Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Observational Study
. 2023 Jan 6:12:e81869.
doi: 10.7554/eLife.81869.

Multimodal brain age estimates relate to Alzheimer disease biomarkers and cognition in early stages: a cross-sectional observational study

Peter R Millar  1 , Brian A Gordon  2 , Patrick H Luckett  3 , Tammie L S Benzinger  2   3 , Carlos Cruchaga  4 , Anne M Fagan  1 , Jason J Hassenstab  1 , Richard J Perrin  1   5 , Suzanne E Schindler  1 , Ricardo F Allegri  6 , Gregory S Day  7 , Martin R Farlow  8 , Hiroshi Mori  9 , Georg Nübling  10   11 ; Dominantly Inherited Alzheimer Network; Randall J Bateman  1 , John C Morris  1 , Beau M Ances  1   2
Collaborators, Affiliations
Observational Study

Multimodal brain age estimates relate to Alzheimer disease biomarkers and cognition in early stages: a cross-sectional observational study

Peter R Millar et al. Elife. .

Abstract

Background: Estimates of 'brain-predicted age' quantify apparent brain age compared to normative trajectories of neuroimaging features. The brain age gap (BAG) between predicted and chronological age is elevated in symptomatic Alzheimer disease (AD) but has not been well explored in presymptomatic AD. Prior studies have typically modeled BAG with structural MRI, but more recently other modalities, including functional connectivity (FC) and multimodal MRI, have been explored.

Methods: We trained three models to predict age from FC, structural (S), or multimodal MRI (S+FC) in 390 amyloid-negative cognitively normal (CN/A-) participants (18-89 years old). In independent samples of 144 CN/A-, 154 CN/A+, and 154 cognitively impaired (CI; CDR > 0) participants, we tested relationships between BAG and AD biomarkers of amyloid and tau, as well as a global cognitive composite.

Results: All models predicted age in the control training set, with the multimodal model outperforming the unimodal models. All three BAG estimates were significantly elevated in CI compared to controls. FC-BAG was significantly reduced in CN/A+ participants compared to CN/A-. In CI participants only, elevated S-BAG and S+FC BAG were associated with more advanced AD pathology and lower cognitive performance.

Conclusions: Both FC-BAG and S-BAG are elevated in CI participants. However, FC and structural MRI also capture complementary signals. Specifically, FC-BAG may capture a unique biphasic response to presymptomatic AD pathology, while S-BAG may capture pathological progression and cognitive decline in the symptomatic stage. A multimodal age-prediction model improves sensitivity to healthy age differences.

Funding: This work was supported by the National Institutes of Health (P01-AG026276, P01- AG03991, P30-AG066444, 5-R01-AG052550, 5-R01-AG057680, 1-R01-AG067505, 1S10RR022984-01A1, and U19-AG032438), the BrightFocus Foundation (A2022014F), and the Alzheimer's Association (SG-20-690363-DIAN).

Keywords: Alzheimer disease; brain aging; human; machine learning; medicine; neuroscience; resting-state functional connectivity; structural MRI.

Plain language summary

The brains of people with advanced Alzheimer’s disease often look older than expected based on the patients’ actual age. This ‘brain age gap’ (how old a brain appears compared to the person’s chronological age) can be calculated thanks to machine learning algorithms which analyse images of the organ to detect changes related to aging. Traditionally, these models have relied on images of the brain structure, such as the size and thickness of various brain areas; more recent models have started to use activity data, such as how different brain regions work together to form functional networks. While the brain age gap is a useful measure for researchers who investigate aging and disease, it is not yet helpful for clinicians. For example, it is unclear whether the machine learning algorithm could detect changes in the brains of individuals in the initial stages of Alzheimer’s disease, before they start to manifest cognitive symptoms. Millar et al. explored this question by testing whether models which incorporate structural and activity data could be more sensitive to these early changes. Three machine learning algorithms (relying on either structural data, activity data, or combination of both) were used to predict the brain ages of participants with no sign of disease; with biological markers of Alzheimer’s disease but preserved cognitive functions; and with marked cognitive symptoms of the condition. Overall, the combined model was slightly better at predicting the brain age of healthy volunteers, and all three models indicated that patients with dementia had a brain which looked older than normal. For this group, the model based on structural data was also able to make predictions which reflected the severity of cognitive decline. Crucially, the algorithm which used activity data predicted that, in individuals with biological markers of Alzheimer’s disease but no cognitive impairment, the brain looked in fact younger than chronological age. Exactly why this is the case remains unclear, but this signal may be driven by neural processes which unfold in the early stages of the disease. While more research is needed, the work by Millar et al. helps to explore how various types of machine learning models could one day be used to assess and predict brain health.

