jpad journal

AND option

OR option



A. Thomas1, S. Lefèvre-Arbogast1, C. Féart1, A. Foubert-Samier1,2, C. Helmer1, G. Catheline3, C. Samieri1


1. Univ. Bordeaux, Inserm, BPH, U1219, F-33000 Bordeaux, France; 2. Institut des Maladies Neurodégénératives, Bordeaux Univ. Hospital, F-33000 Bordeaux, France;
3. INCIA, EPHE, Université PSL, Univ. Bordeaux, CNRS, F-33000 Bordeaux, France

Corresponding Author: Aline Thomas, Inserm U1219, Isped, Univ. Bordeaux, CS 61292, 146 rue Léo-Saignat, F-33076 Bordeaux cedex, France; Phone: + (33) 05 57 57 12 99 Fax: + (33) 05 57 57 14 86; e-mail:

J Prev Alz Dis 2022;
Published online August 9, 2022,



Background: Adherence to the Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet, which combines higher consumption of vegetables, berries, nuts, whole grains, olive oil, fish, beans and poultry, with lower consumption of meat, sugars and saturated fats, is a promising strategy to prevent dementia. However, evidence in populations with non-US food culture, especially from Europe, is limited.
Objectives: To evaluate the association of a French-adapted MIND diet score with gray matter volumes, white matter microstructure and incident dementia.
Design and setting: This longitudinal study included participants from the population-based Three-City Bordeaux cohort (≥65 years), with a follow-up from June 2001 to February 2018.
Participants: Dementia-free participants at dietary assessment, in 2001-2002, who underwent systematic detection of incident dementia (over up to 7 visits). A subset of the cohort was included in an ancillary MRI study in 2010-2011.
Measurements: A French-adapted MIND diet score (range, 0-15) was computed from a 148-item Food Frequency Questionnaire and a 24-hour recall administered at home. Incident dementia and its subtypes were adjudicated by an expert committee; and gray matter volumes and white matter microstructure were assessed by 3D-T1 MRI and diffusion-MRI.
Results: Among 1,412 participants (mean age, 75.8 [SD, 4.8]; 63% women), followed for a median of 9.7 years (maximum 16.3 years), 356 (25.2%) developed incident dementia. In multivariable-adjusted Cox model, a higher French MIND diet score was associated with lower risks of dementia and AD (hazard ratios for 1-point of score = 0.89 [95% confidence interval, 0.83-0.95] and 0.88 [0.81-0.96], respectively). In Tract-Based Spatial Statistics analysis of 175 participants included in the MRI sub-study, a higher MIND diet score was associated with lower diffusivity values in the splenium of the corpus callosum (P < .05 after Family-Wise Error-correction). In contrast, there was no significant association of the adapted MIND diet score with gray matter volumes in Voxel-Based Morphometry analysis.
Conclusion: In this cohort of French older adults, higher adherence to the French MIND diet was associated with a lower dementia risk and with preserved white matter microstructure. These results provide further evidence for a role of the MIND diet in the prevention of dementia.

Key words: MIND diet, dementia, magnetic resonance imaging, prospective studies, risk factors in epidemiology.



There is increasing evidence for a role of healthy diets to lower the risk of dementia. Combining the traditional Mediterranean diet and the Dietary Approaches to Stop Hypertension (DASH), the hybrid Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet has been specifically designed for the prevention of Alzheimer’s disease (AD) and dementia (1). The MIND diet emphasizes neuroprotective foods such as fish (2), green leafy vegetables (3, 4) and berries (5, 6), which provide nutrients with specific neurobiological targets (long-chain omega-3 fatty acids, folate and anthocyanins, respectively) (7, 8). In the Rush Memory and Aging Project (MAP), the MIND diet was associated with a lower risk of AD (1) and lower cognitive decline (9). Since these seminal studies, beneficial associations have been reported in the literature but most subsequent research has been cross-sectional (10–12) or with limited follow-ups (13–19) (with a notable exception in Australia (20)). Moreover, studies have been mostly conducted in populations with North-American food culture and prospective studies in European populations are lacking. Differences in dietary habits and guidelines across countries (21), with, for example, higher intake of fish or lower intake of nuts in France compared to the US (1, 2), might result in differences observed in the relationships between the overall diet and health outcomes. Thus, evaluation of “universal” preventive strategies (i.e., that are based on a unique set of lifestyle-based recommendations), as the MIND diet, across different cultures is critical for design and implementation of efficient prevention strategies that may apply to the variety of cultural backgrounds around the globe.
In this study, we took advantage of a large population-based cohort of French older adults to evaluate the association of a MIND diet score adapted to French habits with incident dementia over up to 16 years. Moreover, in a structural neuroimaging sub-study assessed on average 9 years after dietary ascertainment, we interrogated potential mechanisms through investigation of gray matter (GM) volumes and white matter (WM) microstructure.



