The term Cognitive Reserve (CR) refers to a set of factors and mentally stimulating activities which may protect against cognitive impairment and explain inter-individual differences in the relationship between brain damage and clinical outcomes (Stern et al., 2019). CR is evaluated by means of single proxies (i.e., years of education, occupational status, or leisure activities), or composite measures and questionnaires, such as the Cognitive Reserve Index Questionnaire (CRIq, Nucci et al., 2012) or the Cognitive Reserve Scale (Altieri et al., 2018), simultaneously evaluating multiple proxies (for a review on available psychometric scales see Kartschmit et al., 2019).

The positive effect of cognitive reserve in clinical and non-clinical populations

CR is thought to exert a protective role in several clinical and non-clinical populations. It could contribute to inter-individual variability in severity of cognitive deficits and disability after stroke (Umarova et al., 2019) and mild traumatic brain injury (Stenberg et al., 2020). In multiple sclerosis, high levels of CR relate to better cognitive performance in memory, attention, and executive tasks (Santangelo et al., 2019a, 2019b; Santangelo et al., 2019a, 2019b). In psychopathological and psychiatric conditions, CR is associated with delayed onset and reduced symptom severity in patients with schizophrenia (Herrero et al., 2020), contributes to the course and prognosis of patients with bipolar disorder (Lin et al., 2020), and moderates the effect of major depressive disorder on neurocognitive test performance (Venezia et al., 2018). Moreover, is well established that normal aging is characterized by decline in cognitive processing speed, memory and reasoning abilities which start in midlife and accelerate with age (Ghisletta et al., 2019; Salthouse, 2019; Schaie, 1994). According to Baker et al. (2017), CR could represent a protective factor against physiological cognitive decline, despite the age-associated reduction in white matter fiber bundle length observed in aging.

Cognitive reserve and reduction of risk of developing dementia

CR could be associated with a reduced risk of developing dementia (Stern, 2012). As life expectancy increases and demographic ageing occurs, the global prevalence of dementia is expected to continue to rise. In UK, epidemiologic models suggest a 57% increase in the number of people with dementia from 2016 to 2040, such that by 2040 there will be up to 2 million people with dementia (Ahmadi-Abhari et al., 2017). Similarly, according to 2021 Alzheimer’s disease facts and figures, in the United States the number of patients with Alzheimer Disease might increase from 6.2 million to 13.8 million by 2060 (2021 Alzheimer’s Disease Facts and Figures, 2021). Worldwide the number of dementia patients is projected to reach around 152 million by mid-century (Alzheimer’s Disease International, 2018). In this context, healthcare services and activities could target cognitively unimpaired individuals, with a main focus on risk profiling and personalized risk reduction interventions rather than diagnosing and treating clinically-evident disease (Livingston et al., 2020; Panico et al., 2020; Ranson et al., 2021; Zhang et al., 2021). Focusing on modifiable risk factors could theoretically prevent or delay onset of dementia (Livingston et al., 2020).

The importance of studying the relationship between cognitive reserve and cognition in adulthood

A first crucial step in the process of dementia profiling and prevention is the timely assessment of modifiable risk factors (Ranson et al., 2021) and the understanding of the relationship between CR and cognition in adulthood. Modifiable risk factors include education and lifestyles habits which affect the relative risk for dementia in later life (> 65 years; Livingston et al., 2020), but are built during adulthood (18–65 years; Livingston et al., 2020). Recent studies suggest that in adulthood some proxies of CR -such as education, vocabulary knowledge, and scores on standardized scales- may positively affect attentional and memory domains (Cansino et al., 2020; Corral et al., 2006; Puccioni & Vallesi, 2012; Roldán-Tapia et al., 2012, 2017). Conversely, in healthy individuals low levels of CR are associated with psychological symptoms such as apathy, which in its turn is considered a risk factor for cognitive decline (Altieri et al., 2020), and with higher levels of perceived stress during challenging times (Panico et al., 2022). Furthermore, education -just one of the proxies of CR, usually acquired during the first decades of life- may contribute to individual differences in cognitive skills from adulthood to old age, and can provide long-lasting cognitive advantages, even delaying the age at which the threshold for cognitive impairments is reached (for a review see Lövdén et al., 2020).