PubMed Disclaimer

Conflict of interest statement

PM, BG, PL, RP, RA, HM, GN, JM, BA No competing interests declared, TB received doses (AV45, AV1451) and partial support for PET scanning through an investigator-initiated research grant awarded to Washington University from Avid Radiopharmaceuticals (a wholly-owned subsidiary of Eli Lilly and Company). The author received consulting fees from Eisai, Siemens, and received payment for Biogen speaker's bureau. Tammie Benzinger acts as site investigator in clinical trials sponsored by Avid Radiopharmaceuticals, Eli Lilly and Company, Biogen, Eisai, Jaansen and Roche. The author has no other competing interests to declare, CC has received research support from Biogen, EISAI, Alector and Parabon. Carlos Cruchaga is a member of the advisory board of Vivid Genetics, Circular Genomics and Alector. The author has no other competing interests to declare, AF has received consulting fees from DiamiR and Siemens Healthcare Diagnostics Inc and has received consulting fees for participation on Scientific advisory boards for Roche Diagnostics, Genentech and Diadem. The author has received travel support for in-person attendance at ABC-DS Meeting/Retreat and travel support/honorarium for in-person attendance at Scientific Advisory Board meeting for South Texas Alzheimer's Disease Research Center (ADRC). The author has no other competing interests to declare, JH has received consulting fees from Roche and Parabon Nanolabs. The author has no other competing interests to declare, SS received personal honoraria for presenting lectures from the University of Wisconsin, St. Luke's Hospital, Houston Methodist Medical Center, personal Honoraria for serving on the Alzheimer Disease Center Clinical Task Force from University of Washington and personal honoraria for serving on the National Centralized Repository for Alzheimer's Disease biospecimen review committee from University of Indiana. The author received travel support from National Institute on Aging grant R01AG070941, and is a board member of the Greater Missouri Alzheimer's Association. The author received plasma Ab42/Ab40 data provided by C2N Diagnostics at no cost. No payments/research funding was provided by C2N Diagnostics. No gifts/financial incentives of any kind have been provided to Dr. Schindler by C2N Diagnostics. The author has no other competing interests to declare, GD received fees for consulting and for acting as Dementia Topic Editor from DynaMed (EBSCO Health) and received fees for consulting, grant writing / implementation Parabon Nanonlabs. The author received payment for CME Content development from PeerView Media and Continuing Education Inc, payment for educational content development and focus group participation from Eli Lilly Co, and payment for continuum manuscript authorship from the American Academy of Neurology. The author received payment for expert testimony in the case of Wernicke encephalopathy from Barrow Law. Gregory S Day acts as Clinical Director for Anti-NMDA Receptor Encephalitis Foundation, Inc. The author has stock holdings at ANI Pharmaceuticals, Inc and stock options at Parabon Nanolabs. The author has no other competing interests to declare, MF received grants from AbbVie, Eisai, Novartis, ADCS Posiphen, Genentech and Suven Life Sciences (no grant numbers available). The author has received consulting fees from Artery Therapeutics, Avanir, Biogen, Cyclo Therapeutics, Green Valley, Lexeo, McClena, Nervive, Oligomerix, Pinteon, Prothena, Vaxinity, Athira, AZTherapies, Cognition Therapeutics, Gemvax, Ionis, Longeveron, Merck, Neurotrope Biosciences, Otsuka, Proclara and SToP-AD. The author has no other competing interests to declare, received funding and non-financial support for the DIAN-TU-001 trial from Avid Radiopharmaceuticals, and funding for the DIAN-TU-001 trial from Janssen, Hoffman La-Roche/Genentech, Eli Lilly & Co., Eisai, Biogen, AbbVie and Bristol Meyer Squibb. The author has equity ownership interest in C2N Diagnostics and receives royalty income based on technology (stable isotope labeling kinetics and blood plasma assay) licensed by Washington University to C2N Diagnostics. The author received International Conference Lecture Honoraria from Korean Dementia Association and Conference Lecture Honoraria from Weill Cornell Medical College. The author received support for travel expenses from Alzheimer's Association Roundtable and Duke Margolis Alzheimer's Roundtable. The author participates on an unpaid Advisory Board for Roche Gantenerumab Steering Committee and Biogen - Combination Therapy for Alzheimer's Disease, and participates on an unpaid Scientific Advisory Board for UK Dementia Research Institute at University College London and Stanford University, Next Generation Translational Proteomics for Alzheimer's and Related Dementias. The author receives an income from C2N Diagnostics for serving on the scientific advisory board. The author has received equipment and materials from Avid Radiopharmaceuticals, Eli Lilly & Co, Hoffman La-Roche, Eisai and Janssen. Unrelated to this article, Randall Bateman serves as principal investigator of the DIAN-TU, which is supported by the Alzheimer's Association, GHR Foundation, an anonymous organization and the DIAN-TU Pharma Consortium (Active: Eli Lilly and Company/Avid Radiopharmaceuticals, F. Hoffman-La Roche/Genentech, Biogen, Eisai, and Janssen. Previous: Abbvie, Amgen, AstraZeneca, Forum, Mithridion, Novartis, Pfizer, Sanofi, and United Neuroscience). In addition, in-kind support has been received from CogState and Signant Health. Unrelated to this article Randall Bateman has submitted the US nonprovisional patent application "Methods for Measuring the Metabolism of CNS Derived Biomolecules In Vivo" and provisional patent application "Plasma Based Methods for Detecting CNS Amyloid Deposition". The author has no other competing interests to declare, RB has received consulting fees from Barcelona Brain Research Center BBRC and Native Alzheimer Disease-Related Resource Center in Minority Aging Research, Ext Adv Board. The author has received payment or honoraria for lectures from Montefiore Grand Rounds, NY and Tetra-Inst ADRC seminar series, Grand Rds, NY. The author has participated on the Research Strategy Council for the Cure Alzheimer's Fund, the Diverse VCID Observational Study Monitoring Board and the LEADS Advisory Board, Indiana University. The author has no other competing interests to declare