Study population

The Three-City (3C) Bordeaux study is an ongoing prospective cohort of 2,104 non-institutionalized community dwellers aged 65 years or older selected from the electoral rolls of Bordeaux (France) in 1999-2000 (22). At baseline, data collected at home by a trained psychologist included sociodemographic, lifestyle and medical information, anthropometric and blood pressure measurements, and cognitive evaluation through a battery of neurocognitive tests. Seven follow-up visits were performed every 2 to 3 years, with face-to-face interviews, including a repeated evaluation of cognitive performances and ascertainment of incident dementia cases until year 2018.
In 2001-2002, a nutritional survey was performed by trained dieticians during a home-interview, including a food frequency questionnaire (FFQ) and a 24-hour recall (23). Among the 1,659 participants of the survey (95% of those examined at the corresponding follow-up visit), 1,584 had no missing dietary data. We excluded 70 individuals with prevalent dementia at the time of dietary assessment and 102 not subsequently followed for dementia, leaving 1,412 participants for the main analysis (eFigure 1). Among them, 203 were part of an ancillary neuroimaging study conducted in 2010-2011, with a 3D-T1 and a diffusion tensor imaging (DTI) sequences for investigation of GM and WM microstructure, respectively. After excluding 28 individuals with brain tumors or major cerebrovascular pathologies at MRI, 175 participants were included in the secondary analysis of brain structure.

Standard protocol approvals, registrations, and patient consents

The protocol of the 3C study was approved by the Consultative Committee for the Protection of Persons participating in Biomedical Research at Kremlin-Bicêtre University Hospital (Paris, France), and all participants provided written informed consent.

Dietary assessment and the French MIND diet score

In the FFQ, the frequency of consumption of 148 foods and beverages was recorded in 11 classes and converted into a number of eating occasions per week. In the 24h-recall, participants reported all meals and beverages, with quantities, consumed the day prior the interview (23). The 24h-recall was previously validated by correlating the estimated energy and macronutrient intakes with biological measurements of standard lipids (23). The FFQ was validated against the 24h-recall with statistically significant correlations between food intakes estimated by both questionnaires (eTable 1). Moreover, in a subset of 717 participants who both answered the dietary survey and provided blood samples for measurement of nutrient biomarkers, food intakes from the FFQ were reasonably correlated with plasma levels of specific nutrients (e.g., fish and plasma long-chain polyunsaturated omega-3 fatty acids, correlation coefficient r = 0.25, p<0.001; vegetable and total plasma carotenoids, r = 0.21, p<0.001).
The MIND diet score is a 15-item score combining the consumptions of 10 brain healthy food groups (green leafy vegetables, other vegetables, nuts, berries, beans, whole grains, fish, poultry, olive oil and wine [moderate]) and 5 unhealthy food groups (red meats, butter/margarine, cheese, pastries/sweets, fried/fast food) (9). For each component, a score of 0, 0.5 or 1 is given according to the frequency of consumption. The total MIND diet score is computed by summing the 15 component sub-scores (range, 0-15).
There are differences in US and French dietary habits (e.g., average fish intake was higher in 3C than in Rush MAP) (1, 2), thus the thresholds from the original MIND scoring system (developed in the US) were not always applicable in our French population. To ensure sufficient sample size in each category of component intake, we adapted the original thresholds for component’s sub-scoring to French dietary habits and guidelines (24, 25). Food intakes were estimated from the FFQ, whenever possible. Three items (green leafy, whole grains, nuts), not assessed in the FFQ, were ascertained through the 24h-recall (in grams per day). Berry consumption is much lower in France than in North-America and very few consumers were reported in our 24h-recall (< 15%). Since berries were originally included in the MIND diet due to their high content in polyphenols, we replaced berry intake by total polyphenol intake (6) (we also considered using the more berry-specific anthocyanins intake in an alternative score) (Table 1 and eTable 2).

Table 1. Definition of French-adapted MIND diet components* and scoring system

* MIND diet components were selected according to Morris et al. (9), and thresholds were adapted to French dietary habits and guidelines (24); † Data derived from the 24-hour recall, with a binary threshold for components not frequently consumed among study participants (whole gains, nuts), or thresholds set according to tertiles of daily intake (green leafy vegetables, polyphenols [as main active nutrient in berries]).


Diagnosis of dementia

Incident dementia cases were ascertained at each follow-up visit through a three-step procedure: (i) assessment of neuropsychological performances by a psychologist during home interviews; (ii) examination of suspected cases by a neurologist to establish clinical diagnosis; and (iii) review of all potential cases by an independent committee of neurologists to validate diagnosis and etiology (including probable/possible AD) according to criteria of the Diagnostic and Statistical Manual of Mental Disorders (26) and the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders Association criteria (27). Incident cases were adjudicated from June 2001 (first visit) until February 2018 (seventh follow-up).

Assessment of brain structure

MRI examination was performed on a 3-T Achieva (Philips Medical System, the Netherlands). Brain volumes (GM, WM, and cerebrospinal fluid) were assessed using Voxel-Based Morphometry (VBM) procedure implemented in SPM12 software (Statistical Parametric Mapping 12). WM microstructure integrity was assessed through DTI parameters processed by Tract-Based Spatial Statistics (TBSS) pipeline within FMRIB Software Library (FSL): (i) fractional anisotropy, representing the directionality of water diffusivity along WM fibers; (ii) mean, axial and radial diffusivities, which quantify the diffusion of water molecules globally and along the principal and perpendicular direction of fibers, respectively. Higher directionality (i.e., higher fractional anisotropy) and lower magnitude of diffusion (i.e., lower diffusivities) generally indicate preserved architecture of WM fiber tracts (28). Details on MRI acquisition and processing are provided in the eMethods.


