Long-term studies in cognitive training for older adults: a systematic review

ABSTRACT. Studies show that aging is accompanied by losses in cognitive functions and that interventions can increase performance and/or support the maintenance of cognitive skills in the elderly. Objective: The objective of this study was to carry out a systematic review of long-term studies involving cognitive training (CT) in older adults without dementia and/or with mild cognitive impairment (MCI). Methods: A systematic review of controlled studies was published in scientific journals from 2000 onward, with duration ≥6 months, CT intervention, cognitively normal (CN) or MCI participants aged ≥60 years, and assessments using cognitive and/or neuropsychological tests. Results: A total of 32 studies were reviewed, comprising 10 on study protocols, 14 in CN older adults (no MCI and/or dementia), and 8 in older adults with MCI or at risk for dementia. Conclusions: The studies reported improvements in cognitive performance for some motor abilities, among older participants of CT with or without booster sessions, including multimodal interventions or otherwise.


INTRODUCTION
A ccording to the estimates of the United Nations 2019 Revision of World Population Prospects 1 , 17.8% of people in the world will be above age 65 by 2060, up from 9.6% in 2021. In Brazil, this proportion will increase from 9.9 to 27%. During the normal cognitive aging process, the organism undergoes periods of stability and change. These changes partially stem from physiological and anatomical components. Most notable of the normal changes accompanying healthy aging are the aspects relating to the brain and cognitive functioning 2 , which may affect more complex everyday tasks, such as driving, paying bills, and remembering dates and appointments 3 .
Studies show that aging is accompanied by losses in cognitive functions and that interventions can promote performance gains and/or support the maintenance of cognitive abilities in healthy older persons 3,4 . According to the literature, the existing cognitive interventions, such as cognitive training (CT), have the potential for promoting health by optimizing cognitive and neural plasticity. These effects may be increased by combining CT with other types of interventions, such as physical activities and a balanced diet, with the aim of improving the performance of the individual and neural reorganization as a result of the intervention 5,6 . Both neural and cognitive plasticity are an inherent part of the life course of an individual. Although diminishing with age, plasticity supports the learning of mnemonic techniques, as well as the expansion and integration of knowledge related to cognitive functions 7 .
The effects of CT can extend to other domains, such as health promotion 8,9 , as well as functioning 10 . A plethora of cognitive modalities have been tested in both healthy and cognitively impaired elderly, displaying similar positive effects on cognitive performance and other variables, such as psychological well-being. According to Ngandu et al. 6 , intervention might not be too late for presymptomatic and predementia disease stages and also for at-risk states, such as in mild cognitive impairment (MCI). For example, recent studies by Peng et al. 11 , Valdés et al. 12 , Lee et al. 13 , and Djabelkhir et al. 14 . reported cognitive gains in older adults with MCI. The research by Lee et al. 13 indicated benefits even for mild dementia.
One of the first CT studies of the long-term type, i.e., lasting 6 consecutive months or longer, was the multicenter study conducted by Ball et al. 15 , involving a large number of participants: 2,832 subjects aged 65-94 years. Participants received one of the following types of cognitive intervention: (a) verbal episodic memory training (group 1), (b) logical reasoning training (group 2), (c) processing speed training (group 3), or (d) control group without training (group 4). A total of 10 training sessions were given and 60% of the sample received 4 booster sessions after 11 months. The results showed that 87% of participants from group 3 improved performance on processing speed, 74% from group 2 improved logical reasoning, and 26% from group 1 improved memory. Booster sessions were effective for maintaining processing speed and logical reasoning, but not for episodic memory. The positive effects of training were not detected during the daily routine of participants, but persisted after 2 years, suggesting that the intervention effects can remain consistent over the long term.
Ngandu et al. 6 , in a study involving a multidomain intervention of nutritional diet, physical exercise, CT, and vascular risk monitoring, reported this to be more effective and efficient, showing 20-150% of improvement in cognitive performance of participants. The Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) described the effects of this approach, showing that the intensity of the intervention, the target public, the type of approach, and the fact of being long-term training explained the beneficial effects for cognition observed in participants. These results reveal the importance of healthy dietary habits and regular physical exercise in conjunction with cognitive interventions.
Training studies differ not only in duration but also in strategies trained and methodology employed. Results reported in the literature varied widely regarding the strength of effects, generalization to untrained tasks, and long-term maintenance of improvements 16 . This heterogeneity justifies an analysis of the literature on long-term cognitive interventions with the aim of adding to the knowledge on CT and providing consistent, in-depth information on the cognitive benefits associated with long-term interventions and the strategies they employ. The objective of this study was to carry out a systematic review of long-term studies involving CT in older adults without dementia and/or with MCI.

