Pith. sign in

REVIEW 5 major objections 7 minor 76 references

Comprehensively stratifying MCIs into distinct risk subtypes based on brain imaging genetics fusion learning

T0 review · 5 major / 7 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read This paper proposes that MCI patients can be split into low- and high-risk subtypes for Alzheimer's progression using only baseline brain scans and genetic data, and presents BigFirst as the method that does it.

desk verdict Novel architecture for MCI risk subtyping, but the key outcome validation is unquantified and the GWAS is circular — worth a serious referee, not yet a citable finding. read the letter →

arxiv 2508.18058 v1 pith:VXGXTAVE submitted 2025-08-25 q-bio.QM

Muheng Shang , Jin Zhang , Junwei Han , Lei Du This is my paper · ORCID
classification q-bio.QM
keywords MCIsubtypingAlzheimer'sdiseaseriskstratificationimaginggeneticsmultimodalfusionbrainheterogeneityconversionprediction
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

This paper proposes a method, BigFirst, that sorts people with mild cognitive impairment (MCI) into low-risk and high-risk subtypes for progression to Alzheimer's disease, using only baseline brain imaging and genetic data. The core claim is that a model trained on the difference between healthy controls and Alzheimer's patients can serve as a risk ruler for the intermediate MCI population, exposing two biologically distinct subgroups. If the claim holds, clinical trials could pre-screen MCI participants by risk instead of waiting two years to see who converts, reducing the heterogeneity that often weakens trial results. The authors report that the high-risk group has greater brain atrophy, altered amyloid markers, worse memory and executive function, faster longitudinal decline in tau and neurofilament light, and two-year conversion rates largely consistent with the standard 'progressive MCI' label.

What carries the argument

BigFirst is a three-module pipeline. The risk genetic information extraction module uses a selective state-space sequence model to compress roughly 360,000 brain-tissue-related genetic loci into disease-relevant genetic representations. The risk genetic guided pseudo-brain generation module sends those representations through adversarial generators to create per-modality attention masks, called pseudo-brains, that gate real imaging data from three modalities: amyloid PET, glucose PET, and structural MRI. The cross-modality pseudo-brain fusion module projects the gated modalities into a shared space with supervised contrastive learning and uses a Gaussian mixture model to separate healthy fro

What would settle it

If an independent cohort with long follow-up showed that MCI patients labeled low-risk convert to Alzheimer's at the same rate as those labeled high-risk, or that the two predicted subtypes have identical longitudinal atrophy, tau, and neurofilament-light slopes, the risk-axis claim would collapse. A direct test would be to take the model's decision scores on an external set and compare conversion-free survival of the two predicted subtypes; a hazard ratio near one would mean the stratification conveys no risk information.

Watch

Extended reading notes

Core claim

The central discovery is that deploying an imaging-genetics model trained only on healthy controls versus Alzheimer's patients onto an MCI population yields a stable two-cluster partition: 86 low-risk MCI and 331 high-risk MCI. The two clusters are not arbitrary; they differ on many independent measures, including brain atrophy patterns, cerebrospinal fluid amyloid levels, cognitive performance, clinical symptoms, and longitudinal trajectories of plasma tau and neurofilament light. The high-risk group behaves physiologically more like Alzheimer's patients and the low-risk group more like healthy controls, which is exactly the axis the model was trained on. The subtypes also show substantial

Load-bearing premise

The whole scheme assumes that the biological axis separating healthy controls from Alzheimer's patients is the same axis that separates low-risk from high-risk MCI, so training only on healthy versus Alzheimer's brains gives a valid risk ruler for MCI.

Editorial extensions

If this is right

  • MCI risk subtypes can be assigned from a single cross-sectional visit, so trials can enrich for high-risk converters earlier than follow-up-based labels allow.
  • Because the two subtypes differ on structural atrophy but not on amyloid PET, the results suggest brain atrophy is a more sensitive discriminator at the MCI stage and may precede metabolic and amyloid changes.
  • The differing longitudinal trajectories of tau and neurofilament light imply the high-risk subtype is the natural target for early intervention trials, while low-risk individuals may be spared aggressive therapy or used as controls.
  • Genetic loci in CACNA1C and ABCA13, plus lifestyle traits including physical activity, social engagement, and diet, associate with subtype membership and offer candidate risk markers.
  • The subtype labels overlap strongly with two-year stable versus progressive MCI labels but are not identical, so the method predicts conversion at the group level, not certainty for individuals.

Reading between the lines

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

  • A direct extension would be to give each MCI patient a calibrated conversion probability from the model's fused representation, using follow-up conversion as supervision; the paper states its current method only assigns subtype, not risk probability.
  • The healthy-versus-Alzheimer's boundary is what defines 'risk', so an independent external cohort with long follow-up is the natural test of whether the boundary picks out the same high- and low-risk people; internal clustering metrics alone cannot establish biological validity.
  • Because the method learns a single axis from the two endpoint groups, it may collapse genuinely distinct MCI etiologies that share endpoints; applying the same fusion pipeline without those endpoint priors could reveal whether more than two MCI subtypes exist on other axes.
  • The genetic filtering step depends on brain-tissue expression and splicing annotations; alternative tissue-specific or functional annotation schemes could change the genetic representation and therefore the subtypes, so the result should be tested with alternative annotation choices.
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

5 major / 7 minor

Summary. The paper proposes BigFirst, a three-module imaging-genetics fusion framework (RGE, RG2PG, CMPF) trained on HC vs. AD data and then deployed on 417 MCI subjects to learn a risk representation. Applying GMM to the learned multimodal representation yields two MCI subtypes, 86 'low-risk' and 331 'high-risk.' The authors report that the two subtypes differ in brain imaging QTs, CSF Aβ, cognitive domains, clinical symptoms, and covariates at baseline, in longitudinal trajectories of fluid biomarkers and cognition, and in GWAS/PheWAS genetic associations. They further claim high concordance with sMCI/pMCI conversion over two years. The main methodological components are a Mamba-based SNP encoder, a GAN-based pseudo-brain generator, and supervised contrastive fusion with GMM clustering.

