Review Reports - Adherence to Healthy Dietary Patterns and Cognitive Performance: Evidence from DASH and Mediterranean Diets

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript addresses an important topic, “DASH and Mediterranean diet adherence in relation to cognitive performance,” but in its current form, the study is weakened by serious methodological, statistical, interpretive, and presentation problems. The central issue is that the authors repeatedly use causal language, such as “effect” and “positive effect,” even though the study is cross-sectional and correlational. Therefore, the manuscript cannot support causal claims. The results should be reframed as associations only.

Major Comments

The title and abstract imply that DASH and Mediterranean diets influence cognitive performance, but the design is cross-sectional. Replace phrases such as “effect on cognitive performance” with “association with cognitive performance.”
The paragraph describing the Mediterranean diet and cognitive impairment is repeated almost verbatim on page 2. This is unacceptable and suggests poor manuscript preparation. The duplicate paragraph should be removed.
The authors claim originality because they used DASH, Mediterranean diet, Oktem-VMPT, and TMT together. This is not a strong novelty claim unless the authors clearly explain why this combination addresses a specific knowledge gap in Turkish adults or in midlife cognitive assessment.
Age, sex, body mass index, education, smoking, alcohol use, chronic disease, medication use, physical activity, sleep, socioeconomic status, and total energy intake can strongly influence both diet adherence and cognition. The current regression models appear to include only DASH and MEDAS scores, which is insufficient. Multivariable models adjusted for key confounders are essential.
Participants range from 18 to 65 years. Cognitive performance, dietary behavior, and risk factors differ greatly across this range. The authors should perform age-stratified analyses, for example, 18-35, 36-50, and 51-65 years, or include age interaction terms.
TMT and verbal memory tests are strongly affected by education. Without adjustment for education level, the diet-cognition association may simply reflect educational or socioeconomic differences.
DASH-Q assesses intake during the last 7 days. It is biologically weak to link 7-day diet quality with cognitive performance unless the authors clearly acknowledge that this reflects recent dietary behavior rather than habitual diet.
The regression models should report R², adjusted R², model assumptions, variance inflation factor, residual diagnostics, and whether DASH and MEDAS were entered simultaneously. The manuscript reports beta coefficients but does not provide enough information to judge model quality.
The authors describe Spearman correlations as percentage increases or decreases, for example, “11.1% decrease.” This is statistically incorrect. A correlation coefficient of −0.111 is not an 11.1% decrease. These statements must be corrected throughout the Results.
Table 3 labels “Unstandardized Coefficients” but uses β. This is confusing because β usually denotes standardized beta. Use “B” for unstandardized coefficients and “β” only for standardized coefficients.
The authors test many cognitive subcomponents against DASH and MEDAS scores. Without correction for multiple comparisons, several weak associations may be false positives. Apply false discovery rate correction or Bonferroni correction.
Many correlations are very weak. The authors should avoid overstating public health implications unless they discuss the small effect sizes and whether the observed differences are clinically meaningful.
Women had higher MEDAS and DASH scores and better cognitive scores in several domains. This may confound the diet-cognition association. Sex-stratified models or interaction analyses are needed.
The manuscript does not clearly explain how participants were recruited, whether sampling was random, convenience-based, community-based, or clinic-based. This affects external validity.
The authors should specify whether participants with neurological disease, psychiatric illness, dementia, major depression, substance use disorder, traumatic brain injury, visual impairment, or medications affecting cognition were excluded.
TMT-A usually requires connecting numbers 1-25, not 1-23. The authors should verify the Turkish version used and, if necessary, correct the description.
The manuscript says delayed recall was assessed “approximately 30 or 40 minutes later.” This should be standardized and reported precisely because delayed recall is sensitive to interval duration.
The Discussion largely repeats the Results and cites supportive literature without critically evaluating inconsistencies, confounding, reverse causation, or limitations of the cognitive tools.
The authors should briefly discuss plausible mechanisms linking dietary patterns to cognition: vascular health, inflammation, oxidative stress, insulin sensitivity, gut microbiota, polyphenols, omega-3 fatty acids, and blood pressure regulation.
Several references are local theses or non-core sources, while key high-impact literature on Mediterranean diet, DASH, MIND diet, cognitive aging, and dementia prevention is underused. More recent systematic reviews, meta-analyses, and prospective cohort/intervention studies should be added.

Minor Comments and Language Corrections:

In the abstract, use p < 0.05, not p<0,05.
Remove the extra space before the comma in “TMT-A) , Ability…”.
Add a period after “p < 0.05)” in the abstract.
Keywords include “diet” twice. Remove duplication.
“mediterranean” and “dash” should be capitalized properly: Mediterranean, DASH.
“Processing Speed Based on Visual Scanning Ability” and “Ability to Shift Set Between Stimulus Sets and Follow Sequencing” are too long for repeated use; define once and then use TMT-A and TMT-B.
“It only be shared” in the Data Availability Statement is grammatically incorrect; revise to “It will only be shared…”
Author Contributions ends with .”. Remove the extra period.
The abbreviation list is incorrect: MEDAS is listed together with DASH-Q. Each abbreviation should be defined separately.
Table formatting is poor, with broken words and crowded values. Tables should be reformatted before publication.
“Oktem-VMPT” contains a formatting artifact and should be corrected.

Author Response

The title and abstract imply that DASH and Mediterranean diets influence cognitive performance, but the design is cross-sectional. Replace phrases such as “effect on cognitive performance” with “association with cognitive performance.”

Thank you for this important comment. We agree that the cross-sectional design of the study does not allow causal inferences. Therefore, we revised the title, abstract, and relevant sections throughout the manuscript to replace causal expressions such as “effect on cognitive performance” with more appropriate associative terminology such as “association with cognitive performance” or “was associated with cognitive performance.”

The paragraph describing the Mediterranean diet and cognitive impairment is repeated almost verbatim on page 2. This is unacceptable and suggests poor manuscript preparation. The duplicate paragraph should be removed.

Thank you for pointing this out. It was noticed that the paragraph regarding the Mediterranean diet and cognitive impairment had been unintentionally repeated during manuscript preparation. The duplicate paragraph in the Introduction section has been removed.

