Edwin Thanarajah S.
Midlife Dietary Quality and Body Composition Relevance for Brain Connectivity and Cognitive Performance in Later Life.
JAMA Netw Open. 2025 Mar 3;8(3):e250181.
PubMed.
The study by Jensen and colleagues is carefully done and in line with the literature showing that midlife obesity—and waist-hip ratio, which is a better measure—are associated with future lower cognition. It also shows for the first time an association with good hippocampal connectivity and higher white-matter integrity. The results on obesity or waist-hip ratio add another layer of evidence to the literature, showing a mechanism of action through brain changes, and reinforce the conclusion that midlife weight is important. In particular, white-matter changes mediated the association with cognition.
The diet component is less clear. Diet is measured through the alternate healthy eating index, incorporating not only healthy eating—vegetables, fruit, whole grains, nuts and legumes, long-chain omega-3 fatty acids that include docosahexaenoic acid and eicosapentaenoic acid, and polyunsaturated fatty acids—but also moderate alcohol, low sugar and low sodium. Diet was associated with the hippocampal but not the white-matter findings. This diet may act directly on the brain, be related to weight or waist-hip ratio, or reflect alcohol intake or sodium that increases hypertension, or a combination of these. It is impossible to know. It clearly is a wide measure with better scores reflecting a healthier lifestyle.
I do not agree that because the participants did not have alcohol dependence, intake was unimportant, as the effect of alcohol on cognition can be important even in the absence of dependence. I agree with the authors’ caveat that the cohort was mainly male, white, British who survived to middle age. Even so, the findings are from a large cohort and add to our understanding, particularly that lower white-matter integrity mediates the effect (i.e., is the pathway of effect) of obesity/waist-hip ratio on cognition.
I’d like to note a COI with the Whitehall paper as some of its authors are colleagues.
We would all like to believe that we could control our own health by eating healthy. Dietary epidemiological studies such as this one of Jensen et al. sought to validate the protective effects of a healthy diet on cognition and its presumed neurophysiological correlates. The authors found that those who reported following quality diets in midlife and who had lower waist-to-hip ratios had better functional and structural MRI measurements as well as slightly better cognitive performance. While these results are gratifying, the lack of attention to the socioeconomic status of participants lessens my confidence for making causal claims about the benefits of diet on later life cognition. Dietary choices are strongly influenced by socioeconomic circumstances, which themselves are unquestionably strongly related to all health outcomes including cognitive health. If the authors had been able to control for socioeconomics, which they do not mention, would the associations of diet be attenuated?
The straightforward observations of Dhana et al. showing that plasma NfL concentrations were associated with a cardiovascular health score shows that the known effects of vascular disease on brain health can be proxied by the plasma biomarkers. This is reinforced by an article that just appeared in Neurology, which also showed in a large multidisciplinary cohort that NfL and AD specific biomarkers associated with brain volume loss and white-matter hyperintensity burden (Sanchez et al., 2025).
One of the valuable insights that has emerged in the past few years about the relationship between cerebrovascular disease and brain health is that it is not just overt strokes that matter. Cerebrovascular disease also exerts its pathologic effects via gradual, insidious, covert means, that are proxied by brain volume loss and nonspecific plasma biomarkers, such as NfL.
References:
Sanchez E, Coughlan GT, Wilkinson T, Ramirez J, Mirza SS, Baril AA, Dilliott AA, Frank A, Lang AE, Hassan A, Pollock BG, Scott CJ, Marras C, Fischer CE, Seitz D, Andriuta D, Dowlatshahi D, Grimes DA, Tang-Wai DF, Sahlas DJ, Rogaeva EA, Finger E, Robinson JF, Tan K, Binns MA, Tartaglia MC, Borrie MJ, Strong MJ, Ozzoude M, Nanayakkara ND, Goncalves RA, Bartha R, Hegele RA, Farhan SM, Black SE, Kumar S, Symons SP, Haddad SM, Pasternak SH, Arnott SR, Rajji TK, Steeves T, Swardfager W, Ashton NJ, Kvartsberg H, Zetterberg H, Munoz DP, Masellis M, ONDRI Investigators.
Association of Plasma Biomarkers With Longitudinal Atrophy and Microvascular Burden on MRI Across Neurodegenerative and Cerebrovascular Diseases.
Neurology. 2025 Apr 8;104(7):e213438. Epub 2025 Mar 10
PubMed.
We thank Drs. Livingston and Knopman for their comments on our study and their evaluation. We agree with Dr Knopman that socioeconomic status is strongly related to dietary quality. Therefore, we’d like to clarify that we did correct all analyses for education status. Thus, in the "Statistical Analysis" section of the paper we state "All analyses were adjusted for age, sex, years of education, MRI scanner model, physical activity, and the Montreal Cognitive Assessment score, all of which were measured at the time of MRI scan. Resting-state fMRI analyses were corrected for head motion and voxelwise gray matter density. AHEI-2010 analyses were adjusted for total energy intake (kcal/d) (eMethods 7 in Supplement 1).
References:
Jensen DE, Ebmeier KP, Akbaraly T, Jansen MG, Singh-Manoux A, Kivimäki M, Zsoldos E, Klein-Flügge MC, Suri S.
Association of Diet and Waist-to-Hip Ratio With Brain Connectivity and Memory in Aging.
JAMA Netw Open. 2025 Mar 3;8(3):e250171.
PubMed.