Significance. If the central claim holds, BigFirst would be a potentially useful tool for cross-sectional MCI risk stratification using multimodal imaging and genetic data, with practical relevance for clinical trial enrichment. The paper has notable strengths: the architecture explicitly combines whole-genome SNP information with three imaging modalities, the authors report ablation results (Supplementary A.4.1), clustering quality metrics (Table 2), parameter sensitivity (Fig. A1), and external genetic annotation via PheWAS. The downstream biomarker comparisons are extensive. However, the two pillars of the central claim—that the subtypes are truly 'risk' subtypes and that the model captures MCI-to-AD conversion risk—are currently unsupported by quantitative outcome statistics. The low/high-risk labels are not defined operationally, and the only direct outcome validation (sMCI/pMCI concordance) is asserted without a contingency table or statistical test. These are correctable but load-bearing omissions.

major comments (5)
  1. [Section 2.6 / Fig. 7] The central validation of the risk subtypes is the claimed 'substantial concordance' with sMCI/pMCI conversion, but no quantitative result is reported. The text gives no contingency table, odds ratio, risk ratio, p-value, confidence interval, or sensitivity/specificity for two-year conversion. Without these statistics, the statement that BigFirst has 'a strong ability in AD conversion prediction' is unsupported. Please report the full 2x2 table for low/high-risk vs. stable/progressive MCI, an effect size with CI, and a test of association; if appropriate, also adjust for age, sex, and APOE.
  2. [Section 4.2.3 / Section 2.1] The paper never specifies how the two GMM clusters are assigned the labels 'low-risk' and 'high-risk.' Section 4.2.3 says only that GMM is applied to the learned neuroimaging representations. Since all downstream analyses depend on this labeling, the assignment rule must be explicit. For example, are clusters assigned by distance of cluster centroids to HC vs. AD representations, by the sign of a learned discriminant score, or by some monotonic ordering of cluster means? If the assignment is arbitrary or data-driven post hoc, the labels are not biologically grounded and the reported subtype differences could be artifacts of the ordering.
  3. [Section 2.5.1 / Table A3] The GWAS analysis has two serious problems. First, the significance threshold was lowered from 1e-8 to 1e-5 after no SNP passed the initial threshold; with 359,997 SNPs tested, approximately 3.6 hits are expected by chance at p < 1e-5, so the 20 reported loci (Table A3) are not adequately controlled for multiple testing. Second, and more fundamentally, the subtype labels are generated by a model whose RGE module consumes the same SNP data. A GWAS on those labels may therefore rediscover SNPs that the model used to make its representations, rather than independent biological associations. The claims about CACNA1C and ABCA13 need either replication in an independent MCI cohort, a permutation-based null that breaks the SNP-label relationship, or a clear statement that these are model-derived candidates and not genome-wide significant associations.
  4. [Section 2.4 / Fig. 5] The longitudinal findings are presented only as 'trajectories fitted from the mean values' with no statistical testing. Statements such as 'the progression of NfL was slower in the low-risk group' and 'the high-risk group exhibiting faster progression in later stages' are not supported by mean-curve inspection. The authors should fit a mixed-effects model (e.g., linear mixed model) to the individual longitudinal measurements, with a time-by-subtype interaction term, and report the interaction p-values and effect sizes. This applies to Fig. 5(a), (b), and (c).
  5. [Section 2.3.4] The paper reports three of 28 clinical symptoms as significant (headache p=6.50e-3, urinary frequency p=2.08e-2, falling p=1.51e-2) with no multiple-testing correction. With 28 χ2 tests, the Bonferroni threshold is 0.05/28 = 1.79e-3, and none of the reported p-values reaches this threshold. The claim that these symptoms 'reaching the significance level' is therefore incorrect under the stated correction policy. Either use an FDR/Bonferroni correction and report adjusted p-values, or clearly label these as uncorrected exploratory findings.
minor comments (7)
  1. [Section 4.2.2, Eq. (3)] The generator loss contains a typo: 'Dm(Y′m − 1)' should presumably be '(Dm(Y′m) − 1)'. Please correct the equation.
  2. [Section 2.3.5] The text states 'three discontinuous covariates ... with t-test and two continuous covariates ... with χ2-test,' but the correct assignment is the reverse: discrete covariates require χ2 and continuous covariates require t-test. The subsequent p-values suggest the intended tests were used, but the wording is wrong.
  3. [Section 2.2] The interpretation step for genetic weights is underspecified: 'We performed a linear regression between the input (SNPs) and output of Mamba.' Please state the regression target explicitly (e.g., Mamba output for each tissue group), the standardization of SNPs, and whether coefficients were averaged across cross-validation folds.
  4. [Section 4.2.3] The GMM application is described as predicting 'disease status' when deployed on MCI. Clarify whether GMM is clustering the fused representations or performing a soft classification relative to HC/AD class centers, and define the input dimensionality and number of components selected.
  5. [Table 2] For the CH index, higher values are better, yet the complete model ('with RGE') has a slightly lower CH (419.80) than the reduced model ('without RGE', 420.38). The text says the proposed framework obtained 'higher CH values' than other methods; this is consistent for k-means/GMM but not for the ablation comparison. Please clarify or report the variance/standard error of these indices.
  6. [General] There is no code or data availability statement for the method. Since the paper introduces a new algorithmic framework, sharing code (or at least a detailed pseudo-code for the GMM labeling step) would substantially improve reproducibility.
  7. [General] Numerous typographical errors remain, including 'data avaliability' (Section 5), 'MCIs with different level risks' (Abstract/introduction), 'were first accessed' (Section 2.1), 'comment genetic basis' (Section 2.5.2), and 'ore suitable' (Discussion). A careful proofread is needed.

Circularity Check

2 steps flagged · score 6.0 of 10

GWAS and imaging-QT validations are double-dipped: subtype labels were generated from the same SNP and imaging data; outcome-based risk claim is asserted without reported statistics.