The authors claim originality because they used DASH, Mediterranean diet, Oktem-VMPT, and TMT together. This is not a strong novelty claim unless the authors clearly explain why this combination addresses a specific knowledge gap in Turkish adults or in midlife cognitive assessment.

Thank you for this valuable comment. We agree that the originality of the study should be more clearly justified beyond the combined use of the DASH diet, the Mediterranean diet, the Oktem-VMPT, and the TMT. We have therefore revised the manuscript to better explain the specific knowledge gap addressed by this study. Previous studies have generally evaluated these dietary patterns separately, whereas the present study simultaneously examined adherence to both dietary models in relation to measures of cognitive performance. Furthermore, given the regional variations in dietary habits across Türkiye, this study provides context-specific evidence for Turkish adults and contributes to the limited literature on dietary patterns and midlife cognitive performance in this population. These points have been clarified in the ‘Strengths of the Study’ section.

Age, sex, body mass index, education, smoking, alcohol use, chronic disease, medication use, physical activity, sleep, socioeconomic status, and total energy intake can strongly influence both diet adherence and cognition. The current regression models appear to include only DASH and MEDAS scores, which is insufficient. Multivariable models adjusted for key confounders are essential.

Thank you for this important comment. In response, Table 1 has been revised and expanded to provide a more comprehensive overview of potential confounding variables that may influence both dietary adherence and cognitive performance, including educational attainment, income level, smoking, alcohol consumption, chronic disease status, medication use, supplement use, and the habit of skipping meals. These revisions provide a clearer characterisation of the study population.

Participants range from 18 to 65 years. Cognitive performance, dietary behavior, and risk factors differ greatly across this range. The authors should perform age-stratified analyses, for example, 18-35, 36-50, and 51-65 years, or include age interaction terms.

We thank the reviewer for this valuable methodological suggestion. We acknowledge that the wide age range (18–65 years) of the study sample may introduce heterogeneity in cognitive performance and dietary behavior. To address this concern, we have incorporated age as a continuous covariate in all multivariate regression models (Table 4). Age emerged as a statistically significant predictor across all cognitive outcome models (p<0.001), confirming that its influence on cognitive performance has been accounted for in the analyses. Additionally, DASH and MEDAS scores were compared across the three age subgroups (18–35, 36–50, and 51–65 years) using non-parametric tests, and statistically significant differences were observed between groups (Table 3).

We recognize that fully stratified regression analyses by age group — or the inclusion of diet × age interaction terms — would provide additional insight into whether the diet–cognition relationship differs across age strata. However, given the sample size constraints within each subgroup and the risk of reduced statistical power, we opted to control for age as a continuous variable in the regression framework rather than conducting separate stratified models. We have explicitly acknowledged this as a limitation of the study in the revised manuscript.

We believe the current analytical approach adequately addresses the confounding role of age while maintaining sufficient statistical power across the full sample.

TMT and verbal memory tests are strongly affected by education. Without adjustment for education level, the diet-cognition association may simply reflect educational or socioeconomic differences.

We thank the reviewer for raising this important point. We would like to clarify that education level and income status were included as covariates in all multivariate regression models presented in Table 4. Educational status emerged as a statistically significant predictor of cognitive performance across all models (p<0.001), confirming that its confounding effect on the diet–cognition association has been controlled for in our analyses. Income level was additionally included to account for broader socioeconomic differences. Therefore, the observed associations between dietary adherence (DASH and MEDAS scores) and cognitive performance reflect relationships independent of educational and socioeconomic background.

DASH-Q assesses intake during the last 7 days. It is biologically weak to link 7-day diet quality with cognitive performance unless the authors clearly acknowledge that this reflects recent dietary behavior rather than habitual diet.

Thank you for this important comment. We agree that the DASH-Q evaluates dietary intake over the previous 7 days and therefore reflects recent dietary behavior rather than long-term habitual dietary patterns. Since cognitive performance is more likely to be influenced by long-term nutritional exposure, this issue represents an important limitation when interpreting the observed associations.

In response to this suggestion, we clarified this point in the Methods section by specifying the short-term assessment period of the DASH-Q and added a statement to the Discussion/Limitations section emphasizing that the identified associations should be interpreted as relationships with recent dietary behavior rather than long-term dietary habits.

8. The regression models should report R², adjusted R², model assumptions, variance inflation factor, residual diagnostics, and whether DASH and MEDAS were entered simultaneously. The manuscript reports beta coefficients but does not provide enough information to judge model quality.

Thank you for this important methodological comment. In response to the reviewer’s suggestion, additional model evaluation and diagnostic information were added to the revised Table 4. Specifically, R², adjusted R², F statistics, confidence intervals, tolerance values, and variance inflation factor (VIF) statistics were included to improve the transparency and interpretability of the regression models. Furthermore, the manuscript was clarified regarding the variables entered into the regression analyses and multicollinearity assessment.

The authors describe Spearman correlations as percentage increases or decreases, for example, “11.1% decrease.” This is statistically incorrect. A correlation coefficient of −0.111 is not an 11.1% decrease. These statements must be corrected throughout the Results.

We thank the reviewer for this important correction. In the original manuscript, correlation coefficients were incorrectly described using percentage language (e.g., "11.1% decrease"), which does not accurately reflect the nature of Spearman's rho. In the revised manuscript, all such expressions have been corrected throughout the Results section. Correlation coefficients are now described using appropriate statistical language, such as "a statistically significant, negative, and very weak correlation was found (rs = −0.111; p<0.01), indicating that as age increased, DASH scores tended to decrease." We have ensured that this correction has been applied consistently across all relevant sections of the manuscript.

Table 3 labels “Unstandardized Coefficients” but uses β. This is confusing because β usually denotes standardized beta. Use “B” for unstandardized coefficients and “β” only for standardized coefficients.

Thank you for this valuable comment. In accordance with the reviewer’s suggestion, the notation in Table 3 has been revised for clarity and statistical accuracy. The symbol “β”, previously used for unstandardised coefficients, has been replaced with “B”, and the table heading has been amended accordingly to accurately reflect unstandardised regression coefficients.

The authors test many cognitive subcomponents against DASH and MEDAS scores. Without correction for multiple comparisons, several weak associations may be false positives. Apply false discovery rate correction or Bonferroni correction.

