**Background:** Cognitive reserve (CR) is a latent construct describing the brain's ability to optimize performance and cope with brain damage through adaptive recruitment of neural networks and cognitive strategies. It is distinguished from brain reserve (structural features like brain volume and synaptic integrity). CR is theorized to explain the discrepancy between the severity of brain pathology and its clinical manifestations. Mild Cognitive Impairment (MCI) represents a transitional state between normal aging and dementia, characterized by cognitive decline greater than expected for age but not meeting dementia criteria. MCI can be amnesic or non-amnesic, single- or multi-domain. Given the aging population and lack of effective pharmacological treatments for dementia, identifying modifiable protective factors is critical. This systematic review aimed to investigate the role of CR as a protective factor against MCI and associated cognitive decline, addressing limitations of prior meta-analyses by including both cross-sectional and longitudinal studies with healthy control groups.
**Methods:** The review was conducted according to the PRISMA 2020 Statement. A systematic search was performed in Psycinfo, MEDLINE, Scopus, and Web of Science on December 28, 2022, for peer-reviewed articles in English or Italian with no year, gender, or ethnicity restrictions. Eligibility criteria included randomized controlled, cross-sectional, and longitudinal studies assessing CR and its relationship to cognitive functioning in healthy individuals and MCI patients; clearly defined MCI diagnostic criteria using standardized instruments; participants aged ≥50; and clear reporting of recruitment methods. Exclusion criteria included neurological diseases (e.g., multiple sclerosis, Parkinson's, epilepsy), metabolic diseases (e.g., diabetes), autoimmune diseases, cardiovascular diseases (e.g., stroke), oncological diseases, dementia (frontotemporal, vascular, Lewy body, Alzheimer's), psychiatric disorders, subjective memory disorders, animal models, and non-original research (reviews, meta-analyses, editorials, commentaries, posters, conferences, single cases, clinical trials). The initial search yielded 5922 articles; after removing 515 duplicates, 5407 articles were screened by title and abstract. Sixty-nine full texts were assessed, with 59 excluded (31 missing control group, 18 no MCI sample, 10 inadequate CR measurement), leaving 10 studies for final analysis. Quality assessment used an adapted Cochrane Collaboration tool evaluating attrition bias, reporting bias, sample bias, and measurement bias.
**Key Results:** The 10 selected studies included 5 longitudinal and 5 cross-sectional designs, conducted in Europe (4), Asia (3), and North America (3), with 6145 total participants (2180 MCI, 3965 healthy controls). MCI participants' mean age ranged from 63 to 83 years; healthy controls ranged from 55.7 to 80 years. Quality assessment showed low risk of attrition and reporting bias across all studies; 4 studies had high sample bias risk (small samples, predominantly female or younger participants); 2 had high measurement bias risk (non-validated CR measures or non-specific cognitive tests). MCI diagnostic criteria varied: 4 studies used Petersen's criteria (1999, 2001, or 2004 versions), 2 used Winblad's criteria, and others used DSM-IV, Mayo Clinic Revised, Albert et al., or NINCDS-ADRDA criteria. CR assessment primarily used education level (8 studies), job position (4 studies), leisure activities (4 studies), and IQ measures (NART, WAIS-III, Raven's matrices). Structured instruments included the Cognitive Reserve Index Questionnaire (CRIq) and Lifestyle Activities Questionnaire (LAQ). All 5 longitudinal studies found high CR significantly associated with better cognitive functioning and lower risk of developing MCI. Liu et al. found significant relationships between years of schooling and MoCA, MMSE, and WMS scores. Malave's study found cognitively stimulating activities and IQ associated with lower MCI incidence. Soldan et al. showed individuals with lower CR had poorer cognitive functioning and higher MCI risk; those with high CR who later developed MCI showed better pre-symptom cognitive performance but more rapid decline after onset, consistent with Stern's CR model. Xu et al. found high CR significantly associated with lower MCI risk longitudinally. Kim et al. found healthy participants had higher CRI-leisure time scores and greater cognitive functioning than MCI participants. All 5 cross-sectional studies found positive associations between CR and cognitive functioning. One study found MCI participants with high CR performed better on implicit associative memory tasks. Solé-Padullès et al. found MCI participants had lower CR and worse amnestic recognition performance. Another study found CR mediated episodic memory performance in aMCI participants. A study on cognitively stimulating activities found reading books/newspapers, craft activities, computer use, cognitively stimulating games, and less television watching were significantly associated with lower MCI risk. One study found single-domain aMCI patients had better cognitive functioning correlated with education level compared to multi-domain aMCI patients.
**Clinical Implications:** This systematic review provides consistent evidence that high cognitive reserve is significantly associated with reduced risk of MCI and better cognitive functioning in both healthy older adults and those with MCI. The findings support theoretical models proposing that CR acts through neural capacity and efficiency mechanisms to compensate for cognitive decline. Given that MCI represents a potentially reversible transitional stage before dementia, interventions aimed at enhancing CR—through education, cognitively stimulating occupations, and leisure activities—could serve as preventive strategies. The authors emphasize the need for universal standardized MCI diagnostic criteria and a standardized CR assessment protocol to improve comparability across studies. Limitations include significant methodological heterogeneity in MCI diagnosis, CR assessment, and cognitive domains evaluated across studies; absence of statistical matching for demographic and biological confounders; and exclusion of dementia patients, limiting assessment of CR's impact across the full cognitive decline continuum.