  1. fitted input called prediction [Section 2.5.1 (GWAS), with Section 2.1 and Section 4.2.1 (RGE module)]
    "BigFirst was first trained on HCs and ADs populations ... Then the well-trained model was deployed on 417 MCI patients and yielded 86 low-risk MCIs and 331 high-risk MCIs. ... The risk genetic information extraction (RGE) module extracts disease-related genetic representations from the whole-genome. We conducted a genome-wide association study (GWAS) to investigate the genetic basis underpinning the difference of the two MCI risk subtypes."

    The RGE module feeds the same 359,997 SNPs into Mamba (Eq. 1), and the learned representations are then clustered by GMM (Section 4.2.3) to assign MCI subjects to low/high-risk subtypes. The GWAS in Section 2.5.1 tests SNPs against these subtype labels. Because the labels are a learned function of the SNP data itself, any SNP with non-zero model weight will tend to associate with the labels; the reported loci (CACNA1C, ABCA13) are therefore not an independent genetic validation but a reflection of the model's SNP input. This is double-dipping: the 'genetic underpinning' result is substantially forced by the RGE module's use of the same SNPs.

  2. fitted input called prediction [Section 2.3.1 and Section 2.7.1, with Section 2.1 and Section 4.2.3]
    "BigFirst is trained on HC and AD subjects to ensure that the differences between MCI subtypes are disease-related. ... We investigated the differences between two MCI subtypes via testing brain imaging QTs derived from AV45-PET, FDG-PET, and VBM-MRI, using t-test with Bonferroni correction."

    These three imaging modalities are the same data used to train the HC/AD classifier and to construct the MCI representation space that is clustered by GMM (Section 4.2.3). Therefore the two subtype labels are a function of the imaging QTs themselves; testing those same QTs for group differences is not an independent validation. The existence of widespread differences is expected from the clustering construction, so the 'substantive differences in neuroimaging QTs' claim does not by itself show that the subtypes capture disease risk. The regional pattern may add information, but the validation is partially circular for these modalities.

full rationale

The main derivation chain is: (1) train a multimodal model on HC vs AD using AV45-PET, FDG-PET, VBM-MRI and 359,997 SNPs; (2) apply it to MCI subjects and cluster their learned representations with GMM; (3) name the clusters low/high-risk; (4) validate by showing the clusters differ in genetics, imaging, biomarkers, cognition, trajectories and sMCI/pMCI concordance. Steps (1)-(2) mean the subtype labels are functions of the SNP and imaging inputs. Therefore two of the validation analyses are double-dipped: the GWAS (2.5.1) tests the same SNPs that the RGE module used to build the representations, and the imaging-QT comparisons (2.3.1, 2.7.1) test the same imaging modalities that defined the representation space. Those results are not independent confirmation. The fluid biomarkers, cognitive composites, and longitudinal trajectories were not training inputs, so they provide genuinely independent evidence, and the Discussion candidly states the HC/AD-similarity assumption. However, the key outcome-level validation—concordance with sMCI/pMCI conversion over two years—is asserted only as 'substantial concordance' with no contingency table, risk ratio, p-value, or CI (Section 2.6); and the paper never specifies how the two GMM clusters were assigned 'low' vs 'high' risk. Thus the independent, non-circular support for the central risk claim is substantially under-reported. The only self-citation (ref. 72) is a literature citation for frontal/cingulum ROIs and is not load-bearing. Overall, the paper has two clear by-construction/double-dipping validation steps, so partial circularity (score 6), but the central risk-stratification claim retains independent (if under-quantified) longitudinal content.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central methodological assumptions are the HC/AD proxy for risk and the GTEx-derived SNP selection. The main free parameters are the grid-searched network dimensions, the temperature, and the post-hoc GWAS threshold and cluster count. No new physical entities are introduced; 'pseudo-brain' maps are internal representations, not new entities.

free parameters (6)
  • GWAS significance threshold = 1e-5
    Set post-hoc because no SNP reached the conventional 1e-8 genome-wide threshold (Section 2.5.1).
  • Number of GMM components = 2
    The two MCI subtypes are defined by a 2-component GMM on learned representations; no model selection criterion is given (Section 4.2.3).
  • SSM state dimension (Mamba) = 2
    Selected by grid search on the same ADNI data (Supplementary A.1, A.3).
  • Multi-modal representation dimension = 64
    Selected by grid search (Supplementary A.1, A.3).
  • MLP structure = [128, 256]
    Selected by grid search (Supplementary A.1).
  • Contrastive temperature tau = 0.07
    Set as suggested in reference [70] (Section A.1).
assumptions (4)
  • ad hoc to paper Training a classifier on HC vs AD produces a valid monotonic risk axis for MCI
    The method is 'grounded in the assumption that the physiological characteristics of low-risk MCI populations resembled those of HCs, whereas high-risk MCI populations were more akin to AD patients' (Discussion).
  • domain assumption Brain tissue related sQTLs from GTEx v8 capture the relevant genetic influence on brain imaging phenotypes
    Section 4.1.2 restricts the 359,997 SNPs to sQTLs of 13 brain tissues from GTEx v8, assuming these loci are the ones influencing brain imaging.
  • ad hoc to paper The learned representations support a two-cluster structure in MCI
    The pipeline applies a 2-component GMM to MCI representations without a stated model selection procedure (Section 4.2.3).
  • ad hoc to paper Statistical thresholds are valid without correction for all exploratory tests
    The authors apply Bonferroni correction to imaging QTs and some fluid biomarkers, but not to the 28 clinical symptoms (Section 2.3.4) or the four cognitive domains (Section 2.3.3).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Comprehensively stratifying MCIs into distinct risk subtypes based on brain imaging genetics fusion learning." pith.science (2026). https://pith.science/paper/VXGXTAVE

@misc{pith2026250818058,
  author       = {Pith},
  title        = {Pith review of: Comprehensively stratifying MCIs into distinct risk subtypes based on brain imaging genetics fusion learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VXGXTAVE}},
  note         = {Machine review of arXiv:2508.18058}
}
read the original abstract