This evidence paves the way for a comprehensive analysis of the relationship between CR and cognition in healthy adult individuals. To date, the literature on this issue is sparse and inconsistent due to the heterogeneity of the targeted cognitive processes and of the measures used to assess CR. To clarify whether the protecting role of CR on cognition is evident during adulthood would provide a rationale for a primary prevention strategy against cognitive decline. Indeed, young individuals may be encouraged to increase their CR throughout their life by means of timely multi-domain lifestyle intervention, which may include intellectual, physical, and social activities (Hertzog et al., 2008; Kivipelto et al., 2020; Rosenberg et al., 2018; Zülke et al., 2019), thus buffering against adverse effects of physiological aging and brain injury.

Study objectives

The aim of the present paper is to summarize the status of research on the relationship between CR and cognitive functions during adulthood. To this purpose, we performed a systematic analysis of the literature from the main scientific databases. We included the studies assessing one or more proxies of CR, composite indexes and standardized scales for measuring CR, and adopting neuropsychological and experimental tasks for the evaluation of cognitive functioning. For each study we considered the experimental design, reported the main results and discussed strengths and limitations for clarifying the issue at hand.


The review has been registered to the international prospective register of systematic reviews (PROSPERO) with registration number CRD42021233427.

Search strategy

We conducted a systematic literature review on PubMed, PsycINFO and Scopus databases until January 2021 following the PRISMA guidelines (Liberati et al., 2009; Moher et al., 2009). We used the following keywords: cognitive reserve AND cognit* AND healthy (OR normal); cognitive reserve AND attent* AND healthy (OR normal); cognitive reserve AND memory AND healthy; cognitive reserve AND frontal AND healthy (OR normal); cognitive reserve AND executive AND healthy; cognitive reserve AND psychomotor speed AND healthy (OR normal); cognitive reserve AND inhibitory control AND healthy (OR normal); cognitive reserve AND verbal fluency AND healthy (OR normal); cognitive reserve AND visuospatial AND healthy (OR normal); cognitive reserve AND visual AND healthy (OR normal); cognitive reserve AND spatial AND healthy (OR normal); cognitive reserve AND constructional AND healthy (OR normal); cognitive reserve AND healthy (OR normal). We used Zotero software (version 5.0.92) to retrieve the studies and remove the duplicates.

Inclusion criteria

We considered the studies eligible for our review if they: (a) were written in English; (b) included healthy individuals (18–65 years; Livingston et al., 2020), and in any case included a sample whose mean age was below 60 years; (c) used proxies of CR, such as education, occupational attainment, vocabulary knowledge, composite indexes, or specific assessment scales; (d) reported scores on at least one cognitive test assessing at least one of the following aspects of cognition: general cognitive functioning, attention, inhibitory control, verbal fluency, constructional abilities, verbal and non-verbal memory. These aspects of cognition were considered as covering the main cognitive domains described in neuropsychological literature (Lezak et al., 2004). We planned to exclude a) book chapters and theses, b) review papers, c) neuroimaging studies not reporting behavioral data.

Data extraction

We extracted the following data from the selected articles: authors, inclusion and exclusion criteria, sample characteristics, study design, methods used for assessing CR and cognitive functions, and results.

Study selection

The study selection process is described in Fig. 1 (Moher et al., 2009). Among the 8259 articles from the primary search, we excluded 6801 articles after checking for duplicates, 1217 following title checking and other 228 after abstract and text reading. The final study sample included 13 papers. Two reviewers (FP and LS) coded each outcome measure into one of the selected cognitive domains; in case of inconsistency, a third reviewer with specific expertise on neuropsychological assessment was involved (LT or GS).