Thank you for this valuable comment. In response to the concern regarding the risk of a Type I error due to multiple comparisons, we performed Bonferroni correction analyses for the comparisons of cognitive subdomains. The corrected significance threshold was calculated and applied accordingly (Bonferroni-corrected p-value = 0.002). The corrected results have been incorporated into the manuscript and are now presented in the newly added Table 5. Relevant explanations have also been included in the Statistical Analysis and Results sections.

Many correlations are very weak. The authors should avoid overstating public health implications unless they discuss the small effect sizes and whether the observed differences are clinically meaningful.

We thank the reviewer for this important methodological observation. We fully acknowledge that the majority of statistically significant correlations in this study were very weak in magnitude (Spearman's r ranging approximately between 0.13 and 0.28), and that statistical significance — particularly in a sample of 600 participants — does not necessarily imply clinical or practical significance. This limitation has now been explicitly addressed in the Discussion. We have added a paragraph noting that while the observed associations between DASH/MEDAS scores and cognitive performance subscales reached statistical significance, the small effect sizes suggest that dietary adherence alone explains a modest proportion of variance in cognitive outcomes, and that sociodemographic factors such as age, educational level, and smoking status appear to be stronger determinants. Accordingly, we have moderated the public health implications stated in the Discussion and Conclusion sections to more accurately reflect the magnitude of the observed effects. We now emphasize that the findings should be interpreted as preliminary and hypothesis-generating, and that the clinical meaningfulness of these associations requires confirmation through longitudinal studies with larger effect sizes, neuroimaging endpoints, and intervention designs capable of establishing causal directionality. We thank the reviewer for this important methodological observation. We fully acknowledge that the majority of statistically significant correlations in this study were very weak in magnitude (Spearman's r ranging approximately between 0.13 and 0.28), and that statistical significance — particularly in a sample of 600 participants — does not necessarily imply clinical or practical significance. This limitation has now been explicitly addressed in the Discussion. We have added a paragraph noting that while the observed associations between DASH/MEDAS scores and cognitive performance subscales reached statistical significance, the small effect sizes suggest that dietary adherence alone explains a modest proportion of variance in cognitive outcomes, and that sociodemographic factors such as age, educational level, and smoking status appear to be stronger determinants. Accordingly, we have moderated the public health implications stated in the Discussion and Conclusion sections to more accurately reflect the magnitude of the observed effects. We now emphasize that the findings should be interpreted as preliminary and hypothesis-generating, and that the clinical meaningfulness of these associations requires confirmation through longitudinal studies with larger effect sizes, neuroimaging endpoints, and intervention designs capable of establishing causal directionality.

Women had higher MEDAS and DASH scores and better cognitive scores in several domains. This may confound the diet-cognition association. Sex-stratified models or interaction analyses are needed.

Thank you for this valuable comment. Since sex differences may influence both dietary adherence and cognitive performance, the descriptive analyses in Table 1 were presented separately for men and women. This revision allows clearer evaluation of sex-based differences within the study population.

Thank you for this valuable comment. Since sex differences may influence both dietary adherence and cognitive performance, the descriptive analyses in Table 2 were presented separately for men and women. This revision allows clearer evaluation of sex-based differences within the study population.

Thank you for highlighting the potential confounding effect of sex on the relationship between dietary adherence and cognitive performance. In response to this suggestion, we conducted sex-stratified analyses to evaluate the associations separately in female and male participants. The findings of these analyses have been added as a new Table 5 in the revised manuscript. This additional analysis provides a clearer interpretation of the diet–cognition relationship according to sex.

The manuscript does not clearly explain how participants were recruited, whether sampling was random, convenience-based, community-based, or clinic-based. This affects external validity.

Thank you for this important comment. We agree that the participant recruitment process should be described more clearly to improve the interpretation of the study’s external validity. In response to this suggestion, we added a detailed explanation to the Methods section specifying that participants were recruited from the community using a convenience sampling approach among adults living in Afyonkarahisar who volunteered to participate in the study.

The authors should specify whether participants with neurological disease, psychiatric illness, dementia, major depression, substance use disorder, traumatic brain injury, visual impairment, or medications affecting cognition were excluded.

Thank you for this important comment. In response to this suggestion, we clarified the inclusion and exclusion criteria in the Methods section. Participants with physician-diagnosed psychiatric disorders, physical, mental, or cognitive impairments, conditions affecting eating behavior, pregnancy or breastfeeding, and the use of physician-prescribed medications were not included in the study. These criteria were applied to minimize potential factors that could influence dietary assessment and cognitive performance.

TMT-A usually requires connecting numbers 1-25, not 1-23. The authors should verify the Turkish version used and, if necessary, correct the description.

Thank you for this important observation. We have carefully reviewed the version of the Trail Making Test used in the present study and confirmed that the standard Turkish adaptation of the test was administered. Accordingly, in Part A, participants were required to connect numbers from 1 to 25 in ascending order. The manuscript has been corrected to accurately reflect this procedure and ensure consistency with the validated version of the instrument.

The manuscript says delayed recall was assessed “approximately 30 or 40 minutes later.” This should be standardized and reported precisely because delayed recall is sensitive to interval duration.

Thank you for this important comment. We agree that the delayed recall interval should be reported precisely due to its sensitivity to memory performance. In the revised manuscript, we have corrected the description and clarified that the delayed recall was assessed approximately 30 minutes after the learning phase, with a maximum interval of 30 minutes applied consistently across all participants.

The Discussion largely repeats the Results and cites supportive literature without critically evaluating inconsistencies, confounding, reverse causation, or limitations of the cognitive tools.

We thank the reviewer for this constructive criticism. The Discussion section has been substantially revised. Redundant restatement of results has been reduced, and critical evaluation of findings has been expanded. Inconsistencies between our results and the existing literature are now explicitly discussed with potential explanations including confounding factors, reverse causation, and methodological differences across studies. A paragraph acknowledging the limitations of the cognitive tools employed (Oktem-VMPT and TMT) — including cross-sectional sensitivity and the absence of neuroimaging validation — has also been added.

The authors should briefly discuss plausible mechanisms linking dietary patterns to cognition: vascular health, inflammation, oxidative stress, insulin sensitivity, gut microbiota, polyphenols, omega-3 fatty acids, and blood pressure regulation.