Mild cognitive impairment (MCI) is the prodromal stage of Alzheimer's disease (AD) and thus enrolling MCI subjects to undergo clinical trials is worthwhile. However, MCI groups usually show significant diversity and heterogeneity in the pathology and symptom, which pose great challenge to accurately select appropriate subjects. This study aimed to stratify MCI subjects into distinct subgroups with substantial difference in the risk of transitioning to AD by fusing multimodal brain imaging genetic data. The integrated imaging genetics method comprised three modules, i.e., the whole-genome-oriented risk genetic information extraction module (RGE), the genetic-to-brain mapping module (RG2PG), and the genetic-guided pseudo-brain fusion module (CMPF). We used data from AD Neuroimaging Initiative (ADNI) and identified two MCI subtypes, called low-risk MCI (lsMCI) and high-risk MCI (hsMCI). We also validated that the two subgroups showed distinct patterns of in terms of multiple biomarkers including genetics, demographics, fluid biomarkers, brain imaging features, clinical symptoms and cognitive functioning at baseline, as well as their longitudinal developmental trajectories. Furthermore, we also identified potential biomarkers that may implicate the risk of MCIs, providing critical insights for patient stratification at early stage.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

76 extracted references · 75 canonical work pages

  1. [1]

    Nature review s Disease primers 7(1), 33 (2021)

    Knopman, D.S., Amieva, H., Petersen, R.C., Ch´ etelat, G., Holtzman , D.M., Hyman, B.T., Nixon, R.A., Jones, D.T.: Alzheimer disease. Nature review s Disease primers 7(1), 33 (2021)

    work page 2021

  2. [2]

    : Mild cognitive impairment (mci) in medical practice: a critical review of the concept and new diagnostic procedure

    Portet, F., Ousset, P., Visser, P., Frisoni, G., Nobili, F., Scheltens, P., Vellas, B., Touchon, J., Alzheimer’s Disease (EADC, M.W.G., et al. : Mild cognitive impairment (mci) in medical practice: a critical review of the concept and new diagnostic procedure. report of the mci working group of the eur opean consortium on alzheimer’s disease. Journal of N...

    work page 2006

  3. [3]

    Alzh eimer’s & dementia 14(4), 535–562 (2018)

    Jack Jr, C.R., Bennett, D.A., Blennow, K., Carrillo, M.C., Dunn, B., Hae berlein, S.B., Holtzman, D.M., Jagust, W., Jessen, F., Karlawish, J., et al.: Nia-aa research framework: toward a biological definition of alzheimer’s disease. Alzh eimer’s & dementia 14(4), 535–562 (2018)

    work page 2018

  4. [4]

    CNS spectrums 24(1), 78–87 (2019)

    Anderson, N.D.: State of the science on mild cognitive impairment (m ci). CNS spectrums 24(1), 78–87 (2019)

    work page 2019

  5. [5]

    : The heterogeneity of mild cognitive impairment: A neuropsychological analysis

    Libon, D.J., Xie, S.X., Eppig, J., Wicas, G., Lamar, M., Lippa, C., Bettch er, B.M., Price, C.C., Giovannetti, T., Swenson, R., et al. : The heterogeneity of mild cognitive impairment: A neuropsychological analysis. Journal of th e International Neuropsychological Society 16(1), 84–93 (2010)

    work page 2010

  6. [6]

    Alzheimer’s & Dementia 6(3), 239–246 (2010)

    Aisen, P.S., Petersen, R.C., Donohue, M.C., Gamst, A., Raman, R., Th omas, R.G., Walter, S., Trojanowski, J.Q., Shaw, L.M., Beckett, L.A., et al.: Clinical core of the alzheimer’s disease neuroimaging initiative: progress and plans . Alzheimer’s & Dementia 6(3), 239–246 (2010)

    work page 2010

  7. [7]

    : Early versus late mci: Improved mci staging using a neuropsycholo gical approach

    Edmonds, E.C., McDonald, C.R., Marshall, A., Thomas, K.R., Eppig, J., Weigand, A.J., Delano-Wood, L., Galasko, D.R., Salmon, D.P., Bondi, M.W., et al. : Early versus late mci: Improved mci staging using a neuropsycholo gical approach. Alzheimer’s & Dementia 15(5), 699–708 (2019)

    work page 2019

  8. [8]

    : Mild cognitive impairment–beyond controversies, towards a consensus: repor t of the interna- tional working group on mild cognitive impairment

    Winblad, B., Palmer, K., Kivipelto, M., Jelic, V., Fratiglioni, L., Wahlund, L .- O., Nordberg, A., B¨ ackman, L., Albert, M., Almkvist, O., et al. : Mild cognitive impairment–beyond controversies, towards a consensus: repor t of the interna- tional working group on mild cognitive impairment. Journal of interna l medicine 256(3), 240–246 (2004)

    work page 2004

Show all 76 references
  1. [9]

    Dementia and geriatric cognitive disorders 28(6), 541–549 (2010)

    Brambati, S.M., Belleville, S., Kergoat, M.-J., Chayer, C., Gauthier, S ., Joubert, S.: Single-and multiple-domain amnestic mild cognitive impairment: two sid es of the same coin. Dementia and geriatric cognitive disorders 28(6), 541–549 (2010)

    2010

  2. [10]

    : Genetic data and cognitively defined late-onset alzheimer’s disease subgroups

    Mukherjee, S., Mez, J., Trittschuh, E.H., Saykin, A.J., Gibbons, L .E., Fardo, 26 D.W., Wessels, M., Bauman, J., Moore, M., Choi, S.-E., et al. : Genetic data and cognitively defined late-onset alzheimer’s disease subgroups. Molec ular psychiatry 25(11), 2942–2951 (2020)

    2020

  3. [11]

    Hughes, T.F., Snitz, B.E., Ganguli, M.: Should mild cognitive impairment be subtyped? Current opinion in psychiatry 24(3), 237–242 (2011)

    2011

  4. [12]

    Alzheimer’s & Dementia 10(5), 511– 521 (2014)

    Nettiksimmons, J., DeCarli, C., Landau, S., Beckett, L.: Biological heterogeneity in adni amnestic mild cognitive impairment. Alzheimer’s & Dementia 10(5), 511– 521 (2014)

    2014

  5. [13]

    Nature communications 15(1), 354 (2024)

    Yang, Z., Wen, J., Abdulkadir, A., Cui, Y., Erus, G., Mamourian, E., M elhem, R., Srinivasan, D., Govindarajan, S.T., Chen, J., et al.: Gene-sgan: discovering disease subtypes with imaging and genetic signatures via multi-view weakly-su pervised deep clustering. Nature communi...