Covariates were collected during the inclusion interview (an average 1.9 years before dietary assessment), except age, body mass index (BMI), total energy intake from the 24h-recall and high depressive symptoms (Center of Epidemiologic Studies-Depression scale (29) score ≥17 for men and ≥23 for women, or being too depressed to answer) (30) that were ascertained at the time of the dietary survey. Sociodemographic and lifestyle information included sex, educational level, tobacco consumption and engagement in regular physical activity (defined as practicing sports or intensive leisure activity [e.g., hiking or swimming] ≥1 hour/week and/or engaging in more moderate activity [e.g., walking or household] ≥1 hour/day). Apolipoprotein E (APOE) ε4 genotype was considered dichotomously (at least one versus no ε4 allele). Vascular risk factors included history of cardiovascular or cerebrovascular disease, hypertension (blood pressure ≥140/90 mmHg, or treatment), hypercholesterolemia (plasma total cholesterol ≥6.2 mmol/L, or treatment) and diabetes (fasting blood glucose ≥7.0 mmol/L, or treatment).

Statistical analyses

Association of French MIND diet with dementia risk

Associations of the French MIND diet score with the risk of all-cause dementia and AD were estimated using multivariable-adjusted Cox proportional hazards models with age as a time scale and delayed entry. Baseline characteristics and incidence rates of dementia and AD were described across tertiles of MIND diet score. Hazard Ratios (HR) for dementia and AD were estimated for each 1-point increase in French MIND diet score, adjusted for covariates cited above.
The proportional hazards assumption was investigated with Schoenfeld residuals and the log-linearity hypothesis was assessed using restricted cubic splines (31). Statistical analyses were performed using SAS v9.4 (SAS Institute Inc), and R version 3.6.1 (R Foundation). Two-sided P-values were used with α = 0.05 threshold for statistical significance. Missing data for covariates were imputed by multiple imputations (using chained equations with fully conditional specification method; M = 5 imputations).
The interactions of French MIND diet score with sex and APOEε4 status on dementia risk were examined. The associations of each individual component of the MIND diet with the risk of dementia were secondarily explored.
Moreover, we conducted several sensitivity analyses. First, we excluded participants with cognitive impairment at dietary assessment (Mini-Mental State Examination <26) to address the possibility of reverse causality. Second, we investigated the association of the French MIND diet score with mortality risk to evaluate the possibility of competing risk by death in interval-censored time-to-event analyses. Third, we evaluated the impact of imputing missing data by running a complete cases analysis.

Association of French MIND diet with brain structure

We examined the relation of the French MIND diet with both brain volumes and WM microstructure (from DTI) using voxel-based statistics. For GM and WM volumes, we used SPM12 linear regressions, with significant threshold cluster of 100 voxels. For DTI parameters, we employed FSL linear regressions (32), with statistical inference based on permutation-based statistics (5000 permutations and threshold-free cluster enhancement) (33). Analyses were adjusted for the full set of covariates as in the main analysis, and models were controlled for multiple comparisons with family-wise error (FWE) correction at a statistical thresholds of P <0.05. GM clusters and WM tracts associated with the French MIND diet score were labeled using the Automated Anatomical Labeling 3 (34) and the Johns Hopkins University atlas (35), respectively.



The 1,412 participants included in the study were 75.8 (SD, 4.8) years-old on average, at dietary assessment, and 63% were women (Table 2). The mean French MIND diet score was 7.2 (SD, 1.5; range, 2.5 to 12) points. Participants with higher French MIND diet score were younger, had lower BMI, and slightly lower tobacco consumption; they tended to practice more physical activity, to be less often diabetics, and to report less depressive symptoms. No differences were observed for sex, educational level, APOE ε4 status, hypertension, hypercholesterolemia or history of cardiovascular diseases.

Table 2. Participants’ characteristics across tertiles of the French-adapted MIND diet score, the 3C Bordeaux study, 2001-2002 (N = 1,412)

Abbreviations: 3C = Three-City; APOEε4 = ε4 allele of the apolipoprotein E gene; BMI = Body Mass Index; FFQ = Food Frequency Questionnaire; MIND = Mediterranean-DASH Diet Intervention for Neurodegenerative Delay; MMSE = Mini-Mental State Examination; SD = standard deviation; * Means and percentages are of non-missing values. Values were missing for: 16.6% of the sample for physical activity; 9.6% for ApoE4; 8.9% for diabetes; 6.1% for MMSE; 5.1% for hypercholesterolemia; 3.5% for total energy intake; 2.1% for BMI; 1.4% for tobacco consumption; 0.5% for education level; and 0.1% for hypertension.


