Pith. sign in

REVIEW 3 major objections 5 minor 51 references

Region-wise stacking ensembles for estimating brain-age using MRI

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A two-level stacking ensemble, pairing per-region voxel-based age predictors with a second-stage fusion model, estimates brain age from structural MRI more accurately than does the conventional regional-mean gray-matter baseline.

desk verdict Region-wise stacking is a genuine contribution and the core accuracy claim holds up outside the OOS setups, but the headline number conflates stacking with target-domain adaptation. read the letter →

arxiv 2501.10153 v1 pith:STOLTLKY submitted 2025-01-17 cs.LG cs.AI

Georgios Antonopoulos , Shammi More , Simon B. Eickhoff , Federico Raimondo , Kaustubh R. Patil This is my paper · ORCID
classification cs.LGcs.AI
keywords brain-agepredictionstackingensemblemagneticresonanceimaginggraymattervolumevoxel-basedmorphometrydatafusionleave-one-site-outvalidationprivacy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes a two-level stacking ensemble for estimating a person's age from structural MRI. Instead of averaging gray-matter volume across each brain region, the first level trains a separate voxel-based predictor of age for each of 873 regions, and the second level learns how to combine those 873 regional age predictions into one estimate. Across four datasets and multiple fusion setups, the stacking ensembles beat the classical regional-mean baseline; the best setup reaches a mean absolute error of 4.75 years versus 5.68 years for the baseline. The paper also reports that regional age predictions carry a stronger and more consistent aging signal than regional volume averages, and that they make it harder to identify which dataset a scan came from, which it interprets as a privacy benefit. The intended significance is a more accurate, interpretable, and privacy-preserving route to brain-age biomarkers for studying healthy and pathological aging.

What carries the argument

The central object is the two-level stacking ensemble. At level 0, each of 873 predefined brain regions gets its own elastic-net regression model, trained on the voxel-wise gray matter volume values inside that region to predict age; this replaces the uniform averaging of voxels used in the baseline. At level 1, a second elastic-net model takes the 873 regional age predictions as input features and produces the final age estimate. The mechanism works because the first-level models weight each voxel by how well it tracks aging instead of treating all voxels equally, and the second-level model learns how much to trust each region rather than applying a predefined weight. The paper also relies on an out-of-sample scheme for the best setup: L0 predictions for a target site are obtained by three-fold cross-validation on that site's own data, with L1 trained on independent data from other sites.

What would settle it

Train the first-level regional models strictly on the three held-in sites, with no target-site data, and compare the same mean absolute error against the regional-mean gray-matter baseline; if the stacking gain shrinks or reverses, the claimed superiority over averaging would be refuted.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that a two-level stacking ensemble improves brain-age prediction compared with the standard practice of averaging voxel-wise gray matter volume within brain regions. At the first level (L0), an elastic-net model is trained per region on the voxels of that region to predict age; at the second level (L1), another elastic-net model learns to combine the 873 regional age predictions. The paper systematically compares eight data-fusion scenarios across four datasets with leave-one-site-out validation. The best-performing scenario obtains out-of-sample L0 predictions at the application site and trains L1 on independent site-specific data, achieving a mean absolute error of 4.75 years versus 5.68 years for the regional-mean baseline; all stacking setups outperform the gray-matter-volume baselines on mean absolute error, $\mathrm{R}^2$, and correlation. The paper further claims that L0 predictions correlate more strongly and more stably with chronological age than regional mean gray matter volume (mean absolute correlation 0.6 versus 0.32), and that a dataset-of-origin classifier achieves lower balanced accuracy on L0 predictions (0.63) than on gray-matter-volume features (0.87), indicating improved privacy.

Load-bearing premise

The claim that stacking beats regional averaging stands on the assumption that the accuracy gain comes from the stacking design itself, not from the best-performing setup's use of the target site's own data, with known ages, to train the first-level models.