    2024

  6. [14]

    Brain 140(3), 735–747 (2017)

    Dong, A., Toledo, J.B., Honnorat, N., Doshi, J., Varol, E., Sotiras, A., Wolk, D., Trojanowski, J.Q., Davatzikos, C., Initiative, A.D.N.: Heterogeneit y of neu- roanatomical patterns in prodromal alzheimer’s disease: links to co gnition, progression and biomarkers. Brain 140(3...

    2017

  7. [15]

    : Two progression patterns in alzheimer’s disease identified from structural mri in a me mory clinic- based cohort

    Baumeister, H., Ziegler, G., Wolfsgruber, S., Ramirez, A., Peters , O., Priller, J., Schneider, A., Wiltfang, J., Fliessbach, K., B¨ urger, K., et al. : Two progression patterns in alzheimer’s disease identified from structural mri in a me mory clinic- based cohort. Alzheimer’s...

    2022

  8. [16]

    : A general- izable data-driven model of atrophy heterogeneity and progress ion in a memory clinic setting

    Baumeister, H., Vogel, J.W., Insel, P.S., Kleineidam, L., Wolfsgruber , S., Stark, M., Gellersen, H.M., Yakupov, R., Schmid, M., L¨ usebrink, F., et al. : A general- izable data-driven model of atrophy heterogeneity and progress ion in a memory clinic setting. Alzheimer’s & De...

    2024

  9. [17]

    Alzheimer’s & Dementia 20(5), 3442–3454 (2024)

    Edmonds, E.C., Thomas, K.R., Rapcsak, S.Z., Lindemer, S.L., Delano -Wood, L., Salmon, D.P., Bondi, M.W.: Data-driven classification of cognitively no rmal and mild cognitive impairment subtypes predicts progression in the na cc dataset. Alzheimer’s & Dementia 20(5), 3442–3454 (2024)

    2024

  10. [18]

    : Fdg-pet markers of heterogeneity and different risk of progression in amnestic mci

    Caminiti, S.P., De Francesco, S., Tondo, G., Galli, A., Redolfi, A., Pera ni, D., Initiative, A.D.N., Project, I., Cappa, S.F., Cotelli, M., et al. : Fdg-pet markers of heterogeneity and different risk of progression in amnestic mci. A lzheimer’s & Dementia 20(1), 159–172 (2024)

    2024

  11. [19]

    Neurology 76(6), 511–517 (2011)

    Petrella, J., Sheldon, F., Prince, S., Calhoun, V., Doraiswamy, P.: D efault mode network connectivity in stable vs progressive mild cognitive impa irment. Neurology 76(6), 511–517 (2011)

    2011

  12. [20]

    Neurology 80(11), 1048–1056 (2013)

    Prestia, A., Caroli, A., Van Der Flier, W.M., Ossenkoppele, R., Van Be rckel, B., Barkhof, F., Teunissen, C.E., Wall, A.E., Carter, S.F., Sch¨ oll, M.,et al.: Prediction 27 of dementia in mci patients based on core diagnostic markers for alz heimer disease. Neurology 80(11), 1...

    2013

  13. [21]

    : A novel grading biomarker for the prediction of conversion from mild cognitive impairment to alzheimer’s disease

    Tong, T., Gao, Q., Guerrero, R., Ledig, C., Chen, L., Rueckert, D ., Initia- tive, A.D.N., et al. : A novel grading biomarker for the prediction of conversion from mild cognitive impairment to alzheimer’s disease. IEEE Transactio ns on Biomedical Engineering 64(1), 155–165 (2016)

    2016

  14. [22]

    Neurology 87(20), 2108–2116 (2016)

    Edmonds, E.C., Eppig, J., Bondi, M.W., Leyden, K.M., Goodwin, B., Dela no- Wood, L., McDonald, C.R., Initiative, A.D.N.: Heterogeneous cortical a trophy patterns in mci not captured by conventional diagnostic criteria. Neurology 87(20), 2108–2116 (2016)

    2016

  15. [23]

    : Genetic heterogeneity of four mci/ad neuroanatomical dimensions discovered via deep learning

    Wen, J., Cui, Y., Yang, Z., Bao, J., Chen, J., Erus, G., Abdulkadir, A ., Mamourian, E., Singh, A., Yang, S., et al. : Genetic heterogeneity of four mci/ad neuroanatomical dimensions discovered via deep learning. Alz heimer’s & Dementia 18, 065223 (2022)

    2022

  16. [24]

    Archives of Gerontology a nd Geriatrics 94, 104376 (2021)

    Katayama, O., Lee, S., Bae, S., Makino, K., Shinkai, Y., Chiba, I., Ha rada, K., Shimada, H.: Lifestyle changes and outcomes of older adults with mild c ognitive impairment: A 4-year longitudinal study. Archives of Gerontology a nd Geriatrics 94, 104376 (2021)

    2021

  17. [25]

    Alzheimer’s & Dementia 19(4), 1428–1439 (2023)

    Katabathula, S., Davis, P.B., Xu, R.: Comorbidity-driven multi-mod al subtype analysis in mild cognitive impairment of alzheimer’s disease. Alzheimer’s & Dementia 19(4), 1428–1439 (2023)

    2023

  18. [26]

    Sauver, J.L ., Mielke, M.M., Geda, Y.E., Machulda, M.M., Knopman, D.S., Petersen, R.C., et al

    Vassilaki, M., Aakre, J.A., Cha, R.H., Kremers, W.K., St. Sauver, J.L ., Mielke, M.M., Geda, Y.E., Machulda, M.M., Knopman, D.S., Petersen, R.C., et al. : Multimorbidity and risk of mild cognitive impairment. Journal of the Ame rican Geriatrics Society 63(9), 1783–1790 (2015)

    2015

  19. [27]

    : Atrophy patterns in early clinical stages across distinct phenoty pes of a lzheimer’s disease

    Ossenkoppele, R., Cohn-Sheehy, B.I., La Joie, R., Vogel, J.W., M¨ o ller, C., Lehmann, M., Berckel, B.N., Seeley, W.W., Pijnenburg, Y.A., Gorno-Tem pini, M.L., et al. : Atrophy patterns in early clinical stages across distinct phenoty pes of a lzheimer’s disease. Human brain m...