French MIND diet and risk of dementia

Over a median follow-up of 9.7 years (range, 0.9 to 16.3), 356 participants were diagnosed with dementia, including 240 participants with AD. The incidence of both dementia and AD decreased with increasing French MIND diet score (Table 3). Compared to the incidence rate for dementia of 3.00 (95% CI, 2.28; 3.18) per 100 person-years among participants in the lowest tertile of MIND diet score, an absolute difference of -0.84 (-1.20; -0.48) was found for those in the highest tertile.
In multivariable analyses, each 1-point increase of the French MIND diet score was associated with a lower risk of both all-cause dementia (HR = 0.90; 95% CI, 0.83-0.96) and AD (HR = 0.89; 0.81-0.97) (Figure 1). When studying the French MIND diet score into tertiles, compared to participants in the lowest tertile of score, the HR for dementia was 0.73 (0.55-0.97) in the highest tertile and 0.93 (0.73-1.17) in the middle tertile (for AD, HR = 0.70 [0.49-1.00] and 0.96 [0.72-1.27] in the highest and middle tertiles, respectively, compared to the lowest tertile). These associations were not modified by APOE ε4 status or sex (P ≥ .32 for interaction terms). Replacing polyphenols intake by anthocyanins in an alternative French MIND diet score did not change the results.
Studying each component individually in fully-adjusted models, although the directions of associations were generally in the expected direction, hazard ratios for all-cause dementia were close to 1 and no single component emerged as leading the association of the MIND diet to dementia risk (eTable 3). A higher intake of cheese was associated with an increased risk of dementia with borderline significance (HR = 1.03; 95%CI, 1.00-1.05 for each additional eating occasion per week), but this association was no longer statistically significant after adjustment for multiple testing and after mutual adjustment for the other food components (eTable 3).
In sensitivity analyses, results were virtually unchanged when excluding the 243 individuals with cognitive impairment (eTable 4). Participants with missing data for covariates were slightly older, had lower French MIND diet scores and were more often diagnosed with dementia (eTable 5); however, running complete case analyses (n = 1,033) did not change the results (eTable 4). There was no association of the French MIND diet score with mortality risk (P = 0.08).

Table 3. Incidence rates of dementia and Alzheimer’s disease by tertile of French-adapted MIND diet score, Three-City Bordeaux study, 2001-2018 (N = 1,412)

Abbreviations: 95% CI = 95% confidence interval; MIND = Mediterranean-DASH Diet Intervention for Neurodegenerative Delay.


Figure 1. Dementia and Alzheimer’s disease-free survival estimated by multivariable Cox models*, according to increasing French-adapted MIND diet score, the 3C Bordeaux study, 2001-2018 (N = 1,412)

* Cox proportional hazard models with delayed entry and age as time scale, adjusted for sex, status for ε4 allele of the apolipoprotein E (APOEε4) gene, educational level, total energy intake, body mass index, tobacco consumption, practice of regular physical activity, diabetes, history of cerebral and cardiovascular diseases, hypertension, hypercholesterolemia and high depressive symptoms; NOTE. Curves were plotted for three representative French-adapted MIND diet scores (min, median, and max) of an average study participant profile (a woman, with no higher than primary education level, APOEε4 non-carrier, who has an average total energy intake (1622 kcal/d), does not smoke, does not practice regular physical activity, with a body mass index of 26 kg/m², without history of cerebral or cardiovascular diseases, diabetes or high depressive symptoms, with hypertension and hypercholesterolemia); Abbreviations: 3C = Three-City; CI = confidence interval; HR = hazard ratio; MIND = Mediterranean-DASH Diet Intervention for Neurodegenerative Delay.


French MIND diet and brain structure

MRI examination was performed an average 8.9 (SD, 0.2) years after dietary assessment for a subset of 175 participants (see eTable 5 for baseline characteristics). In voxel-based multivariable-adjusted linear regressions, the French MIND diet score (continuous) was positively correlated with GM volumes in the left temporal superior pole and left anterior cingulate cortex. However, these associations did not survive FWE correction (corrected-P >0.05 but <0.10; Figure 2). In TBSS analyses of WM microstructure, a higher French MIND diet score was associated with lower mean and radial diffusivities in the splenium of the corpus callosum, after adjusting for the full set of potential confounders and controlling for multiple comparisons (Figure 2). There was no association of the French MIND diet score with WM volumes in any brain area and fractional anisotropy in any WM bundle.

Figure 2. Multivariable-adjusted associations of higher French-adapted MIND diet score (year 2001-2002) with preserved brain structure (year 2010-2011) estimated by voxel-based linear regressions in SPM and Tract-Based Spatial Statistics in FSL, the 3C Bordeaux study (N = 175)

Areas of the brain where each 1-point increase in French-adapted MIND diet score was associated with higher GM volume (left) and with lower diffusivity values (right), after adjustment for age, sex, status for ε4 allele of the apolipoprotein E gene, educational level, total energy intake, body mass index, tobacco consumption, practice of regular physical activity, diabetes, history of cerebral and cardiovascular diseases, hypertension, hypercholesterolemia and high depressive symptoms. Results are displayed with statistical thresholds of P < 0.10 (top) and P < 0.05 (bottom) after TFCE FWE-correction. For GM volumes, correlated clusters indicated in red (not robust to FWE-correction) were located in the left temporal superior pole and left anterior cingulate cortex according to Automated Anatomical Labeling 3. For WM integrity, results displayed on a MNI template indicated areas of the white matter skeleton robust to FWE-correction for mean and axial diffusivities located in the splenium of the corpus callosum. Abbreviations: 3C = Three-City; FSL = FMRIB Software Library; FEW = Family-Wise Error; GM = Gray Matter; MIND = Mediterranean-DASH Diet Intervention for Neurodegenerative Delay; NS = Non-Significant; SPM = Statistical Parametric Mapping; TBSS = Tract-Based Spatial Statistics; TFCE = Threshold-Free Cluster Enhancement; WM = White Matter.