Editorial extensions

If this is right

  • A clinic or consortium can estimate brain age without pooling raw MRI scans, because only the 873 regional age predictions need to be shared.
  • Regional age predictions reveal aging signal in subcortical structures that regional volume averaging dilutes, giving a more detailed map of where the brain ages.
  • Accuracy improves systematically as more training datasets are added, so the approach becomes more valuable as multi-site data accumulate.
  • The two-level design leaves room for other base learners, parcellation schemes, and bias-correction steps, which the paper identifies as future refinements.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the best out-of-sample setup trains L0 models on the target site's own data with known ages, part of the reported gain may come from target-site adaptation rather than from the stacking architecture; a stricter evaluation would train L0 exclusively on the three held-in sites.
  • The same two-level design could be applied to other voxel-wise biomarkers or to longitudinal settings, and age-bias correction could be integrated into the L1 model rather than applied post hoc.
  • A direct ablation comparing the stacking ensemble with a single whole-brain voxel-based elastic net under the same regularization and folds would isolate whether the regional decomposition and the second-level fusion carry the gain.
  • The privacy conclusion rests on a proxy task; a direct re-identification test would strengthen or weaken the claim that L0 predictions protect subject identity.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes a two-level stacking ensemble (SE) for MRI-based brain-age prediction: first-level (L0) models are trained on voxel-wise gray matter volume within each of 873 brain regions, and a second-level (L1) GLMnet model fuses the regional age predictions. Eight data-fusion scenarios are compared under leave-one-site-out (LOSO) validation across four datasets, with baselines using regional mean GMV. The authors report that SE outperforms the regional-mean baseline, with the abstract's headline result of MAE=4.75 for an out-of-sample L0 setup versus MAE=5.68 for the corresponding GMV baseline. They also report stronger age correlations of L0 predictions than of raw GMV, and lower dataset-identification accuracy from L0 predictions, which they interpret as improved biological insight and privacy.

Significance. If the central accuracy claim holds with the appropriate comparison, the paper offers a practically useful and privacy-aware framework for regional brain-age modeling, with an open code repository and a systematic exploration of data-fusion choices. The non-OOS comparisons (e.g., PredL0 P L1 P MAE=5.06 vs GMV P L1 P MAE=6.22) do show an advantage for the stacking architecture over regional-mean GMV, so the qualitative conclusion is defensible. The paper is less convincing on its headline framing, statistical substantiation, and biological-interpretation claims.

major comments (3)
  1. [Data split and training setups; Results; Supplementary Table 2] The abstract's headline comparison (MAE=4.75 vs 5.68) conflates the stacking ensemble with target-domain adaptation. In OOSPred S L1 P, the L0 regional models are trained with 3-fold CV on the test site's own data, including its age labels, while the GMV P L1 P ext baseline only uses test-site folds at L1 and a fixed averaging operation at L0. The paper's statement that LOSO 'mimics the scenario when the models are applied to data from a new scanner not available during training' is therefore invalid for the setup that produces the headline number. The non-OOS setups (e.g., PredL0 S L1 P, MAE=5.12, vs GMV P L1 P, MAE=6.22) preserve the qualitative SE advantage, but the abstract and Results need to be reframed so the 4.75 result is attributed to SE combined with test-site L0 fitting, not to SE alone.
  2. [Results, Biological insights; Figures 4-6] The claim that L0 predictions 'better capture the aging process' is circular: the L0 models are trained to predict age, so their outputs are mechanically correlated with age. The comparison of L0 prediction-age correlations with raw GMV-age correlations therefore does not provide independent evidence of improved biological validity. The paper should either remove or substantially soften this interpretation, or support it with held-out validation, external age-related outcomes, or a non-circular benchmark such as predicting age from GMV in one region using a model trained on the same region's GMV (rather than on age itself).
  3. [Results, Performance; Figure 2; Supplementary Table 2] No uncertainty quantification is provided for the LOSO performance differences. With only four test sites, each setup yields a single aggregate MAE; the reported differences (e.g., 4.75 vs 5.12 vs 5.68) could be within site-to-site noise. The paper should report per-fold MAE values, confidence intervals, or a paired statistical test across the four LOSO folds before claiming a reliable ranking of setups. This is load-bearing for the central claim that SE outperforms the GMV baselines.
minor comments (5)
  1. [Results, Biological insights] The expression 'p<<0.01e13' is not a valid p-value; please correct the notation and report a test statistic or a standard p-value.
  2. [Results, Biological insights vs Stability across datasets] The two sets of mean correlations (r_mean=0.6 vs 0.32 in the Biological insights section and mean=0.86 vs 0.81 in the Stability across datasets section) appear inconsistent; the text should clarify which quantity is averaged and how the two analyses differ.
  3. [Table 1] The duplicate rows marked with an asterisk and the notation 'Mean()' in the L0 operation column are confusing; please clarify whether the GMV setups use the same training splits and data as the corresponding OOS setups.
  4. [Methods, SE model training] The description of GLMnet's 'adaptive resampling' for tuning lambda and alpha is brief; for reproducibility, please specify the resampling scheme (number of folds, repetitions) or cite the package defaults explicitly.
  5. [Discussion] The sentence 'This way of obtaining OOS L0 predictions effectively models the idiosyncrasies of the specific dataset' is vague and risks overinterpreting a target-domain adaptation effect; please revise to state precisely what is being modeled and how.