Fig. 1
figure 1

The PRISMA Flow Diagram (Moher et al., 2009)

Quality assessment

We evaluated quality of the selected studies by using a modified version of the Newcastle–Ottawa Scale (Wells et al., 2011; see supplementary material 1) which considers appropriateness of the recruitment strategy, representativeness of the sample, outcomes and statistical analyses. Two independent reviewers (LS and AM) assessed the risk of bias of each primary study, and a third examiner (FP) was involved in case of disagreement. The scale score ranged 0–10 points, from unsatisfactory to very good studies.


Table 1 describes the main features of the studies included in the analysis. Among the selected studies, eight articles were correlational, i.e. searched for correlations between CR and performance on neuropsychological tests (Cansino et al., 2020; Corral et al., 2006; Mohammad et al., 2020; Moraes et al., 2013; Narbutas et al., 2019; Roldán-Tapia et al., 2012; Soldan et al., 2017, 2020) and five studies were cross-sectional, i.e. compared neuropsychological performance in groups with low or high CR or belonging to different age stages (Elshiekh et al., 2020; Puccioni & Vallesi, 2012; Roldán-Tapia et al., 2017; Ruiz-Contreras et al., 2012; Torrens‐burton et al., 2020).

Table 1 Main Findings of Studies Investigating the Relationship between CR and Cognitive Performance (in Alphabetical Order)

Most of the studies included several measures of CR and of cognitive functioning. As far as CR is concerned, six studies considered education as a proxy of CR (Cansino et al., 2020; Mohammad et al., 2020; Moraes et al., 2013; Puccioni & Vallesi, 2012; Roldán-Tapia et al., 2017; Torrens‐burton et al., 2020), two studies considered IQ (Corral et al., 2006; Puccioni & Vallesi, 2012), two studies used a CR scale (Puccioni & Vallesi, 2012; Roldán-Tapia et al., 2017), and six papers used a composite score (Elshiekh et al., 2020; Narbutas et al., 2019; Roldán-Tapia et al., 2012; Ruiz-Contreras et al., 2012; Soldan et al., 2017, 2020).

As far as cognitive assessment is concerned, three studies included a measure of general cognitive functioning (Narbutas et al., 2019; Soldan et al., 2017, 2020), five articles a measure of attention (Corral et al., 2006; Roldán-Tapia et al., 2012, 2017; Ruiz-Contreras et al., 2012; Torrens‐burton et al., 2020), two articles a measure of inhibitory control (Puccioni & Vallesi, 2012; Roldán-Tapia et al., 2017), two articles a measure of verbal fluency (Moraes et al., 2013; Roldán-Tapia et al., 2012), four articles a measure of verbal memory (Cansino et al., 2020; Corral et al., 2006; Mohammad et al., 2020; Roldán-Tapia et al., 2017), and three articles a measure of spatial memory (Corral et al., 2006; Roldán-Tapia et al., 2017; Torrens‐burton et al., 2020). Only one paper (Roldán-Tapia et al., 2017) included a measure of constructional abilities. Among the selected studies, only one study (Puccioni & Vallesi, 2012) was included in a previous meta-analysis on older individuals (Opdebeeck et al., 2016), confirming that our inclusion criteria allowed to target research articles not considered before.

Quality evaluation

Results from quality assessment are provided in Table 2. We rated study quality as “good” for two articles (Narbutas et al., 2019; Roldán-Tapia et al., 2017), “satisfactory” for seven papers (Corral et al., 2006; Elshiekh et al., 2020; Mohammad et al., 2020; Moraes et al., 2013; Puccioni & Vallesi, 2012; Roldán-Tapia et al., 2012; Torrens‐burton et al., 2020), and “unsatisfactory” for the remaining four papers (Cansino et al., 2020; Ruiz-Contreras et al., 2012; Soldan et al., 2017, 2020). We considered “unsatisfactory” these four studies because of presence of a convenience sample (Soldan et al., 2017, 2020), lack of description of the sampling strategy (Ruiz-Contreras et al., 2012), and lack of justification for sample size and control on relevant confounders (Cansino et al., 2020; Ruiz-Contreras et al., 2012; Soldan et al., 2017, 2020).