We thank the reviewer for this important suggestion. A dedicated mechanistic paragraph has been added to the Discussion, covering the following pathways: reduction of neuroinflammation and oxidative stress via polyphenols and omega-3 fatty acids, improvement of cerebral blood flow through blood pressure regulation and vascular protection, modulation of insulin sensitivity and its downstream effects on hippocampal neurogenesis, and gut microbiota–brain axis signaling. These mechanisms are now linked specifically to the cognitive domains observed in our study — verbal memory, learning, and executive functions.

Several references are local theses or non-core sources, while key high-impact literature on Mediterranean diet, DASH, MIND diet, cognitive aging, and dementia prevention is underused. More recent systematic reviews, meta-analyses, and prospective cohort/intervention studies should be added.

We acknowledge this limitation in our reference list. Several local thesis citations and low-impact sources in the Discussion have been replaced with peer-reviewed, high-impact publications including recent systematic reviews, meta-analyses, and prospective cohort studies on Mediterranean diet, DASH diet, MIND diet, cognitive aging, and dementia prevention. The updated reference list prioritizes international, indexed sources.

Minör

In the abstract, use p < 0.05, not p<0,05.

Thank you for this observation. We have revised the manuscript to ensure consistency with statistical reporting standards by correcting the formatting of p-values (decimal separators) throughout the text.

Remove the extra space before the comma in “TMT-A) , Ability…”.

Thank you for this comment. The manuscript has been carefully revised and the spacing error before the comma in “TMT-A), Ability…” has been corrected throughout the text.

3. Add a period after “p < 0.05)” in the abstract.

Thank you for this comment. The manuscript has been revised and a period has been added after “p < 0.05)” in the abstract to ensure correct punctuation.

Keywords include “diet” twice. Remove duplication.

Thank you for this comment. We have revised the keywords to remove duplication and improve clarity. The keywords “diet, Mediterranean” and “diet, DASH” have been corrected to “Mediterranean diet” and “DASH diet,” respectively.

“mediterranean” and “dash” should be capitalized properly: Mediterranean, DASH.

Thank you for this observation. The keywords have been revised to ensure proper capitalization, and “Mediterranean” and “DASH” are now correctly formatted in accordance with standard terminology.

“Processing Speed Based on Visual Scanning Ability” and “Ability to Shift Set Between Stimulus Sets and Follow Sequencing” are too long for repeated use; define once and then use TMT-A and TMT-B.

Thank you for this comment. The manuscript has been revised so that the full terms are defined at first use and the abbreviations TMT-A and TMT-B are used consistently thereafter.

“It only be shared” in the Data Availability Statement is grammatically incorrect; revise to “It will only be shared…”

Thank you for this observation. The Data Availability Statement has been revised to correct the grammatical error, and “It only be shared” has been updated to “It will only be shared” for clarity and correctness.

Author Contributions ends with .”. Remove the extra period.

Thank you for this observation. The Author Contributions section has been carefully revised to correct punctuation and formatting inconsistencies, including the missing semicolon and overall standardization of the section.

The abbreviation list is incorrect: MEDAS is listed together with DASH-Q. Each abbreviation should be defined separately.

Thank you for this observation. The Abbreviations section has been revised and formatted correctly so that each abbreviation is defined separately, including MEDAS and DASH-Q.

Table formatting is poor, with broken words and crowded values. Tables should be reformatted before publication.

Thank you for this comment. Due to the revisions made in response to the comments regarding the Results section, all tables were updated accordingly. During this revision process, particular attention was paid to table formatting, readability, and layout throughout the manuscript.

“Oktem-VMPT” contains a formatting artifact and should be corrected.

Thank you for this observation. The terminology and abbreviation related to the Oktem Verbal Memory Processes Test have been standardized throughout the manuscript, and all instances have been revised consistently as “Oktem-VMPT”

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review the manuscript of Seher Abasız and Muge Arslanet investigating how adherence to DASH and Mediterranean diet influence the cognitive performance in a cross-sectional study. The Introduction is well written and provides enough background information to allow the reader to understand the topic, the research gap and the research question. The study design is appropriate and the results are interesting. Study implications and limitations are presented, and the conclusions are supported by the results.

However, the description of the methodology, of the results and the discussion of these results in the context of existing literature must be improved. Please find my comments below.

Introduction

Third paragraph is redundant. Please delete it.

Methods

Please describe the inclusion and exclusion criteria used and how the participants were approached/invited to participate. Were consecutive participants included? How many were invited and how many agreed to participate. If a description of the ones who refused to participate is available, please add it in the Results section.
How was data collected? Using an electronical tool or a pen-and-paper form? Was it self-administered or administered by personnel involved in the study?
Please clarify what do you understand by meal pattern, and if weight, waist and hip circumferences were measured and how. How was the presence of any diagnosed chronic disease and prescribed medication established?
Please briefly describe how the scores for LMT were calculated.
Education and socioeconomic status are important determinants of dietary choices. Were these data collected? If not, this should be added as a study limitation.
Second paragraph of section 2.1. would be more suitable for the statistical analysis section as it describes sample size calculation.

Results

How many participants were included in the analysis? Were there any excluded due to incomplete data or exclusion criteria? Please add this information. Also, a study flow chart would answer these questions.
Please add results on other characteristics assessed - waist, waist-to-hip ratio, chronic diseases, medication, alcohol consumption, smoking, dietary supplements used, alcohol consumption, meal pattern. It helps assess the sample structure and other lifestyle components also associated with dietary choices.
Please add in the header of the Table 1 the number for male, female and total sample
What statistical test was used for the results presented on page 8, lines 256-270 and Table 3? It looks like linear regression, but this should be described in the Methods section. Also please state how the data which did not meet the assumption of normal distribution was handled for the linear regression analysis. Also, the adjustment for age and sex would be appropriate for this analysis. If available, cofounding factors for which this analysis should be adjusted are education and economic factors.
Could you explain in the Methods section why in some instances correlations and in other linear regression were used for the assessment of relationship between MEDAS and DASH scores with other variables?

Discussion

A brief discussion on the mechanisms potentially explaining the positive effect of DASH and Mediterranean diets on cognitive performance should be added.
There are studies available which assessed the effect of Mediterranean diet with executive functions also using the Trail Making test. Please discuss the reasons for different results. See Koutsonida M, et al. Adherence to Mediterranean Diet and Cognitive Abilities in the Greek Cohort of Epirus Health Study. Nutrients. 2021 Sep 25;13(10):3363.
Consider moving the study strength and limitations before Conclusion section.