Circularity Check

1 steps flagged · score 6.0 of 10

The MAE comparisons are empirical and non-circular, but the 'biological insights' claim is circular: L0 models are trained to predict age, so strong L0-prediction/age correlations merely reflect the training objective, not independent evidence of aging signal.

  1. fitted input called prediction [Methods — Biological insights; Results — Biological insights (Figs. 4–6); Discussion]
    "To this end, we calculated Pearson’s correlation of real age with L0 predictions and mean regional GMV across subjects. These values were compared to identify the more robust approach providing biological insights. ... The results showed that regional predictions of the L0 models exhibit a stronger alignment to age (Figure 4)."

    L0 regional models are elastic-net regressions trained per region, via 3-fold CV, to predict chronological age from voxel-level GMV. Their outputs are therefore constructed to track age. Reporting corr(L0 prediction, age) as 'biological insight' is an evaluation of the supervised objective, not an independent discovery about aging; the comparison against unsupervised mean-GMV correlations compounds the issue because a supervised feature extractor is expected to align with its training target. No external validation is supplied, so the claim that SE 'better captures the aging process' reduces to the statement that the L0 models were fitted to age.

full rationale

The core performance comparison is not circular: SE and GMV baselines are evaluated in site-separated LOSO folds, and non-OOS SE setups (e.g., PredL0^P L1^P MAE 5.06, PredL0^S L1^P MAE 5.12) beat GMV baselines (MAE 6.22, 6.68), so the stacking architecture has independent empirical content. The circularity is localized to the biological-insight interpretation: L0 is fitted to predict age, so high L0-prediction/age correlation is not evidence that the model 'better captures the aging process' beyond its training objective. Separately, the headline MAE 4.75 is obtained from the OOSPred setup, in which L0 is trained on the test site's own labels via 3-fold CV; this is a target-site adaptation effect and does not mimic a new-scanner scenario, a validity limitation rather than a formal circularity.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

No new entities are introduced. The main free parameters are the regularization parameters of the elastic net, which are tuned automatically. The central assumptions are the choice of parcellation, the validity of LOSO as a generalization estimate, and the use of GMV as the input representation.

free parameters (1)
  • GLMnet alpha and lambda = tuned internally via adaptive resampling
    These are model hyperparameters selected by the glmnet package during training. They are not scientific parameters fitted to external data, but they are still free in the sense that the final model depends on the tuning procedure.
assumptions (3)
  • domain assumption The 873-region parcellation retains enough anatomical specificity for age prediction
    The choice of atlas granularity is a modeling decision; the paper acknowledges other options could be explored, but the results depend on this specific parcellation.
  • domain assumption Leave-one-site-out evaluation approximates cross-scanner generalization
    The paper states this assumption in the Data split section, but the OOS setups train L0 on the test site's own data, violating a strict interpretation of 'new scanner not available during training'.
  • domain assumption Gray matter volume derived from CAT12 preprocessing is a valid and sufficient representation for brain age estimation
    The entire study uses modulated GMV as the input; no alternative tissue types or preprocessing pipelines are compared.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Region-wise stacking ensembles for estimating brain-age using MRI." pith.science (2026). https://pith.science/paper/STOLTLKY

@misc{pith2026250110153,
  author       = {Pith},
  title        = {Pith review of: Region-wise stacking ensembles for estimating brain-age using MRI},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/STOLTLKY}},
  note         = {Machine review of arXiv:2501.10153}
}
read the original abstract