Table 2 Results from Quality Assessment of Studies Included in the Review (in Alphabetical Order; see Supplementary Materials for Details of the NOS)

Detailed evidence description

General cognitive functioning

Studies on the relationship between CR and general cognitive functioning converged in showing that a global measure of CR might be significantly and positively associated with cognitive performance. In a correlational study, Soldan et al. (2017) targeted cognitive unimpaired individuals with familiarity for Alzheimer’s disease. CR was assessed by a composite score, including years of education, reading, and vocabulary measures; general cognition was evaluated by a composite score obtained from the following tests: i) Paired Associates immediate recall of the Wechsler Memory Scale-Revised (Wechsler, 1987), ii) Logical Memory delayed recall of the Wechsler Memory Scale-Revised (Wechsler, 1987), iii) Boston Naming (Goodglass, Kaplan, Weintraub, 1983), and iv) Digit-Symbol Substitution from the Wechsler Adult Intelligence Scale-Revised (Wechsler, 1981). The results revealed that higher CR was associated with higher cognitive performance. This finding has been confirmed also by a more recent study from the same group (Soldan et al., 2020). Similarly, Narbutas et al. (2019) investigated the relationship between CR and general cognition in healthy adults. CR was measured by a composite score including educational level, occupational demands, physical activities, and leisure activities. General cognition was assessed by the Preclinical Alzheimer Cognitive Composite score (PACC5), a composite measure sensitive to early subtle cognitive changes (Donohue et al., 2014) and including tests assessing episodic memory, executive function, and global cognition. The results revealed that higher CR was related to higher cognitive efficiency, also after controlling for sex and age. Although the above studies pointed out a significant association between CR and cognitive performance, they did not allow to identify the proxies of CR which may be primarily linked to cognition. Also, they did not provide relevant information on the cognitive domain mainly involved in this relationship.


The studies assessing attentional abilities provided moderate evidence of a positive relationship between CR (primarily IQ, years of education and participation in multiple leisure activities) and simple attentional tasks during adulthood. For instance, Corral et al. (2006) aimed at exploring the association between CR and performance in several cognitive domains, among which attention (the other domains will be considered below). CR was measured by a single proxy, i.e. the IQ, assessed by the Vocabulary subtest of the Wechsler Adult Intelligence Scale (Wechsler, 1981). Attention was measured by using the WAIS Digit Span subtest. The results showed a significant explanatory effect of CR on the Digit Span, as participants with low CR were more likely to obtain low scores in the Digit Span subtest than participants with high CR. Roldán-Tapia et al. (2012) explored the relationship between a composite score of CR (including years of education, occupational attainment, and vocabulary knowledge), and performance in the Digit Span and the Trail Making Test (Gordon, 1978). CR resulted to be a predictor of the Digit Span performance (forward not backward) and of parts A and B of the Trail Making Test. However, these findings were not confirmed by Roldán-Tapia et al. (2017), who did not observe significant covariation of either educational level or CR (as measured by the CRS; León-Estrada et al., 2017) on the Digit Span (backward) and the Trail Making Test part B. Ruiz-Contreras et al. (2012) considered diversity and frequency of leisure activities as proxies of CR and assessed their relationships with working memory (WM) efficiency, measured by a n-back task with 2 and 3 levels of difficulty, in young adults. The results showed that the participants with a low index of CR were significantly less accurate on the WM task than the participants with a high index. Although the authors did not find significant relationships of diversity and frequency of leisure activities (considered separately) with accuracy or reaction times on WM tasks, the study might envisage that regular participation in multiple leisure activities during youth could relate to WM efficiency. Finally, Torrens-burton et al. (2020) devised an ecological attentional task -the Swansea Test of Attentional Control (STAC), where information processing is adapted on the individual level of accuracy- to investigate whether high-demand attentional performance is associated with CR (measured as years of education) in healthy individuals, but did not observe significant correlations between education and STAC reaction times.