Author Response

Review 2

Introduction

Third paragraph is redundant. Please delete it.

Thank you for your comment. The redundant third paragraph in the Introduction section has been deleted as suggested.

Methods

Please describe the inclusion and exclusion criteria used and how the participants were approached/invited to participate. Were consecutive participants included? How many were invited and how many agreed to participate. If a description of the ones who refused to participate is available, please add it in the Results section.

Thank you for this valuable comment. We have revised the Methods section (Section 2.1) accordingly. The inclusion and exclusion criteria were already described in the original manuscript. In the revised version, we have added information regarding how participants were approached: they were recruited through face-to-face invitations at community centers, public spaces, and university campus areas in Afyonkarahisar. Approximately 700–750 individuals were approached, of whom 600 agreed to participate. Between 100 and 150 individuals declined to participate, primarily citing time constraints and unwillingness to complete the lengthy questionnaire battery. Participation was voluntary and non-consecutive. This information has been added to Section 2.1 of the revised manuscript.

How was data collected? Using an electronical tool or a pen-and-paper form? Was it self-administered or administered by personnel involved in the study?

Thank you for this comment. As stated in Section 2.2, all data were collected face-to-face by the researchers. A pen-and-paper format was used for all forms and scales, which were administered in person by trained personnel involved in the study. We have clarified this in the revised manuscript.

Please clarify what do you understand by meal pattern, and if weight, waist and hip circumferences were measured and how. How was the presence of any diagnosed chronic disease and prescribed medication established?

Thank you for this comment. We have clarified these points in the revised manuscript (Section 2.2.1). Meal pattern referred to which meals participants skipped during the day. Anthropometric measurements were taken by the researcher in a clinical setting using a TANITA body composition analyzer for weight and BMI, and a standard measuring tape for waist and hip circumferences. The presence of chronic disease and medication use was based on participant self-report.

Please briefly describe how the scores for LMT were calculated.

Thank you for this comment. We have clarified in Section 2.2.3 that the TMT score was based on the time taken to complete each form, recorded in seconds. If errors were made, participants were immediately corrected and asked to continue without stopping the clock. This information has been added to the revised manuscript.

Education and socioeconomic status are important determinants of dietary choices. Were these data collected? If not, this should be added as a study limitation.

Thank you for this important comment. We would like to clarify that both educational status and income level (as a proxy for socioeconomic status) were collected as part of the sociodemographic data form and were included in the analyses. As shown in Table 1 and Table 3, statistically significant differences in DASH and MEDAS scores were observed according to both educational status and income level. Therefore, we do not consider this a limitation of the study.

Second paragraph of section 2.1. would be more suitable for the statistical analysis section as it describes sample size calculation.

Thank you for this suggestion. The sample size calculation paragraph has been moved from Section 2.1 to the Statistical Analysis section (Section 2.3) in the revised manuscript.

Results

How many participants were included in the analysis? Were there any excluded due to incomplete data or exclusion criteria? Please add this information. Also, a study flow chart would answer these questions.

Thank you for this comment. We have added a statement at the beginning of the Results section clarifying that all 600 participants who agreed to participate completed the data collection process and were included in the final analysis. No participants were excluded due to incomplete data or failure to meet the eligibility criteria. Given that there were no exclusions, we believe a flow chart would not provide additional information beyond what is already stated in the text; however, we are happy to add one if the reviewer considers it necessary.

Please add results on other characteristics assessed - waist, waist-to-hip ratio, chronic diseases, medication, alcohol consumption, smoking, dietary supplements used, alcohol consumption, meal pattern. It helps assess the sample structure and other lifestyle components also associated with dietary choices.

Thank you for this suggestion. We would like to clarify that the characteristics mentioned are already reported in the manuscript. Smoking status, alcohol consumption, chronic disease status, medication use, nutritional supplement use, meal skipping, and daily water intake are presented in Table 1. Waist circumference and waist-to-hip ratio are presented in Table 2, along with other anthropometric measurements, stratified by gender. We believe the current tables provide a comprehensive description of the sample structure and relevant lifestyle components.

Please add in the header of the Table 1 the number for male, female and total sample

Thank you for this comment. We would like to clarify that the sample sizes for male (n=257), female (n=343), and total (n=600) are already indicated in the column headers of Table 1 in the original manuscript.

What statistical test was used for the results presented on page 8, lines 256-270 and Table 3? It looks like linear regression, but this should be described in the Methods section. Also please state how the data which did not meet the assumption of normal distribution was handled for the linear regression analysis. Also, the adjustment for age and sex would be appropriate for this analysis. If available, cofounding factors for which this analysis should be adjusted are education and economic factors.

Thank you for this comment. Multiple linear regression was used for the analysis presented in Table 4, and this has been clearly stated in the Methods section (Section 2.3). The regression models were adjusted for age, sex, educational status, and income level as covariates. Although some variables did not meet the assumption of normal distribution, multiple linear regression was considered appropriate given the large sample size (n=600), as the method is robust to normality violations under such conditions. This clarification has been added to the Methods section.

Could you explain in the Methods section why in some instances correlations and in other linear regression were used for the assessment of relationship between MEDAS and DASH scores with other variables?

Thank you for this comment. We have added a clarification to the Methods section (Section 2.3) explaining the rationale for using both approaches. Spearman's rank-order correlation was used to examine bivariate associations between DASH and MEDAS scores and continuous variables without controlling for other factors. Multiple linear regression was subsequently used to assess the independent predictive effects of DASH and MEDAS scores on cognitive outcomes while simultaneously adjusting for potential confounding variables, including age, sex, educational status, and income level.

Discussion

A brief discussion on the mechanisms potentially explaining the positive effect of DASH and Mediterranean diets on cognitive performance should be added.

A dedicated mechanistic paragraph has been added to the Discussion as described above in our response to Reviewer 1, Comment 19. This paragraph systematically addresses the neuroprotective pathways through which the DASH and Mediterranean diets may exert their effects on cognitive performance, including vascular, anti-inflammatory, metabolic, and gut–brain axis mechanisms.