Predictive modeling using structural magnetic resonance imaging (MRI) data is a prominent approach to study brain-aging. Machine learning algorithms and feature extraction methods have been employed to improve predictions and explore healthy and accelerated aging e.g. neurodegenerative and psychiatric disorders. The high-dimensional MRI data pose challenges to building generalizable and interpretable models as well as for data privacy. Common practices are resampling or averaging voxels within predefined parcels, which reduces anatomical specificity and biological interpretability as voxels within a region may differently relate to aging. Effectively, naive fusion by averaging can result in information loss and reduced accuracy. We present a conceptually novel two-level stacking ensemble (SE) approach. The first level comprises regional models for predicting individuals' age based on voxel-wise information, fused by a second-level model yielding final predictions. Eight data fusion scenarios were explored using as input Gray matter volume (GMV) estimates from four datasets covering the adult lifespan. Performance, measured using mean absolute error (MAE), R2, correlation and prediction bias, showed that SE outperformed the region-wise averages. The best performance was obtained when first-level regional predictions were obtained as out-of-sample predictions on the application site with second-level models trained on independent and site-specific data (MAE=4.75 vs baseline regional mean GMV MAE=5.68). Performance improved as more datasets were used for training. First-level predictions showed improved and more robust aging signal providing new biological insights and enhanced data privacy. Overall, the SE improves accuracy compared to the baseline while preserving or enhancing data privacy.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 26 canonical work pages

  1. [2]

    Brain-age prediction: A systematic comparison of machine learning workflows,

    S. More, G. Antonopoulos, F. Hoffstaedter, J. Caspers, S. B. Eickhoff, and K. R. Patil, “Brain-age prediction: A systematic comparison of machine learning workflows,” NeuroImage , vol. 270, p. 119947, Apr. 2023, doi: 10.1016/j.neuroimage.2023.119947

    arXiv 2023

  2. [3]

    Age-related gray matter volume changes in the brain during non-elderly adulthood,

    D. Terribilli et al. , “Age-related gray matter volume changes in the brain during non-elderly adulthood,” Neurobiol. Aging , vol. 32, no. 2–6, pp. 354–368, Feb. 2011, doi: 10.1016/j.neurobiolaging.2009.02.008

    work page doi:10.1016/j.neurobiolaging.2009.02.008 2011

  3. [4]

    Assessing age-related gray matter decline with voxel-based morphometry depends significantly on segmentation and normalization procedures,

    D. V. Callaert, A. Ribbens, F. Maes, S. P. Swinnen, and N. Wenderoth, “Assessing age-related gray matter decline with voxel-based morphometry depends significantly on segmentation and normalization procedures,” Front. Aging Neurosci. , vol. 6, 2014, Accessed: Feb. 13, 2022. [Online]. Available: https://www.frontiersin.org/article/10.3389/fnagi.2014.00124

    work page arXiv 2014

  4. [5]

    Gray Matter Volume and Cognitive Performance During Normal Aging. A Voxel-Based Morphometry Study,

    S. Ramanoël et al. , “Gray Matter Volume and Cognitive Performance During Normal Aging. A Voxel-Based Morphometry Study,” Front. Aging Neurosci. , vol. 10, 2018, Accessed: Feb. 20, 2024. [Online]. Available: https://www.frontiersin.org/articles/10.3389/fnagi.2018.00235

    work page arXiv 2018

  5. [6]

    Prefrontal Gray and White Matter Volumes in 17 Healthy Aging and Alzheimer Disease,

    D. H. Salat, J. A. Kaye, and J. S. Janowsky, “Prefrontal Gray and White Matter Volumes in 17 Healthy Aging and Alzheimer Disease,” Arch. Neurol. , vol. 56, no. 3, pp. 338–344, Mar. 1999, doi: 10.1001/archneur.56.3.338

    work page doi:10.1001/archneur.56.3.338 1999

  6. [7]

    Longitudinal pattern of regional brain volume change differentiates normal aging from MCI,

    I. Driscoll et al. , “Longitudinal pattern of regional brain volume change differentiates normal aging from MCI,” Neurology , vol. 72, no. 22, pp. 1906–1913, Jun. 2009, doi: 10.1212/WNL.0b013e3181a82634

    work page doi:10.1212/wnl.0b013e3181a82634 1906

  7. [8]

    Age-related changes in grey and white matter structure throughout adulthood,

    A. Giorgio et al. , “Age-related changes in grey and white matter structure throughout adulthood,” NeuroImage , vol. 51, no. 3, pp. 943–951, Jul. 2010, doi: 10.1016/j.neuroimage.2010.03.004

    work page doi:10.1016/j.neuroimage.2010.03.004 2010

  8. [9]

    Longitudinal progression of Alzheimer’s-like patterns of atrophy in normal older adults: the SPARE-AD index,