METHODS
This systematic review had its protocol registered in the International Prospective Register of Systematic Reviews (PROSPERO) in April 2021 (submitted in February 2021) under registration number CRD42021239130. The protocol can be assessed at https://www.crd.york.ac.uk/prospero/display_record. php?ID=CRD42021239130.
Eligibility criteria were as follows: clinical trial studies with a 6-month duration or longer, intervention involving CT and a control group, cognitively normal (CN) or MCI participants aged ≥60 years, articles published in 2000 or later in scientific journals, and follow-up assessments of intervention effects using cognitive and/or neuropsychological tests. The criterion used for CT studies to be considered long term was to have a duration of 6 consecutive months or more.
Publications of master's dissertations, book chapters, doctoral theses, letters to the editor and case studies, studies whose samples included individuals aged <60 years or with dementia, studies performed at long-term care institutions, and studies failing to report the effects of intervention on cognitive performance were excluded.
The systematic review was conducted between February and April 2021. All manuscripts in Portuguese and English were revised for eligibility criteria. The Scielo, LILACS, and PubMed/MEDLINE scientific databases were searched using the following combinations of the key words: ((idoso OR idosos OR idosa OR idosas) OR (elder OR "older person" OR "older persons" OR "older people" OR "senior citizen" OR "senior citizens" OR elderly OR "aging people" OR "aging person" OR "aging persons")) AND ("treino cognitivo" OR "cognitive training") AND ("longa duração" OR "long term" OR longitudinal OR "follow up") AND (envelhecimento OR aging).
To guide the stages of identification, screening, and eligibility of studies, two pairs of reviewers working independently screened all records retrieved, following the steps of the Statement of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 17 . The initial identification of studies was performed by searching the abovementioned databases. In the screening stage, duplicate studies were excluded, and titles and abstracts were read for the first selection, according to the preestablished inclusion and exclusion criteria. In the eligibility stage, the remaining studies were read in full in order to be selected according to the same criteria. The remaining studies after this stage were the studies included in the review.
The following data were extracted from the articles: study title, authors' name, summary, results, methods, justification, objectives, and conclusion. Excel spreadsheets were used as support tools in this process.
The studies were assessed for quality according to Downs and Black's 18 checklist. This assessment tool consists of 27 questions, which are divided into 5 subscales: report or assessment of adequate information (10 items), external validity (3 items), internal validity of detailed measurements and result bias (7 items), confounding factors (6 items), and power (1 item). Each item that makes up the checklist assigns a score from 0 to 1, except for the item that assesses the description of confounding factors, which can assign up to 2 points, and the item that assesses the description of the study's power (27), which originally assigned from 0 to 5 points, but was modified to assign from 0 to 1 point, as in other studies [19][20][21] so that a score of 1 was given if the article presented power calculation and/or sample size calculation and a score of 0 if it did not present any of these calculations. Thus, the checklist has total scores ranging from 0 to 28 points. For a better understanding of the data obtained, the score was converted into a percentage for each domain, and a final average of the total score of all domains was calculated. Next, the quality of the articles was classified as follows: up to 0.39 was considered bad, 0.40-0.69 considered regular, 0.70-0.79 considered good, and 0.80 or above was considered excellent.

RESULTS
The initial search led to the retrieval of 83 studies, of which 1 was subsequently excluded owing to duplication. Titles and abstracts of the remaining 82 studies were read and screened for relevance to the review topic. After applying inclusion and exclusion criteria, a total of 27 studies were excluded. Thus, 56 articles were read in full; of these, 24 studies that did not meet the eligibility criteria were excluded. The process of study selection for inclusion in the review is shown in Figure 1. The final 32 studies included in the review for analysis are listed in Tables 1-3.
Eight of the study protocols [22][23][24][25][26][27][28][29] had innovative methods and objectives, which, in general, sought to investigate the effects of multiple domain interventions focusing on a range of aspects, such as preventing cognitive impairment, cognitive functions, physical fitness, activities of daily living (ADL), quality of life, gait speed, incidence of falls, and executive functions. For these studies, participants were categorized into CN older adults, elderly at risk for cognitive impairment, subjects with aging-related cognitive impairment, and older patients diagnosed with MCI. Interventions used a variety of resources, such as tablets, computers, physical training (PT), walking, health advice, software such as Fit Brains and Tonic and Phasic Alertness Training (TAPAT), electrostimulation, and vitamins.