In this large cohort of older adults, a greater adherence to the French-adapted MIND diet was associated with lower risk of all-cause dementia and AD in the following 16 years. Each increase of 1-point in French MIND diet score was associated with a risk decreased by approximately 10%. Compared to participants in the lowest tertile of score, those in highest tertile had an approximately 30% lower risk of both dementia and AD. When investigating the relationship of the French MIND diet with brain structure at MRI approximately 9 years after the dietary survey, a higher MIND diet score was significantly associated with a preserved WM microstructure, as reflected by lower diffusivity values in the splenium of the corpus callosum. Atrophy and altered WM microstructural integrity of the corpus callosum, reflecting loss of intra-cortical projecting neurons primarily in the splenium part, have been reported in AD as a potential marker of neocortical neurodegeneration associated with memory impairment (36–38).
Our findings in older adults from the south-west of France support a role of the MIND diet for the prevention of brain aging and dementia already suggested in populations with North-American culture and diet. For example, among 923 participants of the Rush MAP (USA), the highest tertile of MIND diet scores were associated with 53% reduction in the 5-year risk of AD (1). Another study from Australia including 1,220 participants yield similar results for the 12-year risk of mild cognitive impairment and dementia combined (20). Moreover, studies on cognition, including three European cohorts (Sweden, France, Spain), reported associations of the MIND diet with slower decline for specific domains and/or global cognition (follow-ups of 0 to 10 years) (9–15,17,18). The single French study to date found that a higher MIND diet score was associated with lower 6-year risk of subjective memory complains only among participants ≥70 years-old (15).
In accordance with our findings, when investigating potential mechanisms through neuroimaging, the Framingham Heart Study (n = 1,904) found no association of the MIND diet with hippocampal or lateral ventricular volumes, or vascular lesions; although a higher score was related to larger total brain volume (19). Most studies relating dietary patterns with brain structure focused on the Mediterranean diet, with mixed findings overall. Some studies identified associations of greater adherence to Mediterranean diet with larger total or regional GM volumes (39, 40) and cortical thickness (41–43), while others did not (44–48). Studies on WM microstructure have been more limited. Consistently with the present results, higher adherence to the Mediterranean diet was associated with preserved WM microstructure, but not with GM volumes in a previous 3C Bordeaux study (n = 161) (45), and with greater WM integrity within the corpus callosum, superior longitudinal fasciculus and corona radiate in another small study (n = 76) (49). In contrast, in the Lothian Birth Cohort 1936 (n = 358), a Principal Component Analysis-derived Mediterranean-style pattern was not associated with brain volumes or WM microstructure; although an inverse association was reported between a processed dietary pattern and greater fractional anisotropy in the splenium of corpus callosum (yet not robust to false-discovery rate correction) (4).
The MIND diet score was specifically designed to encompass brain-healthy dietary components with well-documented neuroprotective properties. The beneficial effect of fish intake on brain aging, attributed to the anti-inflammatory and vasoprotective properties of long-chain polyunsaturated omega-3 fatty acids, has been reported in numerous studies (2, 7, 50). Polyphenols found in fruits, vegetables, coffee or wine and specific species such as anthocyanins provided by berries may lower the risk of dementia through promotion of normal functioning and plasticity of the brain as well as anti-amyloid properties (5, 6). Green leafy vegetables, rich in antioxidants such as carotenoids (51, 52) and vitamin E (53) and important providers of dietary folates, have also been associated with greater brain health (3, 4). Nuts contain fatty acids, phenolic compounds, vitamins E and B, sphingolipids or choline (a precursor of sphingomyelin), which can lower the risk of dementia through several pathways including: amyloidogenesis, oxidative stress, cholinergic or inflammatory pathways (54). There is clear interest to study the additive and synergic effect of these food components by combining them in a recommended dietary pattern easily implementable in all cultures, such as the MIND diet.
The strengths of our large population-based study include a long follow-up of 16 years for dementia, and a clinical diagnosis by an expert committee. For analyses of brain structure, dietary assessment was performed an average 9 years before high resolution multimodal 3-T MRI, which limits the possibility of reverse causation.
A limitation of our study is the estimation of diet exposure with the use of a single dietary assessment, including a single 24h-recall for some components not available in the FFQ (i.e., whole grains, nuts, polyphenols and green leafy). Although a single recall provides acceptable estimations of average dietary intakes in large populations (55), it also leads to inevitable measurement error and does not fully capture individual variations in dietary habits, potentially causing misclassification. However, for foods ascertained with both the 24h-recall and the FFQ, we found reasonable correlations between average intakes estimated by both surveys (eTable 1). Moreover, the average intake values from the 24-h recall were very similar to those reported based on several 24-h records in other large French population-based cohorts, including estimated intakes of macronutrients (23), polyphenols (6, 56), or B-vitamins (57, 58). All these aspects suggest a single 24-hour recall was reasonably efficient to capture habitual dietary intakes in this large population of older persons.
We also had to accommodate some components and thresholds of the initial score to average intakes in our study population, which may limit external validity of our findings. At the same time, developing a MIND diet score adapted to French dietary habits was necessary for both methodological and practical reasons from a public health perspective. Another limitation is the transversal design of our analysis of brain structure and the focus on structural markers, while other endophenotypes, not available in this study, such as vascular lesions or amyloid-β deposition may also be of interest. Finally, although analyses were controlled for a large set of potential confounders, as in any observational study, residual confounding may still persist.
In this large population-based cohort of older French adults, greater adherence to the French-adapted MIND diet was associated with lower risk of dementia and AD, with suggestion of preservation of WM microstructure. These results provide further evidence for a role of brain-healthy diets such as the MIND to promote healthy brain aging and prevent dementia (59).