Comments
UCL
The study by Jensen and colleagues is carefully done and in line with the literature showing that midlife obesity—and waist-hip ratio, which is a better measure—are associated with future lower cognition. It also shows for the first time an association with good hippocampal connectivity and higher white-matter integrity. The results on obesity or waist-hip ratio add another layer of evidence to the literature, showing a mechanism of action through brain changes, and reinforce the conclusion that midlife weight is important. In particular, white-matter changes mediated the association with cognition.
The diet component is less clear. Diet is measured through the alternate healthy eating index, incorporating not only healthy eating—vegetables, fruit, whole grains, nuts and legumes, long-chain omega-3 fatty acids that include docosahexaenoic acid and eicosapentaenoic acid, and polyunsaturated fatty acids—but also moderate alcohol, low sugar and low sodium. Diet was associated with the hippocampal but not the white-matter findings. This diet may act directly on the brain, be related to weight or waist-hip ratio, or reflect alcohol intake or sodium that increases hypertension, or a combination of these. It is impossible to know. It clearly is a wide measure with better scores reflecting a healthier lifestyle.
I do not agree that because the participants did not have alcohol dependence, intake was unimportant, as the effect of alcohol on cognition can be important even in the absence of dependence. I agree with the authors’ caveat that the cohort was mainly male, white, British who survived to middle age. Even so, the findings are from a large cohort and add to our understanding, particularly that lower white-matter integrity mediates the effect (i.e., is the pathway of effect) of obesity/waist-hip ratio on cognition.
I’d like to note a COI with the Whitehall paper as some of its authors are colleagues.
View all comments by Gill LivingstonMayo Clinic College of Medicine
We would all like to believe that we could control our own health by eating healthy. Dietary epidemiological studies such as this one of Jensen et al. sought to validate the protective effects of a healthy diet on cognition and its presumed neurophysiological correlates. The authors found that those who reported following quality diets in midlife and who had lower waist-to-hip ratios had better functional and structural MRI measurements as well as slightly better cognitive performance. While these results are gratifying, the lack of attention to the socioeconomic status of participants lessens my confidence for making causal claims about the benefits of diet on later life cognition. Dietary choices are strongly influenced by socioeconomic circumstances, which themselves are unquestionably strongly related to all health outcomes including cognitive health. If the authors had been able to control for socioeconomics, which they do not mention, would the associations of diet be attenuated?
The straightforward observations of Dhana et al. showing that plasma NfL concentrations were associated with a cardiovascular health score shows that the known effects of vascular disease on brain health can be proxied by the plasma biomarkers. This is reinforced by an article that just appeared in Neurology, which also showed in a large multidisciplinary cohort that NfL and AD specific biomarkers associated with brain volume loss and white-matter hyperintensity burden (Sanchez et al., 2025).
One of the valuable insights that has emerged in the past few years about the relationship between cerebrovascular disease and brain health is that it is not just overt strokes that matter. Cerebrovascular disease also exerts its pathologic effects via gradual, insidious, covert means, that are proxied by brain volume loss and nonspecific plasma biomarkers, such as NfL.
References:
Sanchez E, Coughlan GT, Wilkinson T, Ramirez J, Mirza SS, Baril AA, Dilliott AA, Frank A, Lang AE, Hassan A, Pollock BG, Scott CJ, Marras C, Fischer CE, Seitz D, Andriuta D, Dowlatshahi D, Grimes DA, Tang-Wai DF, Sahlas DJ, Rogaeva EA, Finger E, Robinson JF, Tan K, Binns MA, Tartaglia MC, Borrie MJ, Strong MJ, Ozzoude M, Nanayakkara ND, Goncalves RA, Bartha R, Hegele RA, Farhan SM, Black SE, Kumar S, Symons SP, Haddad SM, Pasternak SH, Arnott SR, Rajji TK, Steeves T, Swardfager W, Ashton NJ, Kvartsberg H, Zetterberg H, Munoz DP, Masellis M, ONDRI Investigators. Association of Plasma Biomarkers With Longitudinal Atrophy and Microvascular Burden on MRI Across Neurodegenerative and Cerebrovascular Diseases. Neurology. 2025 Apr 8;104(7):e213438. Epub 2025 Mar 10 PubMed.
View all comments by David KnopmanUniversity Hospital Leipzig and Max Planck Institute for Human Cognitive and Brain Sciences
We thank Drs. Livingston and Knopman for their comments on our study and their evaluation. We agree with Dr Knopman that socioeconomic status is strongly related to dietary quality. Therefore, we’d like to clarify that we did correct all analyses for education status. Thus, in the "Statistical Analysis" section of the paper we state "All analyses were adjusted for age, sex, years of education, MRI scanner model, physical activity, and the Montreal Cognitive Assessment score, all of which were measured at the time of MRI scan. Resting-state fMRI analyses were corrected for head motion and voxelwise gray matter density. AHEI-2010 analyses were adjusted for total energy intake (kcal/d) (eMethods 7 in Supplement 1).
References:
Jensen DE, Ebmeier KP, Akbaraly T, Jansen MG, Singh-Manoux A, Kivimäki M, Zsoldos E, Klein-Flügge MC, Suri S. Association of Diet and Waist-to-Hip Ratio With Brain Connectivity and Memory in Aging. JAMA Netw Open. 2025 Mar 3;8(3):e250171. PubMed.
View all comments by Daria JensenMake a Comment
To make a comment you must login or register.