Two other studies included in this review were publications of study protocols. Jobe et al. 31 presented the design of the long-term Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) study, based on a sample of 2,802 older adults. The participants were randomized into a control group, speed of processing training (SOPT) group, a reasoning training group, and a memory group. The intervention consisted of 10 sessions of 60-75 min over a period of 56 weeks, plus booster sessions for 11 months after the primary training. To determine the long-term effects, assessments were carried out after 1 and 2 years.
Similarly, Rebok et al. 10 applied booster sessions for 35 months after the primary intervention, with data collected 3, 5, and 10 years after the pre-intervention assessment. Due to the magnitude of the ACTIVE study, many studies recruited its data, of which 11 were selected for inclusion in this review.
The study by Kivipelto et al. 30 described the protocol of the FINGER study, with 1,200 older persons at risk for cognitive impairment. Participants were randomized into two groups. One group received CT combined with nutritional counseling, PT, social activity, and management of metabolic and vascular risk factors, for two 6-month periods, three times a week, totaling 72 sessions of 10-15 min each. The other group consisted of control and was given regular health advice. Long-term effects were to be measured by assessments planned 1 and 2 years after the intervention. One of the studies derived from the FINGER was also included in this review.

Complete clinical trials
Of the 22 studies whose results were analyzed in this review, 14 involved samples comprising CN older adults, i.e., without MCI and/or early dementia 10,32-44 , whereas 8 involved subjects with MCI or at risk for dementia 6,45-51 .
Ten 10,37-44,51 of 22 studies drew on data from the ACTIVE study. Besides, a number of studies included employed multimodal interventions, also referred to as multifactorial, namely, FINGER, with nutritional interventions based on a specific diet, physical fitness training programs, and cognitive interventions, such as CT, and also with vascular risk monitoring 6,30 ; AgeWell, which includes nutritional counseling, physical activity, CT, optimization of medication, management of vascular risk factors, social activity, and further specific interventions targeting grief and depression 33 ; other studies whose interventions include combined CT and PT 22,23,28,32,34,35 ; and a study with training of memory, reasoning, problem resolution strategies, visuospatial map reading skills, and handicraft making 34 .

Objectives
The objectives of the studies varied widely, in which those involving CN subjects tended to investigate the long-term effects of programs for SOPT, reasoning and/or episodic memory on everyday functioning and cognition 10,41,43 , on the trajectory of cognitive aging 42 , increase in cognitive function of older people 33 , impact of CT on objective measures of physical functioning 38 , on use of cognitive strategies 40,44 , on initial recall and  To determine the effects of cognitive training on daily function and durability of training on cognitive abilities.
Ten-session training for memory (verbal episodic memory), reasoning (inductive reasoning), or speed of processing (visual search and identification); four-session booster training at 11 and 35 months after training in a random sample of those who completed training.
Reasoning training resulted in less functional decline in self-reported IADL. CT cognitive training improved cognitive abilities specific to the abilities trained that continued 5 years after initiation of intervention. 24 Gross Participants were randomly assigned to one of four groups: 10-session group training for memory (verbal episodic memory; n=711), or reasoning (ability to solve problems that follow a serial pattern; n=705), or speed of processing (visual search and identification; n=712); or a no-contact control group (n=704). For the three treatment groups, four-session booster training was offered to a 60% random sample 11 months later.
Initial SOPT effects were maintained over 5 years and amplified by booster sessions.
A single booster session counteracted 4.92 months of agerelated processing speed decline. To investigate the influence of memory training on initial recall and learning.
Each ACTIVE intervention was administered in 10 small-group training sessions, each lasting 60-75 min, offered over a course of 10 weeks. The first of 10 sessions provided didactic training on how memory works and how to maximize benefits of training.
Memory strategy training was associated with significant longterm gains in learning, stemming from both the highly significant effect of the training and from a slower decline, for up to 5 years, in memory span. To determine the influence of CT in the ACTIVE study on the pace of cognitive aging.
Briefly, older adults (aged 65-94) were randomly assigned to one of the three cognitive training or no contact control arms. Training lasted 5-6 weeks, and participants were assessed pre-and postintervention, and at 1, 2, 3, 5, and 10 years after post-test. This analysis considers outcomes through 5 years, as the 10-year main results are currently under analysis.
Reasoning training attenuated aging-related training. Memory gains were maintained but about half of reasoning and speed gains were lost.
All trained groups performed better than controls at 5 years. Performance differences at end of follow-up were equivalent to about 6, 4, and 8 years of aging for memory, reasoning, and speed training, respectively.