    C. Davatzikos, F. Xu, Y. An, Y. Fan, and S. M. Resnick, “Longitudinal progression of Alzheimer’s-like patterns of atrophy in normal older adults: the SPARE-AD index,” Brain , vol. 132, no. 8, pp. 2026–2035, Aug. 2009, doi: 10.1093/brain/awp091

    work page doi:10.1093/brain/awp091 2026

Show all 51 references
  1. [11]

    Predicting brain-age from multimodal imaging data captures cognitive impairment,

    F. Liem et al. , “Predicting brain-age from multimodal imaging data captures cognitive impairment,” NeuroImage , vol. 148, pp. 179–188, Mar. 2017, doi: 10.1016/j.neuroimage.2016.11.005

    2017 doi

  2. [12]

    Ageing, neurodegeneration and brain rejuvenation,

    T. Wyss-Coray, “Ageing, neurodegeneration and brain rejuvenation,” Nature , vol. 539, no. 7628, pp. 180–186, Nov. 2016, doi: 10.1038/nature20411

    2016 doi

  3. [13]

    Gray Matter Age Prediction as a Biomarker for Risk of Dementia,

    J. Wang et al. , “Gray Matter Age Prediction as a Biomarker for Risk of Dementia,” Proc. Natl. Acad. Sci. , vol. 116, no. 42, pp. 21213–21218, Oct. 2019, doi: 10.1073/pnas.1902376116

    2019 doi

  4. [14]

    Predicting Age Using Neuroimaging: Innovative Brain Ageing Biomarkers,

    J. H. Cole and K. Franke, “Predicting Age Using Neuroimaging: Innovative Brain Ageing Biomarkers,” Trends Neurosci. , vol. 40, no. 12, pp. 681–690, Dec. 2017, doi: 10.1016/j.tins.2017.10.001

    2017 doi

  5. [15]

    A voxel-based morphometric study of ageing in 465 normal adult human brains.,

    C. D. Good, I. S. Johnsrude, J. Ashburner, R. N. Henson, K. J. Friston, and R. S. Frackowiak, “A voxel-based morphometric study of ageing in 465 normal adult human brains.,” Neuroimage , vol. 14, no. 1 Pt 1, pp. 21–36, Jul. 2001, doi: 10.1006/nimg.2001.0786

    2001

  6. [16]

    Global and local gray matter loss in mild cognitive impairment and Alzheimer’s disease,

    G. B. Karas et al. , “Global and local gray matter loss in mild cognitive impairment and Alzheimer’s disease,” NeuroImage , vol. 23, no. 2, pp. 708–716, Oct. 2004, doi: 10.1016/j.neuroimage.2004.07.006

    2004 doi

  7. [17]

    What is normal in normal aging? Effects of aging, amyloid and Alzheimer’s disease on the cerebral cortex and the hippocampus,

    A. M. Fjell, L. McEvoy, D. Holland, A. M. Dale, and K. B. Walhovd, “What is normal in normal aging? Effects of aging, amyloid and Alzheimer’s disease on the cerebral cortex and the hippocampus,” Prog. Neurobiol. , vol. 117, pp. 20–40, Jun. 2014, doi: 10.1016/j.pneurobio.2014.02.004

    2014 doi

  8. [18]

    Pseudonymisation of neuroimages and data protection: Increasing access to data while retaining scientific utility ,

    D. Eke et al. , “Pseudonymisation of neuroimages and data protection: Increasing access to data while retaining scientific utility ,” Neuroimage Rep. , vol. 1, no. 4, p. 100053, Dec. 2021, doi: 10.1016/j.ynirp.2021.100053

    2021

  9. [19]

    Estimating the age of healthy subjects from T1-weighted MRI scans using kernel methods: Exploring the influence of various parameters,

    K. Franke, G. Ziegler, S. Klöppel, and C. Gaser, “Estimating the age of healthy subjects from T1-weighted MRI scans using kernel methods: Exploring the influence of various parameters,” NeuroImage , vol. 50, no. 3, pp. 883–892, Apr. 2010, doi: 10.1016/j.neuroimage.2010.01.005

    2010 doi

  10. [20]

    Predicting brain age with deep learning from raw imaging data results in a reliable and heritable biomarker,

    J. H. Cole et al. , “Predicting brain age with deep learning from raw imaging data results in a reliable and heritable biomarker,” NeuroImage , vol. 163, pp. 115–124, 2017, doi: 10.1016/j.neuroimage.2017.07.059

    2017 doi

  11. [21]

