post-training • Memory training was associated with greater improvement in FAS and RAVLT scores (compared to other groups examined) • MCI participants demonstrated more significant increase in scores than normal controls in RAVLT (immediate & delay) and RBMT (screening)	• Memory training resulted in higher change in scores from pre-training values, beyond education trained or inactive controls • MCI individuals appeared to benefit more from training than normal controls, supporting the compensation hypothesis • Memory training was not associated with improvements in cognitive outcomes of fluency (semantic or phonemic), verbal list learning & recall, or general memory performance	−1.401
20.	Polito et al. (2015)	RCT, single-blind, inactive control	n = 44 (MCI) Tx n = 22 Cntl n = 22 Age: Tx: $\bar{x}$ = 74.0 s = 1.4 Cntl: $\bar{x}$ = 74.3 s = 1.7	• 10, 100 min sessions (2×/ week, 5 weeks) • Group-based, 7–8 per group • Group sessions with multicomponent training using body awakening, reality orientation, and multiple compensatory cognitive exercises. Exercises were designed to stimulate attention (auditory and visual), executive reasoning, language (fluency), semantic memory, visual perception, encoding, information storage, nonverbal learning and executive problem solving.	MMSE, MOCA, and CSST	• Participants receiving either cognitive training or sham treatment both demonstrated a significant improvement in MMSE and MOCA scores • Improvements from baseline scores were observed in the MCI trained group post-training, although this did not reach significance for any outcome measure (MMSE, MOCA, & CSST)	• Cognitive training was not associated with an increase in cognitive performance on outcome measures • An increase in performance observed in the sham training group (inactive controls) may be attributed to a placebo effect or, possibly, represent practice effects	0.661
21.	Rapp et al. (2002)	RCT, single-blind, inactive control	n = 19 (MCI) Tx n = 9 Cntl n = 10 Age: Tx: $\bar{x}$ = 73.33 s = 6.61 Cntl: $\bar{x}$ = 75.10 s = 7.03	• 6, 120 min sessions (1× /week for 6 weeks) • Group-based, multicomponent training using education, relaxation training, memory skills training, and cognitive-restructuring	CERAD neuropsychological battery, MMSE, Face-Name, MFQ, Memory Controllability Inventory, POMS	• Mean values noted for word list delay of trained group. The data for pre-test, post-test, and follow-up was [3.56; SD = 2.92], [8.44; SD = 4.22], and [6.71; SD = 3.99]. Follow-up was significant (p < 0.07) • Pt’s rated their memory ability higher than controls (i.e. MCI-Present Ability scale, p = 0.008) • Training led to ↑ expectations for future improvement (i.e. MCI-Potential Improvement: p = 0.005) and ↓ expectations for cognitive decline MCI Inevitable decline: p = 0.06) • No change in objective laboratory memory tasks following training	• Cognitive/ behavioral group intervention targeting memory performance and memory appraisals can be effective at changing perceptions of memory ability in a high-risk population of older adults with MCI • Older adults with MCI may need more skills training to achieve and maintain performance improvements.	−0.713
22.	Rojas et al. (2013)	RCT, inactive control	n = 30 (MCI) Tx n = 15 Cntl n = 15 Age: Tx $\bar{x}$ = 72.00 s = 14.29 Cntl: $\bar{x}$ = 76.93 s = 7.05	• 52, 120 min sessions (2× /week for 6 months) • Group-based multicomponent intervention program including cognitive-stimulation, cognitive training, and education	1° - MMSE, CDR 2° - SMB, SF, BNT, PhF, Verbal Fluency, WASI, Similarities and Matrix reasoning, Block Design, TMT A/B, digit span forward/backward WAI-III, QoLQ, NPI, ADL Scale	• Trained group: significant mean of change for BNT [−2.84, p = .04] and SF [−3.03, p = .004] • Non-trained group: significant mean of change for MMSE [1.77, p = .002], Mem-REC [1.00, p = 0.036], SF [2.40, p = .007], CDR [−.01, p = .02] • No significant differences on secondary outcome measures	• Training group improved on BNT and semantic fluency	−1.401