    2015

  20. [28]

    Medical Clin ics 97(3), 485–494 (2013)

    Chew, J., Silverman, D.H.: Fdg-pet in early ad diagnosis. Medical Clin ics 97(3), 485–494 (2013)

    2013

  21. [29]

    Journal of Molecular Neuroscience 20, 203–206 (2003)

    Hardy, J.: The relationship between amyloid and tau. Journal of Molecular Neuroscience 20, 203–206 (2003)

    2003

  22. [30]

    https://mathworld

    Weisstein, E.W.: Bonferroni correction. https://mathworld. w olfram. com/ (2004)

    2004

  23. [31]

    Neuroimage 115, 117–137 (2015)

    Iglesias, J.E., Augustinack, J.C., Nguyen, K., Player, C.M., Player, A., Wright, 28 M., Roy, N., Frosch, M.P., McKee, A.C., Wald, L.L., et al.: A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution mr i: Application to adaptive segmentation of...

    2015

  24. [32]

    Gigascience 4(1), 13742–015 (2015)

    Chang, C.C., Chow, C.C., Tellier, L.C., Vattikuti, S., Purcell, S.M., Lee, J.J.: Second-generation plink: rising to the challenge of larger and richer datasets. Gigascience 4(1), 13742–015 (2015)

    2015

  25. [33]

    PLoS computational biology 11(4), 1004219 (2015)

    De Leeuw, C.A., Mooij, J.M., Heskes, T., Posthuma, D.: Magma: gen eralized gene-set analysis of gwas data. PLoS computational biology 11(4), 1004219 (2015)

    2015

  26. [34]

    In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

    Li, Y., Hu, P., Liu, Z., Peng, D., Zhou, J.T., Peng, X.: Contrastive clu stering. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, pp. 8547–8555 (2021)

    2021

  27. [35]

    arXiv preprint arXiv:2312.00752 (2023)

    Gu, A., Dao, T.: Mamba: Linear-time sequence modeling with select ive state spaces. arXiv preprint arXiv:2312.00752 (2023)

    2023 arXiv

  28. [36]

    Cells 8(8), 826 (2019)

    Ng, L.F., Kaur, P., Bunnag, N., Suresh, J., Sung, I.C.H., Tan, Q.H., G ruber, J., Tolwinski, N.S.: Wnt signaling in disease. Cells 8(8), 826 (2019)

    2019

  29. [37]

    Biomedic ine & Pharmacotherapy 143, 112187 (2021)

    Tabnak, P., Mafakheri, A., Emsailpoor, Z.H., Kazemi, T., Shekari, N .: Regula- tory interplay between micrornas and wnt pathway in glioma. Biomedic ine & Pharmacotherapy 143, 112187 (2021)

    2021

  30. [38]

    : Genetic analysis of vcp and wash complex genes in a german cohort of sporad ic als-ftd patients

    Tuerk, M., Schroeder, R., Khuller, K., Hofmann, A., Berwanger, C., Ludolph, A.C., Dekomien, G., Mueller, K., Weishaupt, J.H., Thiel, C.T., et al. : Genetic analysis of vcp and wash complex genes in a german cohort of sporad ic als-ftd patients. Neurobiology of Aging 56, 213–1 (2017)

    2017

  31. [39]

    : Heterozygous capza2 mutations cause global developmental delay, hypotonia with epilepsy: a case r eport and the literature review

    Zhang, X.-M., Xu, K.-L., Kong, J.-H., Dong, G., Dong, S.-J., Yang, Z .-X., Xu, S.-J., Wang, L., Luo, S.-Y., Zhang, Y.-D., et al. : Heterozygous capza2 mutations cause global developmental delay, hypotonia with epilepsy: a case r eport and the literature review. Journal of Huma...

    2024

  32. [40]

    Movement disorders: official journal of the Movement Disorder So ciety 24(15), 2277–2281 (2009)

    Lovrecic, L., Kastrin, A., Kobal, J., Pirtosek, Z., Krainc, D., Pete rlin, B.: Gene expression changes in blood as a putative biomarker for huntington ’s disease. Movement disorders: official journal of the Movement Disorder So ciety 24(15), 2277–2281 (2009)

    2009

  33. [41]

    Alzheimer ’s & Dementia 12, 1146–1146 (2016)

    Parrado, A.: P4-294: Identification of genome-wide significant associations in the alzheimer’s disease sequencing project dominican families with var iants in ugt2b28, axin2, syce1, bcat1, fam13a, and dq594761. Alzheimer ’s & Dementia 12, 1146–1146 (2016)

    2016

  34. [42]

    Genes 15(1), 87 (2024)

    Janda, E., Boutin, J.A., De Lorenzo, C., Arbitrio, M.: Polymorphism s and 29 pharmacogenomics of nqo2: the past and the future. Genes 15(1), 87 (2024)

    2024

  35. [43]

    Autophagy 11(7), 1063–1080 (2015)

    Janda, E., Lascala, A., Carresi, C., Parafati, M., Aprigliano, S., Ru sso, V., Savoia, C., Ziviani, E., Musolino, V., Morani, F., et al.: Parkinsonian toxin-induced oxida- tive stress inhibits basal autophagy in astrocytes via nqo2/quinon e oxidoreductase 2: Implications for n...