There are studies available which assessed the effect of Mediterranean diet with executive functions also using the Trail Making test. Please discuss the reasons for different results. See Koutsonida M, et al. Adherence to Mediterranean Diet and Cognitive Abilities in the Greek Cohort of Epirus Health Study. Nutrients. 2021 Sep 25;13(10):3363.

We thank the reviewer for this specific and relevant recommendation. The Koutsonida et al. (2021) study has been incorporated into the Discussion. In that study, higher Mediterranean diet adherence was associated with better TMT performance in a Greek cohort, which partially aligns with our bivariate correlation findings. However, in our multiple linear regression analysis, DASH and MEDAS scores did not emerge as independent predictors of TMT performance after controlling for age, education, income, and smoking status. We now explicitly discuss potential reasons for this discrepancy: their sample consisted of older adults in whom vascular and neuroinflammatory pathways may already produce detectable cognitive differences amenable to dietary modulation, whereas our younger mean age (36 years) and generally intact TMT performance may have limited detectable diet-related variation. Furthermore, the stronger predictive effects of age and educational level on TMT performance in our sample likely attenuated the independent contribution of diet in multivariate models.

Consider moving the study strength and limitations before Conclusion section.

Thank you for this suggestion. The strengths and limitations of the study have been moved to a separate section preceding the Conclusions section in the revised manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

The paper present a timely and well-powered cross-sectional study evaluating the associations between adherence to the DASH and Mediterranean diets and distinct cognitive domains in a Turkish adult cohort. This work offers valuable population-level insights that reinforce the critical role of nutritional patterns in preserving cognitive health.

Major comments：

The regression models presented in Table 3 appear to only include the dietary indices as predictors for cognitive outcomes. Given that the earlier descriptive tables clearly demonstrate significant associations between demographic factors such as age, sex, and BMI with both dietary adherence and likely cognitive performance, it is imperative to include these variables as covariates in the models to adequately adjust for potential confounding effects.
A substantial number of correlation coefficients are reported across various cognitive subdomains and two distinct dietary scales in the correlational analysis. Conducting such a large array of unadjusted statistical tests inherently increases the risk of Type I errors. It is highly recommended to apply an appropriate multiple comparison correction method, such as the False Discovery Rate or Bonferroni correction, to ensure the statistical robustness of these findings.
While the current methodology relies on univariate correlations and standard multiple regression, the relationship between nutritional patterns and brain health is inherently multidimensional. Investigating this dataset using multivariate statistical frameworks, such as canonical correlation analysis, could provide a more sophisticated understanding of how specific clusters of dietary habits map onto distinct cognitive profiles, thereby extracting more nuanced features from your matrices.
The discussion currently describes the general cognitive benefits of the diets but largely lacks a mechanistic exploration of why specific domains, such as visual scanning or verbal memory, are distinctly affected. Expanding this section to bridge the observed behavioral metrics with potential central neurobiological mechanisms, perhaps drawing on how nutrient profiles influence underlying brain network organization, would substantially elevate the scientific narrative of the manuscript.
The discussion could be strengthened by situating the neuroprotective mechanisms of healthy dietary patterns more fully with citations to representative studies exploring molecular pathways in cognitive decline. For example, considering how nutritional interventions might intersect with complex regulatory mechanisms like alternative splicing to slow neurodegeneration would provide a deeper mechanistic context for your behavioral findings https://doi.org/10.1016/j.arr.2026.103133.

The methodology section details a power analysis indicating a minimum required sample size of 181, yet the final recruited cohort includes 600 participants. While a larger sample size is generally advantageous for statistical power, the manuscript would benefit from a brief explanation regarding the sampling strategy that led to this final number and a comment on whether the study was powered to detect smaller effect sizes within specific demographic subgroups.

Minor comments

There is a minor inconsistency in the formatting of decimal separators throughout the manuscript, such as using a comma in the abstract for p-values and a period in the main text. Please carefully proofread the document to ensure a uniform decimal format is applied consistently across all sections and tables.

Author Response

The regression models presented in Table 3 appear to only include the dietary indices as predictors for cognitive outcomes. Given that the earlier descriptive tables clearly demonstrate significant associations between demographic factors such as age, sex, and BMI with both dietary adherence and likely cognitive performance, it is imperative to include these variables as covariates in the models to adequately adjust for potential confounding effects.

Thank you for this important suggestion. In response to the reviewer’s comment, the regression analyses presented in Table 3 have been revised to include potential confounding variables as covariates. Demographic and anthropometric variables, including age, sex and BMI, were added to the regression models to better account for their potential effects on both dietary adherence and cognitive performance. The revised analyses are now presented in the updated Table 3, and the Methods and Results sections have been revised accordingly.

A substantial number of correlation coefficients are reported across various cognitive subdomains and two distinct dietary scales in the correlational analysis. Conducting such a large array of unadjusted statistical tests inherently increases the risk of Type I errors. It is highly recommended to apply an appropriate multiple comparison correction method, such as the False Discovery Rate or Bonferroni correction, to ensure the statistical robustness of these findings.

Thank you for this important methodological suggestion. We agree that conducting multiple statistical comparisons may increase the likelihood of Type I error. Therefore, we applied Bonferroni correction to the correlational analyses involving cognitive subdomains and dietary scale scores. The corrected significance threshold (p = 0.002) and the updated findings are now reported in the newly added Table 5 and described in the revised manuscript.

While the current methodology relies on univariate correlations and standard multiple regression, the relationship between nutritional patterns and brain health is inherently multidimensional. Investigating this dataset using multivariate statistical frameworks, such as canonical correlation analysis, could provide a more sophisticated understanding of how specific clusters of dietary habits map onto distinct cognitive profiles, thereby extracting more nuanced features from your matrices.

We thank the reviewer for this methodologically valuable suggestion. Canonical correlation analysis (CCA) was considered as an analytical approach to explore the multivariate relationships between dietary patterns and cognitive profiles. However, upon examination of the dataset, the necessary statistical assumptions for CCA — including multivariate normality and sufficient sample size relative to the number of variables — were not adequately met. Proceeding with CCA under these conditions would risk producing unstable and misleading canonical variates, potentially compromising the integrity of the findings.