24 Table 1. Long-term studies with cognitively normal older adults.
Continue... To examine the short-and long-term effects of a cognitive training (CT) program in enhancing cognitive function of older people with subjective memory complaints.
A single-blind randomized placebocontrolled trial was carried out in a sample of 223 older adults aged 65 years or above with subjective memory complaints in Hong Kong. They were randomly assigned to either receive CT (intervention group, n=111) or attend health-related educational lectures only (control group, n=112). Participants' cognitive abilities were assessed by the Chinese version of Mattis Dementia Rating Scale at baseline, immediately after the training, and 9 months after the training.
Cognitive training was effective in enhancing the overall cognitive functioning of less educated older adults with subjective memory complaints. The positive effect was durable for at least 9 months in conceptualization and memory. To investigate how a multicomponent memory intervention affected memory for prose.
Participants were randomized into one of the three training arms (i.e., memory, reasoning, and speed of processing) or a no-contact control group; about half of the trained participants received additional booster training 1 and 3 years post intervention.
Multi-factorial memory training can improve verbatim recall for prose, but the effect does not last without continued intervention. To analyze the short-and longterm effects of PT, combined CT, and PT plus CT programs on age-sensitive fluid cognitive abilities. 70 healthy senior citizens (age 60-75) were allocated to a physical, cognitive, combined physical plus cognitive, and waiting control group. The trial assessed information processing speed, short-term memory, spatial relations, concentration, reasoning, and cognitive speed.
Physical, cognitive, and combined physical plus cognitive activity can be seen as cognition-enrichment behaviors in healthy older adults that show different rather than equal intervention effects. To determine the effects of cognitive training on cognitive abilities and everyday function over 10 years.
Ten training sessions for memory, reasoning, or speed of processing; four sessions of booster training 11 and 35 months after initial training. learning 37 , on memory for prose 39 , on performance of working memory (WM), and on untrained tasks 36 . Studies with interventions involving combined CT and PT investigated the long-term effects on cognition 32 and fluid cognitive abilities sensitive to age 35 . Finally, the study by Li et al. 34 examined the relationship between changes in brain activity and cognitive performance after a multimodal or multifactorial intervention. The studies with samples comprising older adults with MCI focused on investigating the longitudinal effects of SOPT on processing speed 51   To examine the relationship between changes in spontaneous brain activity and cognitive performance that occur after CT.
Participants were trained for 1 h, twice a week, for 12 weeks. Cognition was assessed in all participants and magnetic resonance images were obtained at baseline and 1 year after training. To assess spontaneous fluctuations in brain activity, we acquired resting-state fMRI data. Two indices-functional entropy and time-domain entropy-were used to measure the effects of training. Functional entropy increases with aging and indicates disruptions in functional connectivity. Time-domain entropy decreases with aging and indicates structural alterations in the brain and blood-flow reduction. To assess the impact of three CT programs on objective measures of physical functioning across 5 years.
Older adults randomized into a processing speed (n=702), reasoning (n=694), or memory (n=703) training intervention were compared to those randomized into a no-contact control condition (n=698). Intention-to-treat (ITT) and treatmentreceived (time-varying number of training sessions) analyses were conducted. To evaluate the extent to which CCT benefits older adults with MCI and MrNPS and examine its effects on meta-cognitive and noncognitive outcomes.
Home-based CCT with adaptive difficulty and personal tailoring appears superior to more generic CCT in relation to both cognitive and non-cognitive outcomes.
23 Table 2. Long-term studies with older adults with MCI or risk for dementia.
Continue... types of intervention: a program using episodic memory coding strategies 48 , a neuropsychological training (NPT) program in patients treated with cholinesterase inhibitors (ChEIs) 49 , a creative expression therapy (CrExp) 45 , CT plus psychosocial intervention 47 , a computerized cognitive training (CCT) program 46 , and a program of simulated functional tasks (FcTSim) 50 .
With regard to the study in older adults with some cognitive impairment but no MCI diagnosis, the objective was to investigate the longitudinal effects on cognitive functions of a multidomain intervention, combined CT, diet, physical exercise, and cardiovascular risk monitoring 6 .

Main interventions with cognitively normal older adults
Of the 14 studies conducted in CN samples, 4 studies 10,38,41,42 analyzed data from the three intervention groups of the ACTIVE study 31 . Based on the protocol of this study, Kwok et al. 33 conducted a trial in 223 older people with subjective cognitive complaints who received an intervention of 12 sessions of 90 min given weekly, also entailing SOPT, besides memory and reasoning training.