Acknowledgments/Funding: The Three-City Study is conducted under a partnership agreement between the Institut National de la Santé et de la Recherche Médicale (INSERM), the Institut de Santé Publique et Développement of the University of Bordeaux and Sanofi-Aventis. The Fondation pour la Recherche Médicale funded the preparation and initiation of the study. The Three-City Study is also supported by the Caisse Nationale Maladie des Travailleurs Salariés, Direction Générale de la Santé, Mutuelle Générale de l’Education Nationale, Institut de la Longévité, Regional Governments of Aquitaine and Bourgogne, Fondation de France, Ministry of Research-INSERM Programme “Cohortes et collections de données biologiques”, French National Research Agency COGINUT ANR-06-PNRA-005 and ANR 2007LVIE 003, the Fondation Plan Alzheimer (FCS 2009–2012), the Caisse Nationale pour la Solidarité et l’Autonomie (CNSA), and Roche Pharma. A. Thomas was part of the University Research school (École Universitaire de Recherche, EUR) Digital Public Health PhD program, supported within the framework of the French National Research Agency (ANR) “Programme d’Investissement d’Avenir” (Investment for the Future) PIA3 (17-EURE-0019). The sponsors had no role in the design and conduct of the study; in the collection, analysis, and interpretation of data; in the preparation of the manuscript; or in the review or approval of the manuscript.

Conflict of interest disclosure: A. Thomas, S. Lefèvre-Arbogast, A. Foubert-Samier, C. Helmer, G. Catheline and C. Samieri report no conflict of interest. C. Féart received fees for conferences from Laboratoire Lescuyer and Synadiet.

Ethical standards: The protocol of the 3C study was approved by the Consultative Committee for the Protection of Persons participating in Biomedical Research at Kremlin-Bicêtre University Hospital (Paris, France), and all participants provided written informed consent.