    Learning patterns of the ageing brain in MRI using deep convolutional networks,

    N. K. Dinsdale et al. , “Learning patterns of the ageing brain in MRI using deep convolutional networks,” NeuroImage , vol. 224, p. 117401, Jan. 2021, doi: 10.1016/j.neuroimage.2020.117401. 18

    2021

  12. [22]

    Deep learning-based brain age prediction in normal aging and dementia,

    J. Lee et al. , “Deep learning-based brain age prediction in normal aging and dementia,” Nat. Aging , vol. 2, no. 5, Art. no. 5, May 2022, doi: 10.1038/s43587-022-00219-7

    2022 doi

  13. [24]

    Brain age prediction using deep learning uncovers associated sequence variants,

    B. A. Jonsson et al. , “Brain age prediction using deep learning uncovers associated sequence variants,” Nat. Commun. , vol. 10, no. 1, Art. no. 1, Nov. 2019, doi: 10.1038/s41467-019-13163-9

    2019 doi

  14. [25]

    Brain Age Prediction With Morphological Features Using Deep Neural Networks: Results From Predictive Analytic Competition 2019,

    A. Lombardi, A. Monaco, G. Donvito, N. Amoroso, R. Bellotti, and S. Tangaro, “Brain Age Prediction With Morphological Features Using Deep Neural Networks: Results From Predictive Analytic Competition 2019,” Front. Psychiatry , vol. 11, 2021, Accessed: Jan. 22, 2024. [Online]. ...

    2019

  15. [26]

    Accelerated Brain Aging in Schizophrenia and Beyond: A Neuroanatomical Marker of Psychiatric Disorders,

    N. Koutsouleris et al. , “Accelerated Brain Aging in Schizophrenia and Beyond: A Neuroanatomical Marker of Psychiatric Disorders,” Schizophr. Bull. , vol. 40, no. 5, pp. 1140–1153, Sep. 2014, doi: 10.1093/schbul/sbt142

    2014 doi

  16. [27]

    Local Brain-Age: A U-Net Model,

    S. G. Popescu, B. Glocker, D. J. Sharp, and J. H. Cole, “Local Brain-Age: A U-Net Model,” Front. Aging Neurosci. , vol. 13, 2021, Accessed: Feb. 17, 2023. [Online]. Available: https://www.frontiersin.org/articles/10.3389/fnagi.2021.761954

    2021

  17. [28]

    Premature brain aging in humans exposed to maternal nutrient restriction during early gestation,

    K. Franke, C. Gaser, T. J. Roseboom, M. Schwab, and S. R. de Rooij, “Premature brain aging in humans exposed to maternal nutrient restriction during early gestation,” NeuroImage , vol. 173, pp. 460–471, Jun. 2018, doi: 10.1016/j.neuroimage.2017.10.047

    2018 doi

  18. [29]

    Brain Age in Early Stages of Bipolar Disorders or Schizophrenia,

    T. Hajek et al. , “Brain Age in Early Stages of Bipolar Disorders or Schizophrenia,” Schizophr. Bull. , vol. 45, no. 1, pp. 190–198, Jan. 2019, doi: 10.1093/schbul/sbx172

    2019 doi

  19. [30]

    Machine learning for brain age prediction: Introduction to methods and clinical applications,

    L. Baecker, R. Garcia-Dias, S. Vieira, C. Scarpazza, and A. Mechelli, “Machine learning for brain age prediction: Introduction to methods and clinical applications,” eBioMedicine , vol. 72, Oct. 2021, doi: 10.1016/j.ebiom.2021.103600

    2021

  20. [31]

    Advanced brain ageing in Parkinson’s disease is related to disease duration and individual impairment,

    C. R. Eickhoff et al. , “Advanced brain ageing in Parkinson’s disease is related to disease duration and individual impairment,” Brain Commun. , vol. 3, no. 3, p. fcab191, Jul. 2021, doi: 10.1093/braincomms/fcab191

    2021 doi

  21. [32]

    Imaging-based parcellations of the human brain,

    S. B. Eickhoff, B. T. T. Yeo, and S. Genon, “Imaging-based parcellations of the human brain,” Nat. Rev. Neurosci. , vol. 19, no. 11, Art. no. 11, Nov. 2018, doi: 10.1038/s41583-018-0071-7

    2018 doi

  22. [33]