Five studies involving CN elderly were specific interventions for training memory. Three studies analyzed data from memory, training, and control groups in the ACTIVE study, involving a total of 1,401 participants 37,40,44 . The studies of by Gross and Rebok 40 and Gross et al. 44 assessed the impact of the memory training program from the ACTIVE study on the use of strategies. In a specific learning intervention involving the memorizing of short stories, the study by Sisco et al. 39 assessed the impacts of the memory training program from the ACTIVE study in conjunction with the booster intervention, including a total of 1,902 participants 39 . Borella et al. 36 carried out WM training with 36 older adults who underwent a 2-week intervention of 60-min sessions.    The SYNERGIC Trial established the efficacy and feasibility of a multimodal intervention to improve cognitive performance and mobility outcomes in MCI. To determine whether a combination of PT and CT has greater effects on walking speed, dual-task cost in walking speed, fall incidence, and executive functions compared to PT alone.
When completed, this study will provide new knowledge on the effects of physical and cognitive training on the prevention of walking limitations and rate of falls in older people. The expected results will be of value in informing strategies designed to promote safe walking among older people and may have a significant health and socioeconomic impact. To examine the effect of a CCT program, alone and preceded by a brisk walk, on cognitive function and explore the underlying neural mechanism in communitydwelling older adults.
Eight weeks sessions, three times week for 1 h+3 three times 1-h session at home; study groups: If results from this study show benefits for cognition at trial completion, CCT programs, alone or in combination with walking, might be a strategy to promote healthy cognitive aging in older adults. In addition, results from the 1-year followup measurement could provide important information regarding the long-term benefits of these CCT programs. To compare the effect of broad and directed (narrow) technology-based training on basic perceptual and cognitive abilities in older adults and on the performance of simulated tasks of daily living including driving and fraud avoidance.
Web-based brain game suite (Brain HQ) and strategy video game (Rise of Nations) or to directed training for IADL training using web-based programs for both driving and fraud avoidance training, active control; assessments: pre-and post-test; follow-up: 1 year after training; tests: ability tests of IADL (driving simulator test for hazard perception, and a financial fraud recognition test), UFOV, DSST, RAPM, Letter sets, HVLT, RAVLT, and UMCFAB.
The baseline results support that randomization was successful across the intervention conditions.  The study by Ball et al. 43 analyzed the data from the primary SOPT program plus booster sessions of the ACTIVE study, in which 1,400 and 633 older adults participated, respectively, including control groups.
A further two studies in CN older adults applied interventions involving combined CT plus PT 32,35 . The study by Linde et al. 35 , involving 70 senior citizens, applied an intervention comprising weekly 30-to 90-min sessions of PT, CT, and combined PT plus CT, given over a 16-week period. Eggenberger et al. 32 conducted an intervention with 89 older adults comprising 52 sessions for 1 h, given twice weekly over 26 weeks, consisting of a virtual reality (VR) videogame dancing and treadmill walking with and without simultaneous verbal memory training.
Finally, Li et al. 34 recruited 270 CN older persons and performed a CT intervention for 12 weeks, twice a week with 1-h sessions, consisting of training of memory, reasoning, problem resolution strategies, visuospatial map reading skills, and production of handcraft.
Main interventions with older adults with mild cognitive impairment or at risk for dementia Seven of the eight studies with cognitively impaired individuals examined the effects of a variety of forms of CT in older adults with MCI 45-51 and one in participants at risk for dementia, but not diagnosed with MCI 6 .
Among the investigations in samples of older subjects with MCI, the study by Valdes et al. 51 assessed the data of 1,298 participants of the SOPT and control groups of the ACTIVE study. Belleville et al. 47 carried out an 8-week intervention of eight 2-h sessions. A total of 145 older adults participated in memory training, psychosocial intervention, and a control group. A booster session of the same duration was performed 3 months after the intervention. In the study by Rojas et al. 48 , 46 MCI participants were randomized and the intervention group participated in a 6-month intervention in sessions of 2 h, twice weekly, of CT and cognitive stimulation, including episodic memory training and executive control training techniques. Rozzini et al. 49 carried out an intervention with 59 participants and 20 sessions of 1 h, five times a week. The investigation groups were divided into a group receiving ChEIs only, a group receiving ChEIs plus NPT (software training memory, language, attention, abstract reasoning, and visuospatial abilities), and a control group. In their study, Bahar-Fuchs et al. 46 applied an intervention lasting 8-12 weeks with two sessions of 20-30 min/ day, three times a week. Notably, 68 participants were randomized into CCT and active control conditions. In the study by Law et al. 50 , 83 older adults were randomized to receive 13 sessions of functional task exercises (FcTSim) with PT or to an active CT group for 10 weeks. In the study by Zhao et al. 45 , an intervention was performed, comprising 25 sessions of 1 h each over 16 weeks with 93 participants, who were randomized into a CrExp group or a standard CT control group.