Inhibitory control

Only two studies explored the relation between CR and inhibitory control during adulthood (Puccioni & Vallesi, 2012; Roldán-Tapia et al., 2017). Although evidence is still limited, these studies seem to suggest a possible relationship between CR and inhibitory control. In Roldan-Tapia et al. (2017), education and scores on a psychometric scale for CR (CRS; León-Estrada et al., 2017) showed a significant covariation with performance on a task for inhibitory control (Stroop Test; Peña-Casanova et al., 2009). Puccioni and Vallesi (2012) considered education and IQ (Wechsler, 1981) as single proxies of CR and observed that response times on a color-word Stroop task were significantly and inversely correlated with IQ but not with education. The Stroop effect (i.e. the difference in response times between incongruent and congruent conditions) and accuracy showed no correlation with verbal IQ.

Verbal fluency

The few studies available so far on verbal fluency seem to support a positive relationship between CR and verbal fluency. Moraes et al. (2013) investigated the correlation between CR (expressed by years of formal education and frequency of reading and writing) and scores on a semantic judgment task, and on unconstrained and constrained verbal fluency tasks (phonemic and semantic fluency; Protocole Montréal d’Évaluation de la Communication – Protocole MEC; Joanette et al., 2004). Education showed the best predictive value on unconstrained verbal fluency, phonemic verbal fluency and semantic verbal fluency. Moreover, Roldan-Tapia et al. (2012) reported that a composite index of CR (including education, occupation and vocabulary knowledge) significantly correlated with scores on phonemic and semantic verbal fluency tasks (Controlled Oral Word Association Test – COWAT; Lezak et al., 2004).

Verbal memory

Overall, available findings supported a moderate positive relationship of composite measures and single proxies of CR with verbal episodic memory during adulthood. In particular, Mohammad et al. (2020) explored the association between several proxies of CR (education, mental activities, and bilingualism) and episodic memory. Episodic memory was assessed by six subtests from the Psychology Experimental Building Language Test battery (Nilsson et al., 1997) including memorization of names and faces, recall of sentences and actions, cued nouns recall, free choice face recognition, forced choice name recognition, noun recognition and cued recall. A factor analysis of results identified five factors: Sentence Memory, Recall-Attention, Action Memory, Name Recognition, and a composite score for Episodic Memory. The participants with distributed education (i.e., who graduated after age 35), outperformed those with continuous education (i.e., who graduated before age 35) in the Sentence Memory, Name Recognition, and global Episodic Memory measures. No difference between groups was found, instead, on episodic memory measures based on language background. Educational level and mental activities (mind teasers, calculation) positively affected episodic memory performance as well. Episodic verbal memory was targeted also by Corral et al. (2006), who observed low scores on the Rey Auditory-Verbal Learning Test (Rey, 1964) more frequently in participants with low CR (assessed by IQ) than in participants with high CR. Similarly, Soldan et al. (2020) reported that higher levels of CR (composite measure including NART, WAIS vocabulary and education) were significantly correlated with higher scores on episodic memory (composite measure considering Paired Associates immediate recall and Logical Memory delayed recall tests). Roldan-Tapia et al. (2017) used education and a CR scale (CRS; León-Estrada et al., 2017) as a measure of CR and assessed young individuals on the Verbal Learning Spanish-Complutense Test (Benedet et al., 1998). The results showed that the sum of the learning slope, the short-term recall score and the delayed-memory score significantly correlated with CR (no effect of education). Moreover, Cansino et al. (2020) investigated whether speed of processing, processing resources, CR (estimated considering the years of formal education) and vocabulary knowledge (vocabulary subscale of the WAIS-R; Wechsler, 1981) mediated the effects of age on source memory decline in a sample covering the entire adult age span. The source memory task consisted of an encoding phase, in which the participants had to evaluate whether some pictures depicted natural or artificial objects, and of a recall phase, when the participants had to judge whether the stimuli were new or old; if the latter was the case, the participants had to point to the quadrant of the display in which the stimuli appeared in the encoding phase. The results showed that CR mediated the effects of age on source memory in middle-aged adults (41–60 years).