23.	Schmitter-Edgecombe and Dyck (2014)	RCT, single-blind, inactive control	n = 46 (MCI) Tx n = 23 Cntl n = 23 Age: Tx: $\bar{x}$ = 72.96 s = 7.05 Control $\bar{x}$ = 73.35 s = 7.89	• 20, 120 min sessions, (2×/ week for 10 weeks) • Post-Booster: 120 min session (1× /month for 9 months) • Group-based, care-dyad, multicomponent training including workbook lessons, and education workshop	WTAR, TICS, MMAA, EFPT, ADL-PI, RBANS, QOL-AD, CSE, GDS, RBMT-II	• Treatment group performed better than controls on RBMT-II F (1, 43) = 4.20, p < .05, n² _p = .09/ F (1, 22) = 6.84, p = .01, n² _p = .24; RBANS Memory Index Imm F (1, 43) = 4.64, p < .05, n² _p = .10; RBANS Memory Index Delay = ns • MCI pre vs. post show improvements in RBANS Memory Index Imm F (1, 22) = 14.41, p < .001, n² _p = .40; and Delayed F (1, 22) = 9.79, p < .005, n² _p = .31 • Better post-test performance on MMAA and EFPT bill paying subtest. No significant post-test differences in MCI for coping strategies, quality of life, or depression	• Training group demonstrated improvements in everyday memory and immediate memory index as compared to controls • Training group demonstrated gains in memory performance comparing pre- to post- training	0.661
24.	Tsolaki et al. (2011)	RCT, inactive control	n = 201 (MCI) Tx n = 122 Cntl n = 76 Age: Tx: $\bar{x}$ = 68.45 s = 6.99 Cntl: $\bar{x}$ = 66.86 s = 8.79	• 60, 90 min sessions, (3×/ week for 5 months) • Booster session to subset 11 months after training • Group-based, multicomponent training in cognitive strategies, cognitive stimulation, and psycho-therapeutic techniques	HVLT, RAVLT, RBMT, Digit Symbol, FUCAS, BNT, FRSSD, MMSE	• Significant values of experimental group: ↑ general cognitive performance (p = 0.000), abilities of attention (p < 0.001), language (p = 0.006), verbal memory (p = 0.000), executive function (p = 0.000), visual perception (p = 0.000) and activities of daily living (ADL) (p = 0.013) • Ctrl group: ↓ in observed ADLs (p = 0.004)	• ↑ cog performance and generalized benefit. • Trained group demonstrate improvement in verbal memory, visual-constructive abilities, and executive functions • MCI group show improvement in general cognitive performance, attention, language, verbal memory, executive functions, visual perception, and activities of daily living.	−0.026
25.	Valdes et al. (2012)	RCT, single-blind, inactive control	n = 195 (MCI mixed) Tx n = 85 Cntl n = 110 Age: Tx: $\bar{x}$ = 76.95 s = 6.53 Control $\bar{x}$ = 78.34 s = 6.3	• 10 sessions, 60 min/ group sessions (5-week duration) • Group-based, computer administered restorative training using a standardized set of visual attention tasks designed to improve speed of information processing (SOPT)	RBMT, RAVLT, RBMT, LS, WS, Computerized UFOV	• SOPT improved UFOV in MCI relative to controls (F1,185) = 81.83, p < 0.001 • All subtypes of MCI appear to benefit from SOPT relative to controls • aMCI subtype appeared to experience greatest benefit from SOPT • MCI who received SOPT show greater rate of improvement over 5 years relative to controls • There was no difference in the slope of change between MCI subtypes across the 5-year period and gains of SOPT were maintained across 5 years	• Individuals with MCI benefit from SOPT • All MCI subtypes appear to benefit from SOPT relative to controls • Individuals with single non-amnestic and multi-domain subtypes demonstrate the greatest immediate improvement • aMCI individuals show the least UFOC improvement from pre- to immediate post-training • Benefits of SOPT training in MCI remain relatively stable across 5-years	−0.026
26.	Vidovich et al. (2015)	RCT, single-blind, active control	n = 160 (MCI) Tx n = 80 Cntl n = 80 Age: Tx: $\bar{x}$ = 75.1 s = 6.1 Cntl: $\bar{x}$ = 74.9 s = 5.5	• 10, 90 min sessions (2×/ week for 5 weeks) • Group-based, multicomponent training in cognition, activities related to enhance attention, memory, and executive functions as well as methods to adapt these to everyday life	CAMCOG-R, CVLT-II, WAIS-III DS, WAIS-III SS, TMT-A/B, COWAT	• Training did not affect CAMCOG-R scores over time, relative to controls • Training demonstrated marginally better scores in DS-FWD • Training had no effect on any other outcome measure related to cognition	• With the exception of attention (DS), training showed no effect on cognition either immediately, one-year or two-years post training. • Illness progression and changes in cognitive status over time may limit findings	−0.026

Please see Supplemental Table S4 for a list of abbreviations