With regard to the study in older adults at risk for developing dementia 6 , a multidomain intervention combining PT, diet, cardiovascular risk monitoring, and CT with the use of technology was performed. The authors carried out the intervention based on the protocol of the FINGER study 30 .

Risk of bias
Regarding the categories of the methodological quality checklist, according to the scores, no articles obtained less than 0.71 points and 23 reached more than 0.80, which corresponds to a high score, excellent quality in the studies analyzed, and low risk of bias. The total average of articles for all categories was 0.84 out of a total of 1.0, meeting the methodological quality requirements of Downs and Black 18 . When divided by domains, the scores achieved were as follows: report 0.84, external validity 0.77, internal validity 0.74, confusion 1.0, and power 0.75 (Table 4).

DISCUSSION
The aim of this study was to carry out a systematic review of studies investigating the long-term effects of CT programs in older adults without dementia. A total of 32 studies were reviewed, comprising 14 in CN older adults, 8 in older adults with MCI and/or at risk of developing dementia, and 10 on study protocols.
Cognitive training long-term studies with cognitively normal older adults With regard to the CT long-term studies with CN older adults, and among the studies analyzing data from the three intervention groups of the ACTIVE study, Willis et al. 41 and Rebok et al. 10 . found less decline in self-reported IADL, particularly in the reasoning training group 41 . Regarding this finding in the functionality of the elderly, Carvalho et al. 52 highlighted the importance of cognitive interventions, with an emphasis on memory training, to improve the performance on mnemonic tasks.
The results from the studies by Willis et al. 41 and Rebok et al. 10 . also showed improvements in trained cognitive abilities, with benefits sustained for 5 years after the start of the intervention in all three CT groups 41 and for 10 years in the reasoning training and SOPT groups. 10 . The study by Rebok et al. 10 also revealed a medium-to-large effect of the SOPT on processing speed after 10 years. This finding, which represents highly effective maintenance of gains of this type of CT over time, highlights the importance of performing longer follow-up of the effects of CT interventions. In previous year, the study by Ball et al. 43 , also derived from the ACTIVE study, confirmed the maintenance of positive effects of SOPT for 5 years after the intervention, corroborating the results of the study by Willis et al. 41 .
The study by Jones et al. 42 , who drew data from the ACTIVE study, except those related to the booster intervention, showed that memory gains were maintained for 5 years, akin to Willis et al. 41 and Ball et al. 43 , and that reasoning training significantly attenuated the aging-related limitations in this cognitive ability. The results of the study by Li et al. 34 , included in this review, were consistent with this finding. The authors reported, based on neuroimaging analyses, that CT can promote plastic gains in intrinsic activity patterns, particularly through improvements in functional connectivity and in brain structure which, according to the researchers, are probably part of the neural mechanisms underlying the effects of CT. In other words, according to these findings, CT can slow the pace of cognitive aging.
The results of the study by Ross et al. 38 , however, showed that more training sessions in all intervention groups of the ACTIVE study, i.e., SOPT, memory, and reasoning training, enhanced the performance on tests evaluating fine motor coordination (abilities such as drawing and painting) and gross motor coordination (running, jumping, and walking up and down the stairs), visuomotor coordination (observe, recognize, and use of visual information on shapes, figures, and objects), and also motor speed. However, these results suggested that greater training on reasoning increased hand-grip strength, closely associated with ADL. Effects on cognitive-motor abilities were also observed by Theill et al. 53 , although in this case through simultaneous performance of PT and CT, as opposed to the use of CT alone by Ross et al. 38 . Theill et al. 53 found that simultaneous training can promote specific improvement in both cognitive performance and dual-task motor-cognitive performance, providing greater potential for performing ADLs.
In a study based on the ACTIVE protocol, Kwok et al. 33 showed an improvement in general cognitive functioning of low-educated individuals, with effects maintained for at least 9 months in the cognitive areas of conceptualization and memory. The authors proposed this finding might be explained by the ceiling effect, i.e., a tendency of the CT to promote greater gains among subjects with below normal cognition prior to the training and in individuals who received no simultaneous training. In contrast, the results of the study by Teixeira-Fabrício et al. 54 showed that a higher educational level can lead to greater use of strategies, higher self-efficacy for memory, and larger performance gain post-training. In addition, Casemiro et al. 55 emphasized that greater education can be directly associated with ease of learning. The researchers reported that high-educated individuals perform visual search tasks more effectively than subjects with a lower educational level.