Figures

Figure 1.
Figure 1.. Performance of the brain age models in the training (left column) and test sets (right column) for each modality: functional connectivity (FC; A and B), structural MRI (S; C and D) and multimodal models (S+FC; E and F).
Age predicted by each model (y axis) is plotted against true age (x axis). Colored lines and shaded areas represent regression lines and 95% confidence regions. Dashed black lines represent perfect prediction. Model performance is evaluated by Pearson’s r, proportion of variance explained (R2), mean absolute error (MAE), and root-mean-square error (RMSE).
Figure 1—figure supplement 1.
Figure 1—figure supplement 1.. Flow chart of participant inclusion, exclusion, and group assignments.
Figure 1—figure supplement 2.
Figure 1—figure supplement 2.. Tuning curves of σ hyperparameter in training for structural (A) and functional connectivity (B) Gaussian process regression (GPR) models.
Figure 1—figure supplement 3.
Figure 1—figure supplement 3.. Correlation between S-brain age gap (BAG; x axis) and functional connectivity (FC)-BAG (y axis) estimates in the training and validation sets (A) and analysis sets (B).
Both BAG estimates are residualized for age. Dotted black lines represent no difference between predicted and chronological age for each model. Colored lines and shaded areas represent group-specific regression lines and 95% confidence regions. Dashed black lines represent main effect regression lines across all groups.
Figure 1—figure supplement 4.
Figure 1—figure supplement 4.. Violin plot of R2 performance estimates from 1000 bootstrapped samples in which a stacked brain age model combined the fully-trained structural MRI model (R2S) with a reshuffled functional connectivity (FC) model (i.e. FC training features were randomly reassigned in each bootstrap sample).
Most bootstrapped stacked models perform about as well or worse than the unimodal structural MRI model (R2S, black dashed line). The fully-trained stacked multimodal model (R2S+FC, red solid line) outperforms all bootstrapped models, suggesting that the modest increase in model performance observed in the multimodal model over the unimodal structural model is due to meaningful age-related FC signal, rather than capitalizing on noise in a larger feature set.
Figure 2.
Figure 2.. Group differences in functional connectivity (FC; A and B), structural (S; C and D), and multimodal (S+FC; E and F) brain age in the analysis sets.
Comparisons are presented between cognitively normal (Clinical Dementia Rating [CDR] = 0) biomarker-negative controls (CN/A−; blue) vs. CN/A+ (green) vs. cognitively impaired participants (CI, red). Scatterplots (A, C, and E) show predicted vs. true age for each group. Colored lines and shaded areas represent group-specific regression lines and 95% confidence regions. Dashed black lines represent perfect prediction. Violin plots (B, D, and F) show residual FC-brain age gap (BAG; controlling for true age) in each group. p values are reported from pairwise independent-samples t tests.
Figure 2—figure supplement 1.
Figure 2—figure supplement 1.. Quantile-quantile plots of brain age gap, controlling for age, in each of the analysis sets (cognitively normal, amyloid negative [CN/A−]; CN/A+; and cognitively impaired [CI]) for functional connectivity [FC; A], structural [S; B] and multimodal [S+FC; C] models.