As an alternative, multiple regression analyses were conducted with all relevant covariates included simultaneously, allowing for the examination of the joint contribution of dietary and demographic variables to each cognitive outcome. We acknowledge that this approach does not fully capture the multidimensional nature of the diet–cognition relationship as CCA would, and we have noted this as a limitation of the present study, recommending that future studies with larger and more normally distributed samples consider multivariate frameworks such as CCA or structural equation modeling.

The discussion currently describes the general cognitive benefits of the diets but largely lacks a mechanistic exploration of why specific domains, such as visual scanning or verbal memory, are distinctly affected. Expanding this section to bridge the observed behavioral metrics with potential central neurobiological mechanisms, perhaps drawing on how nutrient profiles influence underlying brain network organization, would substantially elevate the scientific narrative of the manuscript.

We thank the reviewer for this valuable suggestion. The revised Discussion now includes domain-specific mechanistic explanations. For verbal memory and learning — reflected in Oktem-VMPT subscales — we discuss how omega-3 fatty acids and polyphenols support hippocampal neurogenesis, synaptic plasticity, and long-term potentiation. For executive functions and visual scanning speed — reflected in TMT performance — we discuss the role of vascular health, prefrontal cortex integrity, and dopaminergic signaling. This domain-specific framing bridges our behavioral findings with underlying neurobiological processes and also addresses the differential pattern of results observed across the two cognitive instruments.

The discussion could be strengthened by situating the neuroprotective mechanisms of healthy dietary patterns more fully with citations to representative studies exploring molecular pathways in cognitive decline. For example, considering how nutritional interventions might intersect with complex regulatory mechanisms like alternative splicing to slow neurodegeneration would provide a deeper mechanistic context for your behavioral findings https://doi.org/10.1016/j.arr.2026.103133.

We thank the reviewer for directing our attention to this important mechanistic dimension. A paragraph discussing how dietary bioactive compounds — particularly polyphenols and omega-3 fatty acids found abundantly in both dietary patterns — may influence RNA splicing regulatory mechanisms involved in neurodegeneration has been incorporated into the Discussion, with reference to the suggested article (doi: 10.1016/j.arr.2026.103133). This molecular-level context strengthens the mechanistic narrative of our findings and highlights a promising direction for future research.

The methodology section details a power analysis indicating a minimum required sample size of 181, yet the final recruited cohort includes 600 participants. While a larger sample size is generally advantageous for statistical power, the manuscript would benefit from a brief explanation regarding the sampling strategy that led to this final number and a comment on whether the study was powered to detect smaller effect sizes within specific demographic subgroups.

Thank you for this valuable comment. Although the minimum required sample size was calculated as 181 participants based on the power analysis, a larger sample was intentionally targeted to improve statistical power, increase the representativeness of the study population, compensate for possible incomplete responses, and allow more reliable subgroup evaluations. This clarification has been added to the Methods section of the revised manuscript.

Minör

There is a minor inconsistency in the formatting of decimal separators throughout the manuscript, such as using a comma in the abstract for p-values and a period in the main text. Please carefully proofread the document to ensure a uniform decimal format is applied consistently across all sections and tables.

Thank you for this observation. The manuscript has been carefully proofread to ensure consistency in decimal formatting throughout all sections and tables. All decimal separators were standardized using periods in accordance with international scientific writing conventions.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review the revised version of the manuscript of Seher Abasız and Muge Arslanet investigating how adherence to DASH and Mediterranean diet influence the cognitive performance in a cross-sectional study.

The manuscript has been substantially improved. However, I still have several small comments

In the Methods section using physician prescribed medication is listed as an exclusion criterion. However, in Table 1 the use of chronic medication is listed. Please clarify why these participants were included in the analysis.
Please add in the Methods section from where the USB error score and IST scores are derived. They appear in the Results without any previous mention, and this is confusing.
Table 4 and associated text should be improved. It seems like some variables are listed on second column instead of column 1 (TMT performance). This will improve the readability. Also please clarify below this table and/or in text what is meant by models 1, 2, 3 and 4. They appear in text but not in Table 4 as such.
I suggest moving the discussion on DASH and MEDAS scores and cognitive performance earlier in the Discussion section as it represents the main study objective (according to the title).
Conclusion

The findings reveal that higher adherence to the DASH and Mediterranean diets was associated with better verbal memory and learning performance, as reflected in Oktem-VMPT subscales, while associations with executive function and visual scanning speed measured by the TMT were attenuated after adjustment for sociodemographic covariates.

This statement is not supported by the results. According to Table 4 the association of Oktem-VMPT with DASH and MEDAS scores were also attenuated in the regression model including the sociodemographic variables.

Author Response

Reviewer 2

In the Methods section using physician prescribed medication is listed as an exclusion criterion. However, in Table 1 the use of chronic medication is listed. Please clarify why these participants were included in the analysis.

We thank the reviewer for this careful observation. We acknowledge the inconsistency between the stated exclusion criterion and the data presented in Table 1. Upon review, 26 participants (4.3%) reported regular use of physician-prescribed medications. These individuals were retained in the analysis as their medications were not deemed likely to influence dietary adherence scores or cognitive test performance. To address this concern transparently, we conducted a sensitivity analysis comparing DASH and MEDAS scores between participants with and without regular medication use. No statistically significant differences were found (DASH: p=0.988; MEDAS: p=0.415), indicating that the inclusion of these participants did not materially affect the study outcomes. A clarifying sentence has been added to the Methods section accordingly.

Please add in the Methods section from where the USB error score and IST scores are derived. They appear in the Results without any previous mention, and this is confusing.

We thank the reviewer for this observation. In the revised manuscript, the abbreviation "USB" (Uzun Süreli Bellek), which had been introduced by the translator, has been replaced with "LTM" (Long-Term Memory), and all IST abbreviations (IST-A, IST-B, IST-Total) have been replaced with their English equivalents (TMT-A, TMT-B, Total TMT) throughout the manuscript. Explanatory footnotes for these abbreviations have also been added to the relevant tables for clarity.

Table 4 and associated text should be improved. It seems like some variables are listed on second column instead of column 1 (TMT performance). This will improve the readability. Also please clarify below this table and/or in text what is meant by models 1, 2, 3 and 4. They appear in text but not in Table 4 as such.