The studies by Gross and Rebok 40 and Gross et al. 44 . assessed the impact of the memory training program from the ACTIVE study on the use of strategies. The results of the first of these two studies 40 indicated that memory training improved the levels of use of strategies and can assist older adults who deploy them in appropriate situations. The authors reported that the effects of training persisted for up to 5 years and that strategies are positively associated with memory performance and daily functioning. The results of the study by Gross et al. 44 suggested that the method of loci (MoL) (post-training) was used by up to 25% of older adults and immediately improved memory, with effects sustained throughout the follow-up period. These results corroborate the notion that a balance occurs between complexity and novelty in strategy selection by the elderly and that the memory training produces, by promoting changes in the strategies used, observable qualitative and quantitative differences in memory performance. Other studies assessing the effects of memory training on the use of strategies in CN older persons are also available in the literature. In contrast to the findings of Gross and Rebok 40 and Gross et al. 44 , the results of the study by Yassuda et al. 16 suggested that the more intense use of memory strategies resulting from training does not necessarily guarantee better performance. Carvalho et al. 52 , however, showed that categorization strategy training led to greater use of the trained strategy and significantly improved the performance on the episodic memory task.
Another study by Gross et al. 37 , also derived from the ACTIVE study, reported an association of memory training with significant long-term gains in learning, stemming from both the highly significant training effect and slower memory decline for up to 5 years. The study by Sisco et al. 9 , which assessed the impacts of the memory training program from the ACTIVE study in conjunction with the booster intervention, suggested in their results that, when carried out in a multifactorial manner together with the booster intervention, the training can improve literal recall for stories.
Borella et al. 36 showed in their study that WM training produced benefits that were maintained over time. The authors suggested that these findings confirmed there is still room for plasticity in the basic mechanisms of cognition in old age, congruent with other studies addressing CT in CN older adults in which these subjects were able to attain a level of current performance closer to their maximum possible performance 56 .
Ball et al. 43 reported that positive initial SOPT effects were amplified by booster sessions. According to these authors, a single booster session counteracted around 5 months of age-related processing speed decline. In line with this finding, the results of the study by Aramaki and Yassuda 57 showed that, besides stability in participants' cognitive performance between the two interventions, additional gains on episodic memory scales were observed after the booster intervention.
Linde et al. 35 revealed in the results of their study that the three types of activities carried out by the participants, i.e., PT, CT, and combined PT plus CT, can be seen as cognition-enrichment behaviors. Eggenberger et al. 32 reported that particular executive functions benefited from simultaneous CT and PT compared to exclusively physical multicomponent training, concluding that cognitive-physical training programs may counteract widespread cognitive impairments in the elderly. These findings are consistent with the recent study by McEwen et al. 58 , who carried out an intervention of simultaneous aerobic exercise and memory training and found that the intervention promoted improvements in memory, attention, and reasoning abilities.
Long-term studies on cognitive training in older adults with mild cognitive impairment or at risk for dementia Of the eight studies involving cognitively impaired individuals, seven examined the longitudinal effects of a variety of forms of CT in older adults with MCI 45-51 , while one was a multidomain intervention in elderly people at risk for dementia, but not diagnosed with MCI 6 .
The study by Valdes et al. 51 revealed that all MCI groups showed an immediate improvement relative to the control group, with an emphasis on the non-amnestic MCI group, in which no significant changes were observed during the 5-year follow-up. Belleville et al. 47 . suggested in their results an improvement on the memory task and strategy use in everyday life of participants of the CT. Consistent with the findings of both studies 47,51 , the study by Olchik 59 , in which older persons with MCI performed memory training, reported that CT can benefit participants in terms of acquisition of strategies for coping with and overcoming cognitive impairment, and even reverse MCI, allowing these individuals to attain a similar level of performance to CN subjects. In addition, the authors believe this training modality represents a cost-effective viable educational intervention that can benefit older persons with MCI.
Rojas et al. 48 reported that CT in individuals with MCI can also represent a promising treatment option for optimizing performance, preventing cognitive decline, or delaying progression to dementia in this patient group. Brum et al. 60 noted that CT in older adults with MCI constitutes a non-pharmacological alternative for preventing cognitive and functional decline and for promoting improvement in cognitive performance.