1. Morris MC, Tangney CC, Wang Y, Sacks FM, Bennett DA, Aggarwal NT. MIND Diet Associated with Reduced Incidence of Alzheimer’s Disease. Alzheimers Dement. 2015;11(9):1007–14.
2. Samieri C, Morris MC, Bennett DA, et al. Fish Intake, Genetic Predisposition to Alzheimer Disease, and Decline in Global Cognition and Memory in 5 Cohorts of Older Persons. Am J Epidemiol. 2018;187(5):933–40.
3. Morris MC, Evans DA, Tangney CC, Bienias JL, Wilson RS. Associations of vegetable and fruit consumption with age-related cognitive change. Neurology. 2006;67(8):1370–6.
4. Corley J, Cox SR, Taylor AM, et al. Dietary patterns, cognitive function, and structural neuroimaging measures of brain aging. Exp Gerontol. 2020;142:111117.
5. Devore EE, Kang JH, Breteler MMB, Grodstein F. Dietary intakes of berries and flavonoids in relation to cognitive decline. Annals of Neurology. 2012;72(1):135–43.
6. Lefèvre-Arbogast S, Gaudout D, Bensalem J, et al. Pattern of polyphenol intake and the long-term risk of dementia in older persons. Neurology. 2018;90(22):e1979–88.
7. Cunnane SC, Plourde M, Pifferi F, Bégin M, Féart C, Barberger-Gateau P. Fish, docosahexaenoic acid and Alzheimer’s disease. Prog Lipid Res. 2009;48(5):239–56.
8. Vauzour D. Polyphenols and brain health. OCL. EDP Sciences. 2017;24(2):A202.
9. Morris MC, Tangney CC, Wang Y, et al. MIND diet slows cognitive decline with aging. Alzheimers Dement. 2015;11(9):1015–22.
10. McEvoy CT, Guyer H, Langa KM, Yaffe K. Neuroprotective Diets Are Associated with Better Cognitive Function: The Health and Retirement Study. J Am Geriatr Soc. 2017;65(8):1857–62.
11. Calil SRB, Brucki SMD, Nitrini R, Yassuda MS. Adherence to the Mediterranean and MIND diets is associated with better cognition in healthy seniors but not in MCI or AD. Clin Nutr ESPEN. 2018;28:201–7.
12. Wesselman LMP, van Lent DM, Schröder A, et al. Dietary patterns are related to cognitive functioning in elderly enriched with individuals at increased risk for Alzheimer’s disease. Eur J Nutr. 2021;60(2):849–60.
13. Shakersain B, Rizzuto D, Larsson SC, Faxén-Irving G, Fratiglioni L, Xu WL. The Nordic Prudent Diet Reduces Risk of Cognitive Decline in the Swedish Older Adults: A Population-Based Cohort Study. Nutrients. 2018;10(2).
14. Berendsen AM, Kang JH, Feskens EJM, de Groot CPGM, Grodstein F, van de Rest O. Association of Long-Term Adherence to the MIND Diet with Cognitive Function and Cognitive Decline in American Women. J Nutr Health Aging. 2018;22(2):222–9.
15. Adjibade M, Assmann KE, Julia C, Galan P, Hercberg S, Kesse-Guyot E. Prospective association between adherence to the MIND diet and subjective memory complaints in the French NutriNet-Santé cohort. J Neurol. 2019;266(4):942–52.
16. Cherian L, Wang Y, Fakuda K, Leurgans S, Aggarwal N, Morris M. Mediterranean-Dash Intervention for Neurodegenerative Delay (MIND) Diet Slows Cognitive Decline After Stroke. J Prev Alzheimers Dis. 2019;6(4):267–73.
17. Munoz-Garcia MI, Toledo E, Razquin C, et al. “A priori” Dietary Patterns and Cognitive Function in the SUN Project. Neuroepidemiology. 2020;54(1):45–57.
18. Mueller KD, Norton D, Koscik RL, et al. Self-reported health behaviors and longitudinal cognitive performance in late middle age: Results from the Wisconsin Registry for Alzheimer’s Prevention. PLoS One. 2020;15(4):e0221985.
19. Melo van Lent D, O’Donnell A, Beiser AS, et al. Mind Diet Adherence and Cognitive Performance in the Framingham Heart Study. J Alzheimers Dis. 2021;82(2):827–39.
20. Hosking DE, Eramudugolla R, Cherbuin N, Anstey KJ. MIND not Mediterranean diet related to 12-year incidence of cognitive impairment in an Australian longitudinal cohort study. Alzheimers Dement. 2019;15(4):581–9.
21. Herforth A, Arimond M, Álvarez-Sánchez C, Coates J, Christianson K, Muehlhoff E. A Global Review of Food-Based Dietary Guidelines. Adv Nutr. 2019;10(4):590–605.
22. 3C Study Group. Vascular factors and risk of dementia: design of the Three-City Study and baseline characteristics of the study population. Neuroepidemiology. 2003;22(6):316–25.
23. Féart C, Jutand MA, Larrieu S, et al. Energy, macronutrient and fatty acid intake of French elderly community dwellers and association with socio-demographic characteristics: data from the Bordeaux sample of the Three-City Study. Br J Nutr. 2007;98(5):1046–57.
24. Programme National Nutrition Santé 2019-2023 [online]. Paris: Ministère des solidarités et de la santé; 2019 Sep p. 94. Available from: Accessed February 2, 2021.
25. Kesse-Guyot E, Amieva H, Castetbon K, et al. Adherence to nutritional recommendations and subsequent cognitive performance: findings from the prospective Supplementation with Antioxidant Vitamins and Minerals 2 (SU.VI.MAX 2) study. Am J Clin Nutr. 2011;93(1):200–10.
26. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders: DSM-V. 5th edition. Washington, DC: American Psychiatric Association; 2013.
27. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology. 1984;34(7):939–44.
28. Sullivan EV, Pfefferbaum A. Diffusion tensor imaging and aging. Neurosci Biobehav Rev. 2006;30(6):749–61.
29. Radloff LS. The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Appl Psychol Meas. 1977;1(3):385–401.
30. Fuhrer R, Rouillon F. La version française de l’échelle CES-D (Center for Epidemiologic Studies-Depression Scale) : description et traduction de l’échelle d’autoévaluation. Psychiatr Psychobiol. 1989;3(4):163–6.
31. Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Stat Med. 2010;29(9):1037–57.
32. Kantarci K, Senjem ML, Avula R, et al. Diffusion tensor imaging and cognitive function in older adults with no dementia. Neurology. 2011;77(1):26–34.
33. Smith SM, Nichols TE. Threshold-free cluster enhancement: addressing problems of smoothing, threshold dependence and localisation in cluster inference. Neuroimage. 2009;44(1):83–98.
34. Rolls ET, Huang CC, Lin CP, Feng J, Joliot M. Automated anatomical labelling atlas 3. Neuroimage. 2020;206:116189.
35. Mori S, Oishi K, Jiang H, et al. Stereotaxic white matter atlas based on diffusion tensor imaging in an ICBM template. Neuroimage. 2008;40(2):570–82.
36. Rose SE, Chen F, Chalk JB, et al. Loss of connectivity in Alzheimer’s disease: an evaluation of white matter tract integrity with colour coded MR diffusion tensor imaging. Journal of Neurology, Neurosurgery & Psychiatry. 2000;69(4):528–30.
37. Zhang Y, Schuff N, Jahng GH, et al. Diffusion tensor imaging of cingulum fibers in mild cognitive impairment and Alzheimer disease. Neurology. 2007;68(1):13–9.
38. Teipel SJ, Bayer W, Alexander GE, et al. Progression of Corpus Callosum Atrophy in Alzheimer Disease. Archives of Neurology. 2002;59(2):243–8.
39. Gu Y, Brickman AM, Stern Y, et al. Mediterranean diet and brain structure in a multiethnic elderly cohort. Neurology. 2015;85(20):1744–51.
40. Karstens AJ, Tussing-Humphreys L, Zhan L, et al. Associations of the Mediterranean diet with cognitive and neuroimaging phenotypes of dementia in healthy older adults. Am J Clin Nutr. 2019;109(2):361–8.
41. Mosconi L, Murray J, Tsui WH, et al. Mediterranean Diet and Magnetic Resonance Imaging-Assessed Brain Atrophy in Cognitively Normal Individuals at Risk for Alzheimer’s Disease. J Prev Alzheimers Dis. 2014;1(1):23–32.
42. Mosconi L, Walters M, Sterling J, et al. Lifestyle and vascular risk effects on MRI-based biomarkers of Alzheimer’s disease: a cross-sectional study of middle-aged adults from the broader New York City area. BMJ Open. 2018;8(3):e019362.
43. Staubo SC, Aakre JA, Vemuri P, et al. Mediterranean diet, micronutrients and macronutrients, and MRI measures of cortical thickness. Alzheimers Dement. 2017;13(2):168–77.
44. Titova OE, Ax E, Brooks SJ, et al. Mediterranean diet habits in older individuals: associations with cognitive functioning and brain volumes. Exp Gerontol. 2013;48(12):1443–8.
45. Pelletier A, Barul C, Féart C, et al. Mediterranean diet and preserved brain structural connectivity in older subjects. Alzheimers Dement. 2015;11(9):1023–31.
46. Luciano M, Corley J, Cox SR, et al. Mediterranean-type diet and brain structural change from 73 to 76 years in a Scottish cohort. Neurology. 2017;88(5):449–55.
47. Berti V, Walters M, Sterling J, et al. Mediterranean diet and 3-year Alzheimer brain biomarker changes in middle-aged adults. Neurology. 2018;90(20):e1789–98.
48. Walters MJ, Sterling J, Quinn C, et al. Associations of lifestyle and vascular risk factors with Alzheimer’s brain biomarker changes during middle age: a 3-year longitudinal study in the broader New York City area. BMJ Open. 2018;8(11).
49. Rodrigues B, Coelho A, Portugal-Nunes C, et al. Higher Adherence to the Mediterranean Diet Is Associated With Preserved White Matter Integrity and Altered Structural Connectivity. Front Neurosci. 2020;14:786.
50. Thomas A, Baillet M, Proust-Lima C, et al. Blood polyunsaturated omega-3 fatty acids, brain atrophy, cognitive decline, and dementia risk. Alzheimers Dement. 2021;17(3):407–16.
51. Feart C, Letenneur L, Helmer C, et al. Plasma Carotenoids Are Inversely Associated With Dementia Risk in an Elderly French Cohort. J Gerontol A Biol Sci Med Sci. 2016;71(5):683–8.
52. Yuan C, Chen H, Wang Y, Schneider JA, Willett WC, Morris MC. Dietary carotenoids related to risk of incident Alzheimer dementia (AD) and brain AD neuropathology: a community-based cohort of older adults. Am J Clin Nutr. 2020;113:200-208.
53. Lloret A, Esteve D, Monllor P, Cervera-Ferri A, Lloret A. The Effectiveness of Vitamin E Treatment in Alzheimer’s Disease. Int J Mol Sci. 2019;20(4):E879.
54. Gorji N, Moeini R, Memariani Z. Almond, hazelnut and walnut, three nuts for neuroprotection in Alzheimer’s disease: A neuropharmacological review of their bioactive constituents. Pharmacol Res. 2018;129:115–27.
55. Willett WC. Nutritional Epidemiology. Oxford University Press: New York, NY, USA; 1998.
56. Pérez-Jiménez J, Neveu V, Vos F, Scalbert A. Systematic analysis of the content of 502 polyphenols in 452 foods and beverages: an application of the phenol-explorer database. J Agric Food Chem. 2010;58(8):4959–69.
57. Lefèvre-Arbogast S, Féart C, Dartigues JF, Helmer C, Letenneur L, Samieri C. Dietary B Vitamins and a 10-Year Risk of Dementia in Older Persons. Nutrients. 2016;8(12):761.
58. ANSES. Food Consumption Data from the Individual and National Study of Food Consumption 2 (INCA2) [online]. Available from: Accessed May 12, 2022.
59. Liu X, Morris MC, Dhana K, et al. Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) study: Rationale, design and baseline characteristics of a randomized control trial of the MIND diet on cognitive decline. Contemp Clin Trials. 2021;102:106270.