    Detect and correct bias in multi-site neuroimaging datasets,

    C. Wachinger, A. Rieckmann, and S. Pölsterl, “Detect and correct bias in multi-site neuroimaging datasets,” Med. Image Anal. , vol. 67, p. 101879, Jan. 2021, doi: 10.1016/j.media.2020.101879

    2021

  23. [34]

    A systematic comparison of VBM pipelines and their application to age prediction,

    G. Antonopoulos, S. More, F. Raimondo, S. B. Eickhoff, F. Hoffstaedter, and K. R. Patil, “A systematic comparison of VBM pipelines and their application to age prediction,” NeuroImage , vol. 279, p. 120292, Oct. 2023, doi: 10.1016/j.neuroimage.2023.120292

    2023

  24. [35]

    Ethics and Law in Research on Algorithmic and Data-Driven Technology in Mental Health Care: Scoping Review,

    P. Gooding and T. Kariotis, “Ethics and Law in Research on Algorithmic and Data-Driven Technology in Mental Health Care: Scoping Review,” JMIR Ment. Health , vol. 8, no. 6, p. e24668, Jun. 2021, doi: 10.2196/24668

    2021 doi

  25. [36]

    Identification of Anonymous MRI Research Participants with Face-Recognition Software,

    C. G. Schwarz et al. , “Identification of Anonymous MRI Research Participants with Face-Recognition Software,” N. Engl. J. Med. , vol. 381, no. 17, pp. 1684–1686, Oct. 2019, doi: 10.1056/NEJMc1908881

    2019 doi

  26. [37]

    Functional connectome fingerprinting: identifying individuals using patterns of brain connectivity,

    E. S. Finn et al. , “Functional connectome fingerprinting: identifying individuals using patterns of brain connectivity,” Nat. Neurosci. , vol. 18, no. 11, Art. no. 11, Nov. 2015, doi: 10.1038/nn.4135

    2015 doi

  27. [38]

    Stacked generalization,

    D. H. Wolpert, “Stacked generalization,” Neural Netw. , vol. 5, no. 2, pp. 241–259, Jan. 1992, doi: 10.1016/S0893-6080(05)80023-1. 19

    1992 doi

  28. [39]

    A Stacking Ensemble Learning Framework for Genomic Prediction,

    M. Liang et al. , “A Stacking Ensemble Learning Framework for Genomic Prediction,” Front. Genet. , vol. 12, p. 600040, Mar. 2021, doi: 10.3389/fgene.2021.600040

    2021

  29. [40]

    A stacking ensemble model for predicting the occurrence of carotid atherosclerosis,

    X. Zhang et al. , “A stacking ensemble model for predicting the occurrence of carotid atherosclerosis,” Front. Endocrinol. , vol. 15, Jul. 2024, doi: 10.3389/fendo.2024.1390352

    2024

  30. [41]

    Application of stacking ensemble learning model in quantitative analysis of biomaterial activity,

    H. Cao et al. , “Application of stacking ensemble learning model in quantitative analysis of biomaterial activity,” Microchem. J. , vol. 183, p. 108075, Dec. 2022, doi: 10.1016/j.microc.2022.108075

    2022

  31. [42]

    BrainAGE: Revisited and reframed machine learning workflow,

    P. Kalc, R. Dahnke, F. Hoffstaedter, C. Gaser, and A. D. N. Initiative, “BrainAGE: Revisited and reframed machine learning workflow,” Hum. Brain Mapp. , vol. 45, no. 3, p. e26632, 2024, doi: 10.1002/hbm.26632

    2024 doi

  32. [43]

    The Cambridge Centre for Ageing and Neuroscience (Cam-CAN) data repository: Structural and functional MRI, MEG, and cognitive data from a cross-sectional adult lifespan sample,

    J. R. Taylor et al. , “The Cambridge Centre for Ageing and Neuroscience (Cam-CAN) data repository: Structural and functional MRI, MEG, and cognitive data from a cross-sectional adult lifespan sample,” NeuroImage , vol. 144, pp. 262–269, Jan. 2017, doi: 10.1016/j.neuroimage.2015.09.018

    2017 doi

  33. [44]

    The NKI-Rockland Sample: A Model for Accelerating the Pace of Discovery Science in Psychiatry,

    K. B. Nooner et al. , “The NKI-Rockland Sample: A Model for Accelerating the Pace of Discovery Science in Psychiatry,” Front. Neurosci. , vol. 6, p. 152, 2012, doi: 10.3389/fnins.2012.00152