We thank the reviewer for this observation. The column structure of Table 4 reflects the intended formatting of the journal template, in which certain variables appear in the second column by design. However, to improve readability, we have reviewed the table layout and made minor adjustments where possible. Additionally, a footnote has been added to Table 4 clarifying the correspondence between model numbers and dependent variables: Model 1: Oktem-VMPT Total Recall; Model 2: TMT-A; Model 3: TMT-B; Model 4: Total TMT.

I suggest moving the discussion on DASH and MEDAS scores and cognitive performance earlier in the Discussion section as it represents the main study objective (according to the title).

The discussion of DASH and MEDAS scores in relation to cognitive performance has been moved to the beginning of the Discussion section, as it represents the primary objective of the study as stated in the title.

Conclusion

Thank you for this important observation. The reviewer is correct that, according to Table 4, the associations between both DASH and MEDAS scores and Oktem-VMPT subscales were also attenuated and did not reach statistical significance in the multiple linear regression model after adjustment for sociodemographic covariates. The original statement was therefore inaccurate and misleading.

The Conclusion section has been revised accordingly. The corrected statement now reads:

"Higher adherence to the DASH and Mediterranean diets was associated with better verbal memory and learning performance in bivariate analyses; however, these associations, along with those with executive function and visual scanning speed measured by the TMT, were attenuated after adjustment for sociodemographic covariates in multivariate models, suggesting that the observed diet–cognition relationship may be partly mediated or confounded by factors such as age, educational level, and income."

We apologize for the oversight and thank the reviewer for the careful reading.

Reviewer 3 Report

Comments and Suggestions for Authors

The revised manuscript effectively addresses the previous concerns, and the authors are to be commended for conducting a methodologically sound study that adds valuable scientific insight into the cognitive benefits of the Mediterranean and DASH diets. I am largely satisfied with the work and recommend the manuscript for publication following the resolution of three minor points.

In the Results section, while the multiple linear regression models demonstrate no multicollinearity, the reporting of variance explained could be more streamlined. It would be beneficial to report the adjusted R-squared values more prominently than the unadjusted ones to provide a more conservative and accurate reflection of the models explanatory power.
The discussion touches upon the limitations of the DASH-Q capturing only recent dietary behaviors, but it would be helpful to briefly acknowledge whether this temporal limitation might disproportionately affect certain demographic groups in your sample. For instance, younger adults might have more variable week-to-week eating patterns, which could introduce specific confounding factors that warrant a brief mention.
The discussion regarding future directions could be strengthened by situating the methodological background more fully with citations to representative multimodal integration studies. Specifically, when mentioning the future incorporation of structural MRI and biomarkers, the authors could draw parallels to recent methodological advancements in multimodal fusion and radiomics that have successfully integrated complex clinical and imaging data for risk stratification in other health domains, demonstrating the potential of these advanced analytical frameworks to elucidate complex diet-cognition pathways e.g., https://doi.org/10.53388/2026926007.
Table 4 is quite extensive and somewhat dense. To improve readability for the broad readership of the journal, please consider whether some of the non-significant covariate rows could be moved to supplementary materials, or at least ensure that the significant predictors are more visually distinct to help readers quickly identify key findings.

Author Response

Reviewer 3

In the Results section, while the multiple linear regression models demonstrate no multicollinearity, the reporting of variance explained could be more streamlined. It would be beneficial to report the adjusted R-squared values more prominently than the unadjusted ones to provide a more conservative and accurate reflection of the models explanatory power.

Thank you for this constructive suggestion. We agree that reporting the adjusted R² more prominently provides a more conservative and accurate reflection of the models' explanatory power.

In response, we have revised the Results section to lead with the adjusted R² values in both the narrative text and the Table 4 footnote. Specifically, the percentage of variance explained now reflects the adjusted R² (e.g., "Model 1 explained 77.7% of the variance, adjusted R²=0.777"), with the unadjusted R² retained in parentheses for completeness. The same ordering has been applied consistently in the Table 4 footnote.

The discussion touches upon the limitations of the DASH-Q capturing only recent dietary behaviors, but it would be helpful to briefly acknowledge whether this temporal limitation might disproportionately affect certain demographic groups in your sample. For instance, younger adults might have more variable week-to-week eating patterns, which could introduce specific confounding factors that warrant a brief mention.

Thank you for this helpful suggestion. We agree that the temporal limitation of the DASH-Q may not affect all demographic groups equally.

Accordingly, we have added the following sentence to the Limitations section:

"Furthermore, this temporal limitation may disproportionately affect younger adults in the sample, whose week-to-week dietary patterns tend to be more variable compared to older age groups. This variability may introduce additional measurement error in younger participants' DASH scores, potentially confounding age-related comparisons of diet quality and cognitive performance."

The discussion regarding future directions could be strengthened by situating the methodological background more fully with citations to representative multimodal integration studies. Specifically, when mentioning the future incorporation of structural MRI and biomarkers, the authors could draw parallels to recent methodological advancements in multimodal fusion and radiomics that have successfully integrated complex clinical and imaging data for risk stratification in other health domains, demonstrating the potential of these advanced analytical frameworks to elucidate complex diet-cognition pathways e.g., https://doi.org/10.53388/2026926007.

We thank the reviewer for this valuable suggestion. In accordance with this recommendation, we have expanded the future directions paragraph in the Discussion section to include reference to recent methodological advances in multimodal fusion and radiomics. Specifically, we have added the following sentence:

"In this regard, recent advances in multimodal radiomics have demonstrated that integrating complementary imaging modalities can substantially enhance predictive accuracy for complex clinical outcomes; adapting such analytical frameworks to diet–cognition research may offer novel mechanistic insights [120]."

The corresponding reference (Yang et al., 2026; https://doi.org/10.53388/2026926007) has been added to the reference list as [120]. We believe this addition appropriately contextualizes the potential of multimodal analytical approaches for future diet–cognition research, as suggested by the reviewer.

Table 4 is quite extensive and somewhat dense. To improve readability for the broad readership of the journal, please consider whether some of the non-significant covariate rows could be moved to supplementary materials, or at least ensure that the significant predictors are more visually distinct to help readers quickly identify key findings.

We thank the reviewer for this suggestion. We note that statistically significant predictors in Table 4 are already marked with asterisks (* p<0.05, ** p<0.01, *** p<0.001), which we believe provides sufficient visual distinction for readers to identify key findings. Nevertheless, we have additionally bolded the significant rows to further enhance readability.