    2015

  36. [44]

    Alzheimer’s & Dementia 14(11), 1460–1469 (2018)

    Schindler, S.E., Gray, J.D., Gordon, B.A., Xiong, C., Batrla-Uterma nn, R., Quan, M., Wahl, S., Benzinger, T.L., Holtzman, D.M., Morris, J.C., et al.: Cerebrospinal fluid biomarkers measured by elecsys assays compared to amyloid ima ging. Alzheimer’s & Dementia 14(11), 1460–1469 (2018)

    2018

  37. [45]

    : Csf biomarkers of alzheimer’s disease concord with amyloid- β pet and predict clinical progression: a study of fully automated immunoassays in biofinder and adni cohort s

    Hansson, O., Seibyl, J., Stomrud, E., Zetterberg, H., Trojanow ski, J.Q., Bittner, T., Lifke, V., Corradini, V., Eichenlaub, U., Batrla, R., et al. : Csf biomarkers of alzheimer’s disease concord with amyloid- β pet and predict clinical progression: a study of fully automated ...

    2018

  38. [46]

    Alzheimer’s & Dementia 20(11), 8097–8112 (2024)

    Lowe, V.J., Mester, C.T., Lundt, E.S., Lee, J., Ghatamaneni, S., Alg eciras- Schimnich, A., Campbell, M.R., Graff-Radford, J., Nguyen, A., Min, H.-K., et al.: Amyloid pet detects the deposition of brain a β earlier than csf fluid biomarkers. Alzheimer’s & Dementia 20(11), 8097–8...

    2024

  39. [47]

    Alzheimer’s research & therapy 10, 1–10 (2018)

    Lewczuk, P., Ermann, N., Andreasson, U., Schultheis, C., Podho rna, J., Spitzer, P., Maler, J.M., Kornhuber, J., Blennow, K., Zetterberg, H.: Plasma ne urofila- ment light as a potential biomarker of neurodegeneration in alzheime r’s disease. Alzheimer’s research & therapy 10, ...

    2018

  40. [48]

    JAMA neurology 76(7), 791–799 (2019)

    Mattsson, N., Cullen, N.C., Andreasson, U., Zetterberg, H., Blen now, K.: Asso- ciation between longitudinal plasma neurofilament light and neurodeg eneration in patients with alzheimer disease. JAMA neurology 76(7), 791–799 (2019)

    2019

  41. [49]

    : Diagnostic value of plasma phosphorylated tau181 in alzheimer’s disease and frontotem poral lobar degeneration

    Thijssen, E.H., La Joie, R., Wolf, A., Strom, A., Wang, P., Iaccarino , L., Bourakova, V., Cobigo, Y., Heuer, H., Spina, S., et al. : Diagnostic value of plasma phosphorylated tau181 in alzheimer’s disease and frontotem poral lobar degeneration. Nature medicine 26(3), 387–397 (2020)

    2020

  42. [50]

    Peptides 23(7), 1205–1214 (2002)

    Andreasen, N., Blennow, K.: β-amyloid (a β) protein in cerebrospinal fluid as a biomarker for alzheimer’s disease. Peptides 23(7), 1205–1214 (2002)

    2002

  43. [51]

    Current Alzheimer Research 11(3), 274–282 (2014)

    Weiler, M., Fukuda, A., HP Massabki, L., M Lopes, T., R Franco, A., P Dama- sceno, B., Cendes, F., LF Balthazar, M.: Default mode, executive fu nction, and language functional connectivity networks are compromised in mild a lzheimer s disease. Current Alzheimer Research 11(3),...

    2014

  44. [52]

    Cortex 48(4), 429–446 (2012)

    Stopford, C.L., Thompson, J.C., Neary, D., Richardson, A.M., Sno wden, J.S.: Working memory, attention, and executive function in alzheimer’s dis ease and 30 frontotemporal dementia. Cortex 48(4), 429–446 (2012)

    2012

  45. [53]

    Neurological Scie nces 45(3), 1017–1030 (2024)

    Cermelli, A., Roveta, F., Giorgis, L., Boschi, S., Grassini, A., Ferran des, F., Lom- bardo, C., Marcinn` o, A., Rubino, E., Rainero, I.: Is headache a risk f actor for dementia? a systematic review and meta-analysis. Neurological Scie nces 45(3), 1017–1030 (2024)

    2024

  46. [54]

    Journal of advanced nursing 65(5), 922–933 (2009)

    H¨ arlein, J., Dassen, T., Halfens, R.J., Heinze, C.: Fall risk factor s in older people with dementia or cognitive impairment: a systematic review. Journal of advanced nursing 65(5), 922–933 (2009)

    2009

  47. [55]

    Dementia and Neurocognitive Disorders 19(3), 77 (2020)

    Na, H.R., Cho, S.T.: Relationship between lower urinary tract dysf unction and dementia. Dementia and Neurocognitive Disorders 19(3), 77 (2020)

    2020

  48. [56]

    Neurology 70(4), 299–303 (2008)

    Ransmayr, G., Holliger, S., Schletterer, K., Heidler, H., Deibl, M., Po ewe, W., Madersbacher, H., Kiss, G.: Lower urinary tract symptoms in demen tia with lewy bodies, parkinson disease, and alzheimer disease. Neurology 70(4), 299–303 (2008)

    2008

  49. [57]

    : A cytogenetic abnormality and rare coding variants identify abca13 as a candidate gene in schizophrenia, bipolar disorder, and depression

    Knight, H.M., Pickard, B.S., Maclean, A., Malloy, M.P., Soares, D.C., Mc Rae, A.F., Condie, A., White, A., Hawkins, W., McGhee, K., et al. : A cytogenetic abnormality and rare coding variants identify abca13 as a candidate gene in schizophrenia, bipolar disorder, and depressio...

    2009

  50. [58]

    Schiz ophrenia bulletin 44(5), 958–965 (2018)

    Moon, A.L., Haan, N., Wilkinson, L.S., Thomas, K.L., Hall, J.: Cacna1c: associa- tion with psychiatric disorders, behavior, and neurogenesis. Schiz ophrenia bulletin 44(5), 958–965 (2018)

    2018

  51. [59]

    Neuroim- age 11(6), 805–821 (2000)

    Ashburner, J., Friston, K.J.: Voxel-based morphometry—the m ethods. Neuroim- age 11(6), 805–821 (2000)

    2000

  52. [60]

    Neuroimage 15(1), 273–289 (2002)

    Tzourio-Mazoyer, N., Landeau, B., Papathanassiou, D., Crivello , F., Etard, O., Delcroix, N., Mazoyer, B., Joliot, M.: Automated anatomical labeling of acti- vations in spm using a macroscopic anatomical parcellation of the mni mri single-subject brain. Neuroimage 15(1), 273–2...