    2012 arXiv

  34. [45]

    Studying variability in human brain aging in a population-based German cohort—rationale and design of 1000BRAINS,

    S. Caspers et al. , “Studying variability in human brain aging in a population-based German cohort—rationale and design of 1000BRAINS,” Front. Aging Neurosci. , vol. 6, 2014, Accessed: Jan. 22, 2024. [Online]. Available: https://www.frontiersin.org/articles/10.3389/fnagi.2014.00149

    2014

  35. [46]

    Diffeomorphic registration using geodesic shooting and Gauss-Newton optimisation.,

    J. Ashburner and K. J. Friston, “Diffeomorphic registration using geodesic shooting and Gauss-Newton optimisation.,” Neuroimage , vol. 55, no. 3, pp. 954–967, Apr. 2011, doi: 10.1016/j.neuroimage.2010.12.049

    2011 doi

  36. [47]

    Regularization Paths for Generalized Linear Models via Coordinate Descent,

    J. H. Friedman, T. Hastie, and R. Tibshirani, “Regularization Paths for Generalized Linear Models via Coordinate Descent,” J. Stat. Softw. , vol. 33, pp. 1–22, Feb. 2010, doi: 10.18637/jss.v033.i01

    2010 doi

  37. [48]

    Local-Global Parcellation of the Human Cerebral Cortex from Intrinsic Functional Connectivity MRI,

    A. Schaefer et al. , “Local-Global Parcellation of the Human Cerebral Cortex from Intrinsic Functional Connectivity MRI,” Cereb. Cortex N. Y. N 1991 , vol. 28, no. 9, pp. 3095–3114, Sep. 2018, doi: 10.1093/cercor/bhx179

    1991 doi

  38. [49]

    The Human Brainnetome Atlas: A New Brain Atlas Based on Connectional Architecture,

    L. Fan et al. , “The Human Brainnetome Atlas: A New Brain Atlas Based on Connectional Architecture,” Cereb. Cortex , vol. 26, no. 8, pp. 3508–3526, Aug. 2016, doi: 10.1093/cercor/bhw157

    2016 doi

  39. [50]

    The organization of the human cerebellum estimated by intrinsic functional connectivity,

    R. L. Buckner, F. M. Krienen, A. Castellanos, J. C. Diaz, and B. T. T. Yeo, “The organization of the human cerebellum estimated by intrinsic functional connectivity,” J. Neurophysiol. , vol. 106, no. 5, pp. 2322–2345, Nov. 2011, doi: 10.1152/jn.00339.2011

    2011

  40. [51]

    Bias-adjustment in neuroimaging-based brain age frameworks: A robust scheme,

    I. Beheshti, S. Nugent, O. Potvin, and S. Duchesne, “Bias-adjustment in neuroimaging-based brain age frameworks: A robust scheme,” NeuroImage Clin. , vol. 24, p. 102063, Jan. 2019, doi: 10.1016/j.nicl.2019.102063

    2019

  41. [52]

    Mind the gap: Performance metric evaluation in brain-age prediction,

    A.-M. G. de Lange et al. , “Mind the gap: Performance metric evaluation in brain-age prediction,” Hum. Brain Mapp. , vol. 43, no. 10, pp. 3113–3129, 2022, doi: 10.1002/hbm.25837

    2022 doi

  42. [53]

    Early diagnosis of Alzheimer’s disease using combined features from voxel-based morphometry and cortical, subcortical, and hippocampus regions of MRI T1 brain images,

    Y. Gupta, K. H. Lee, K. Y. Choi, J. J. Lee, B. C. Kim, and G. R. Kwon, “Early diagnosis of Alzheimer’s disease using combined features from voxel-based morphometry and cortical, subcortical, and hippocampus regions of MRI T1 brain images,” PLoS ONE , vol. 14, no. 10, p. e02224...

    2019 doi

  43. [54]

    Classification of Alzheimer’s Disease and Mild Cognitive Impairment Based on Cortical and Subcortical Features from MRI T1 Brain Images Utilizing Four Different Types of Datasets,

    S. Toshkhujaev et al. , “Classification of Alzheimer’s Disease and Mild Cognitive Impairment Based on Cortical and Subcortical Features from MRI T1 Brain Images Utilizing Four Different Types of Datasets,” J. Healthc. Eng. , vol. 2020, p. 3743171, Sep. 2020, doi: 20 10.1155/20...

    2020 doi

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.