In the study by Rozzini et al. 49 , participants who received ChEIs plus NPT showed significant improvements in cognitive areas and in behavioral disturbances, confirming that a long-term NPT in ChEIs-treated MCI subjects induces additional cognitive and mood benefits. These results are in line with the findings of Olazarán et al. 61 , who reported that patients with MCI, mild Alzheimer's disease (AD), or moderate AD treated with ChEIs and undergoing a long-term cognitive-motor intervention had greater mood and cognitive benefits compared to the control group.
Bahar-Fuchs et al. 46 showed in their study that unsupervised home-based CCT with individual tailoring can lead to cognitive and non-cognitive benefits in older adults with MCI. Consistent with these results, the study by Hill et al. 62 in older adults with MCI and dementia revealed the efficacy of CCT on global cognition, selected cognitive domains, and psychosocial functioning of individuals with MCI.
In the study by Law et al. 50 , the results showed that the FcTSim promoted improvements in general cognitive functions, particularly executive function and problem-solving ability, thereby serving as a cost-effective way of promoting brain plasticity, even in patients with MCI. These findings are consistent with the study by Liao et al. 63 , who randomized older adults with MCI into either a VR-based PT with CT group or a combined PT and CT group without VR. Results showed that the VR group improved global cognition, while both groups improved executive function and verbal memory.
Zhao et al. 45 suggested in their study that the CrExp therapy promoted greater gains in general cognitive functioning, memory, executive functions, functional status, and everyday living ability among patients receiving the therapy compared to participants receiving standard CT. The authors reported that improvements were maintained at the 6-month follow-up and concluded that this therapy may serve as a cost-effective adjunct to standard interventions for older adults with MCI.
Law et al. 50 and Zhao et al. 45 showed that these interventions can serve as cost-effective strategies for older adults with MCI, satisfying the premises of the World Health Organization (WHO), which holds that CT should be provided and applied to both CN older adults and individuals with MCI as a preventive action for cognitive decline and development of dementia, irrespective of social class 64 . This is also guaranteed by the Active Aging policy, which highlights the necessity of incentive for care and development of cognitive abilities to maintain the autonomy of an individual 64 .
With regard to the study in older adults at risk for developing dementia but not diagnosed with MCI 6 , results showed that multidomain intervention can improve cognitive functioning in older adults at risk of cognitive decline 6 .
To sum up, the studies reviewed reported a number of cognitive performance benefits, including a role in improving some motor abilities, among older adults without dementia who participated in CT programs with or without booster sessions and who received multimodal interventions or otherwise.
A total of 14 long-term studies were gathered in which cognitively healthy elderly people, without any type of cognitive impairment, were followed up. In view of this, it was possible to report several cognitive performance benefits. Furthermore, the results of these studies documented that such cognitive benefits lasted up to 5 years after starting the intervention. Eight long-term studies were also gathered in which elderly with MCI or at risk for dementia were followed up. Studies have indicated significant sustained improvements in general cognitive function, executive function, and problem-solving ability, in addition to an increase in brain plasticity. Furthermore, it has also been observed that computerized cognitive interventions at the MCI can prevent cognitive decline or slow conversion to dementia. Finally, 10 publications of protocols were analyzed, studies that will describe their methods and plans. Among them, one protocol has demonstrated the potential to significantly improve efforts to ameliorate cognitive decline, providing important information about the feasibility and intervention effects of a combination of exercise and CT for older adults with MCI.
It is important to highlight that the studies with methodology models included in this review were mostly interventions characterized as multicomponent cognitive stimulation and allow the replication of their methods to other research centers, to verify in a contemporary way to the original authors, if the models of proposed interventions can generate cognitive gains in healthy elderly and in elderly people with MCI.
Therefore, different types of CT programs appear to represent highly applicable cost-effective strategies for promoting health and quality of life in older age. There were a vast number of studies addressing the theme and a wide variety of objectives related to the specific subthemes, with consequent heterogeneity in study results. Generally, however, all findings showed positive effects on the cognition of participants.
The limitations of this study included the selection and inclusion of multimodal CT research; the comparison of CT studies whose participants were CN elderly with studies in which older adults with MCI participated; and the citing of cognitive improvements measured using cognitive screening tests as opposed to more specific tests, such as neuropsychological assessment scales.
As presented in this article, some of the studies employed original, innovative methods incorporating a long-term follow-up. Thus, the methodology of these studies should be replicated in different cultures, given some have been published without results, providing fertile ground for future studies. It is also suggested to carry out future systematic review studies of CT only with a focus on multimodal studies and with samples focused on CN elderly and older adults with MCI.