Figure 3.
Figure 3.. Continuous relationships between amyloid biomarkers and functional connectivity (FC-brain age gap [BAG]; A and B), structural (S-BAG; C and D), and multimodal (S+FC BAG; E and F) BAG in the analysis sets.
Scatterplots show amyloid PET (A, C, and E) and CSF AB42/40 (B, D, and F) as a function of residual BAG (controlling for true age) in each group. Colored lines and shaded areas represent group-specific regression lines and 95% confidence regions. Dashed black lines represent main effect regression lines across all groups.
Figure 4.
Figure 4.. Continuous relationships between tau biomarkers and functional connectivity (FC-brain age gap [BAG]; A and B), structural (S-BAG; C and D), and multimodal (S+FC BAG; E and F) BAG in the analysis sets.
Scatterplots show Tau PET summary (A, C, and E) and log-transformed CSF pTau/Aβ40 (B, D, and F) as a function of residual BAG (controlling for true age) in each group. Colored lines and shaded areas represent group-specific regression lines and 95% confidence regions. Dashed black lines represent main effect regression lines across all groups.
Figure 5.
Figure 5.. Continuous relationships between global cognition and functional connectivity (FC-brain age gap [BAG]; A), structural (S-BAG; B), and multimodal (S+FC BAG; C) in the analysis sets.
Scatterplots show global cognition as a function of residual BAG (controlling for true age) in each group. Colored lines and shaded areas represent group-specific regression lines and 95% confidence regions. Dashed black lines represent main effect regression lines across all groups.
See this image and copyright information in PMC

Update of

  • doi: 10.1101/2022.08.25.505251

References

    1. Armitage SG. An analysis of certain psychological tests used for the evaluation of brain injury. Psychological Monographs. 1946;60:i1–i48. doi: 10.1037/h0093567. - DOI
    1. Aschenbrenner AJ, Gordon BA, Benzinger TLS, Morris JC, Hassenstab JJ. Influence of tau PET, amyloid PET, and hippocampal volume on cognition in Alzheimer disease. Neurology. 2018;91:e859–e866. doi: 10.1212/WNL.0000000000006075. - DOI - PMC - PubMed
    1. Bashyam VM, Erus G, Doshi J, Habes M, Nasrallah IM, Truelove-Hill M, Srinivasan D, Mamourian L, Pomponio R, Fan Y, Launer LJ, Masters CL, Maruff P, Zhuo C, Völzke H, Johnson SC, Fripp J, Koutsouleris N, Satterthwaite TD, Wolf D, Gur RE, Gur RC, Morris J, Albert MS, Grabe HJ, Resnick S, Bryan RN, Wolk DA, Shou H, Davatzikos C. Mri signatures of brain age and disease over the lifespan based on a deep brain network and 14 468 individuals worldwide. Brain. 2020;143:2312–2324. doi: 10.1093/brain/awaa160. - DOI - PMC - PubMed
    1. Bocancea DI, van Loenhoud AC, Groot C, Barkhof F, van der Flier WM, Ossenkoppele R. Measuring resilience and resistance in aging and Alzheimer disease using residual methods: a systematic review and meta-analysis. Neurology. 2021;97:474–488. doi: 10.1212/WNL.0000000000012499. - DOI - PMC - PubMed
    1. Brier MR, Thomas JB, Ances BM. Network dysfunction in alzheimer’s disease: refining the disconnection hypothesis. Brain Connectivity. 2014a;4:299–311. doi: 10.1089/brain.2014.0236. - DOI - PMC - PubMed

Publication types