Non-verbal memory (and constructional abilities)

Limited and inconsistent evidence is available on the possible relationships between spatial memory and CR in adulthood. On the one hand, Corral et al. (2006) assessed the effect of CR on visuo-spatial memory, expressed by the number of correctly reproduced stimuli in the Benton Visual Retention Test (BVRT; form C, administration A; Benton et al., 1994) and reported that higher verbal IQ scores were associated with better performance on the BVRT. Similarly, in their study cited above, Roldan-Tapia et al. (2017) investigated in young individuals whether education and scores on a CR scale (CRS; León-Estrada et al., 2017) covaried with performance in copying, and in short-term and delayed recall of the Rey-Osterrieth Complex Figure Test (Peña-Casanova et al., 2009). A significant covariation of education with the performance on the copying and the delayed recall tasks was found, but not with the short-term recall task; CR scale scores did not covary with the three tasks. Finally, in a fMRI investigation, Elshiekh et al. (2020) looked at functional compensatory activity patterns related to aging and their association with CR during a spatial memory task in a wide sample covering the adult lifespan. CR was measured by a composite score obtained from years of education and crystallized IQ, whereas the memory task required to encode and then retrieve spatial and temporal information. In keeping with the previous study, the behavioral results showed no modulatory effect of CR on task accuracy or reaction times. However, CR related to age-invariant and task-general activity in several cortical regions (Elshiekh et al., 2020).


In the next decades, worldwide estimates forecast thousand millions of individuals with dementia (Alzheimer’s Disease International, 2018). Faced with this perspective, the main objective is to profile at risk individuals through timely assessment of modifiable risk factors (Ranson et al., 2021), some of which are related to education and lifestyle habits (Livingston et al., 2020). The present review explored the relationships between CR and performance in the domains of general cognitive functioning, attention, inhibitory control, verbal fluency, constructional abilities, and verbal and spatial long-term memory during adulthood. The evidence collected here showed that such relationships do exist, at least for selected cognitive domains, thus suggesting that CR could be considered as one of the modifiable risk factors for dementia. In this sense, the present findings complemented evidence showing a significant positive relationship between CR and cognition in later life (Baker et al., 2017; Opdebeeck et al., 2016) and in specific brain diseases (Santangelo et al., 2019a, 2019b; Stern, 2012; Umarova et al., 2019).

Of note, in their recent meta-analysis on healthy elderly people over 60, Opdebeeck et al. (2016) reported that the proxies of CR such as educational level, occupational status, and engagement in cognitively stimulating activities -alone and in combination- were positively associated with scores in the domains of memory, executive function, visuospatial ability, and language. In our study, we targeted an earlier stage of life, adulthood, to understand when the association between CR and cognition develops, and whether it is possible to enhance cognitive resilience against pathological aging, in line with recent guidelines (Livingston et al., 2020; Ranson et al., 2021; Zhang et al., 2021).

Association between cognitive reserve and cognition across lifespan

Our systematic literature search identified only a small number of papers. These few papers differed in terms of the targeted cognitive domains and measures of CR, thus precluding the implementation of a quantitative approach to analyze data (Forero et al., 2019; Siddaway et al., 2019). Almost all of the selected papers (twelve out of thirteen) had not been considered in previous reviews (Opdebeeck et al., 2018). This confirmed that the adopted inclusion criterion based on the age of the sample allowed us to select specific studies. Quality assessment of these studies generally revealed satisfactory appropriateness of recruitment strategy, representativeness of the sample, and statistical analyses and outcomes. Nonetheless, the sampling strategies should be further improved, by including a clear justification of the sample size, and controlling for relevant confounders.