    2002

  53. [61]

    Nature neuroscience 24(9), 1313–1323 (2021)

    Hartl, C.L., Ramaswami, G., Pembroke, W.G., Muller, S., Pintacuda, G ., Saha, A., Parsana, P., Battle, A., Lage, K., Geschwind, D.H.: Coexpression n et- work architecture reveals the brain-wide and multiregional basis of disease susceptibility. Nature neuroscience 24(9), 1313–...

    2021

  54. [62]

    Science 369(6509), 1318–1330 (2020)

    Consortium, G.: The gtex consortium atlas of genetic regulator y effects across human tissues. Science 369(6509), 1318–1330 (2020)

    2020

  55. [63]

    : Eye-gaze guided multi-modal alignment for medical representa- tion learning

    Ma, C., Jiang, H., Chen, W., Li, Y., Wu, Z., Yu, X., Liu, Z., Guo, L., Zhu, D., 31 Zhang, T., et al. : Eye-gaze guided multi-modal alignment for medical representa- tion learning. Advances in Neural Information Processing Systems 37, 6126–6153 (2024)

    2024

  56. [64]

    IEEE Transactio ns on Medical Imaging 41(9), 2348–2359 (2022)

    Ko, W., Jung, W., Jeon, E., Suk, H.-I.: A deep generative–discrimin ative learning for multimodal representation in imaging genetics. IEEE Transactio ns on Medical Imaging 41(9), 2348–2359 (2022)

    2022

  57. [65]

    In: 202 1 IEEE International Conference on Data Mining (ICDM), pp

    Ko, W., Jung, W., Jeon, E., Mulyadi, A.W., Suk, H.-I.: Engine: Enhanc ing neuroimaging and genetic information by neural embedding. In: 202 1 IEEE International Conference on Data Mining (ICDM), pp. 1162–1167 (2021). IEEE

    2021

  58. [66]

    In: Proceedings of the IEEE I nternational Conference on Computer Vision, pp

    Mao, X., Li, Q., Xie, H., Lau, R.Y., Wang, Z., Paul Smolley, S.: Least sq uares generative adversarial networks. In: Proceedings of the IEEE I nternational Conference on Computer Vision, pp. 2794–2802 (2017)

    2017

  59. [67]

    Molecular psychiatry 28(6), 2423–2432 (2023)

    Avelar-Pereira, B., Belloy, M.E., O’Hara, R., Hosseini, S.H., Initiative , A.D.N.: Decoding the heterogeneity of alzheimer’s disease diagnosis and pro gression using multilayer networks. Molecular psychiatry 28(6), 2423–2432 (2023)

    2023

  60. [68]

    Advances in neural information processing systems 33, 18661–18673 (2020)

    Khosla, P., Teterwak, P., Wang, C., Sarna, A., Tian, Y., Isola, P., M aschinot, A., Liu, C., Krishnan, D.: Supervised contrastive learning. Advances in neural information processing systems 33, 18661–18673 (2020)

    2020

  61. [69]

    In: Pr oceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Yuan, X., Cheng, G., Yan, K., Zeng, Q., Han, J.: Small object dete ction via coarse-to-fine proposal generation and imitation learning. In: Pr oceedings of the IEEE/CVF International Conference on Computer Vision, pp. 631 7–6327 (2023)

    2023

  62. [70]

    In: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern R ecognition, pp

    Wang, F., Liu, H.: Understanding the behaviour of contrastive lo ss. In: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern R ecognition, pp. 2495–2504 (2021)

    2021

  63. [71]

    Advances in neural information processing systems 37, 103031–103063 (2025)

    Liu, Y., Tian, Y., Zhao, Y., Yu, H., Xie, L., Wang, Y., Ye, Q., Jiao, J., Liu , Y.: Vmamba: Visual state space model. Advances in neural information processing systems 37, 103031–103063 (2025)

    2025

  64. [72]

    IEEE transactions on medical imag ing 39(11), 3416–3428 (2020)

    Du, L., Liu, F., Liu, K., Yao, X., Risacher, S.L., Han, J., Saykin, A.J., S hen, L.: Associating multi-modal brain imaging phenotypes and genetic risk fa ctors via a dirty multi-task learning method. IEEE transactions on medical imag ing 39(11), 3416–3428 (2020)

    2020

  65. [73]

    NeuroImage: Clinical 31, 102750 (2021)

    Vanhoutte, M., Landeau, B., Sherif, S., Sayette, V., Dautricou rt, S., Abbas, A., Manrique, A., Chocat, A., Ch´ etelat, G.: Evaluation of the early-pha se [18f] av45 pet as an optimal surrogate of [18f] fdg pet in ageing and alzheimer’s clinical syndrome. NeuroImage: Clinical ...

    2021

  66. [74]

    Molecular neurodegeneration 6, 1–9 (2011)

    Mu, Y., Gage, F.H.: Adult hippocampal neurogenesis and its role in a lzheimer’s 32 disease. Molecular neurodegeneration 6, 1–9 (2011)

    2011

  67. [75]

    Annals of n eurology 49(4), 433–442 (2001)

    Chan, D., Fox, N.C., Scahill, R.I., Crum, W.R., Whitwell, J.L., Leschziner , G., Rossor, A.M., Stevens, J.M., Cipolotti, L., Rossor, M.N.: Patterns of t emporal lobe atrophy in semantic dementia and alzheimer’s disease. Annals of n eurology 49(4), 433–442 (2001)

    2001

  68. [76]

    Archives of neurology 50(9), 949–954 (1993) 33

    Killiany, R.J., Moss, M.B., Albert, M.S., Sandor, T., Tieman, J., Jolesz, F.: Tem- poral lobe regions on magnetic resonance imaging identify patients w ith early alzheimer’s disease. Archives of neurology 50(9), 949–954 (1993) 33

    1993

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.