Overall, the selected studies revealed that CR is generally correlated with performance in the cognitive domains considered. To summarize, the selected papers supported that: i) composite indexes of CR (including education, reading, and vocabulary measures, occupational demands, physical activities, and leisure activities) are positively related to general cognitive functioning; ii) IQ and composite scores of CR are related to performance in sustained attentional tasks, whereas frequency and/or diversity of leisure activities are related to working memory performance; iii) the ability to inhibit automatic responses is positively related to IQ, education and performance on a standardized scale of CR; iv) the single proxy of education and a composite index including education, occupation and vocabulary knowledge positively correlate with verbal fluency performance; v) single proxies of education, IQ, and scores in a standardized scale of CR are related to verbal memory performance; vi) IQ and education as single proxies might be related to non-verbal memory scores.

The positive association of education with several cognitive domains is in line with data showing that years of formal education are positively correlated with cognitive functioning throughout adulthood (Lövdén et al., 2020). No clear conclusion can be reached on constructional abilities, as only one of the selected studies (Roldan-Tapia et al., 2017) investigated the relationship between CR and visuo-constructional functions. Yet, constructional abilities are a relevant cognitive domain in the assessment of cognitive deterioration and dementia (Trojano & Gainotti, 2016).

Limitations and future directions

Although this general framework highlighted a positive effect of CR, and of its singular proxies, on several cognitive domains, some inconsistencies must be pointed out. For instance, Roldan-Tapia et al. (2017) showed a significant covariation between education and inhibitory control assessed by the Stroop Test, whereas Puccioni and Vallesi (2012) found no significant effect of education on the same task. Recent studies on the relationships between CR and attentional (Torrens‐burton et al., 2020) and memory tasks (Elshiekh et al., 2020) reported some null results, too. These inconsistencies warrant attention in future studies, to achieve a clear understanding of the role of CR. In addition, as pointed out by Opdebeeck et al. (2016) the single proxy measures of CR are likely to share an underlying process, even though each might provide an additional exclusive contribution to cognition. This issue pinpoints a limitation of some studies in CR literature that used single or heterogeneous proxies of CR, possibly leading to underestimation of the relationships between CR and cognitive functions and to non-comparability between studies. For these reasons, we encourage in future studies the use of standardized scales of CR, to provide comprehensive and comparable measurement of the construct (Altieri et al., 2018; Kartschmit et al., 2019; Nucci et al., 2012). CR questionnaires could be part of the initial assessment in the workflow for dementia risk profiling within brain health services (Ranson et al., 2021).

Moreover, future studies will benefit from examining how CR and its proxies are associated with structural and functional brain changes in the healthy brain. For instance, recent evidence showed that in healthy elderly individuals CR correlates with levels of activation in a right fronto-parietal network during visual attention tasks (Brosnan et al., 2018). Moreover, Elshiekh et al. (2020) reported that CR was significantly related to changes in activity in several brain regions including the superior temporal, occipital, and left inferior frontal regions, during a non-verbal episodic memory task in adult individuals. It would be worth investigating the neural correlates of CR in the young healthy brain and assessing the effects of specific behavioral trainings or brain stimulation interventions aimed at increasing CR (Passow et al., 2017) on topography of neural activation.


In conclusion, the present systematic review provides the background for early assessment of CR during adulthood. Indeed, it demonstrates that the evaluation of the relationship between CR and cognition in adulthood is feasible and meaningful. At the same time, the review calls for the use of standardized scales simultaneously assessing multiple proxies of CR, and for more extensive research on the relationship between CR and nonverbal cognitive functions. This study provided further support to worldwide guidelines for risk profiling and early intervention (Livingston et al., 2020; Ranson et al., 2021; Zhang et al., 2021) which are expected to inform brain healthcare services in the present and in the next future.