Pith. sign in

REVIEW 5 major objections 5 minor 26 references

Transformer Model for Alzheimer's Disease Progression Prediction Using Longitudinal Visit Sequences

T0 review · 5 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Transformer models that read a patient's full visit history at once flag Alzheimer's stage transitions significantly better than recurrent networks, at a small cost in stable-patient accuracy.

desk verdict A useful, internally consistent Transformer-vs-RNN comparison for AD conversion prediction, but the underspecified model-filling imputation protocol leaves the headline result conditional on a fix. read the letter →

arxiv 2507.03899 v1 pith:RUPJRNPG submitted 2025-07-05 cs.LG cs.AIcs.CV

Mahdi Moghaddami , Clayton Schubring , Mohammad-Reza Siadat This is my paper · ORCID
classification cs.LGcs.AIcs.CV
keywords Alzheimer'sdiseaseprogressionTransformerlongitudinalpredictionconvertersequencesMCItoADconversionrecurrentneuralnetworksmodelfillingimputationTADPOLEdataset
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 Transformer model that takes a patient's sequence of clinical visits as input and predicts the diagnosis at the next visit, comparing it against LSTM, GRU, minimalRNN, and a time-series Transformer on the TADPOLE dataset. The central claim is that attention-based models are significantly better than RNNs at identifying converter sequences—the visits where a patient moves from CN to MCI or from MCI to AD—while being slightly worse on stable sequences. The authors argue that this trade-off is the right one for Alzheimer's care, because missing a conversion is clinically costlier than a false alarm. The paper also finds that cognitive test scores carry most of the predictive signal, and that the converter advantage is strongest for short visit histories, which is exactly the setting where early warning is needed.

What carries the argument

The load-bearing mechanism is multi-head self-attention over visit tokens: each visit in the sequence is embedded into a 256-dimensional vector and attends to all other visits, so the model can compare the current trajectory against earlier states without compressing history into a single hidden state. This is supported by three auxiliary pieces: 'model filling' imputation, which fills missing feature values using predictions from a minimalRNN; a temporal feature representing months until the final visit; and the construction of converter versus stable sequences, where a converter sequence ends at a diagnosis change and a stable sequence ends at the same diagnosis as the previous visit. The evaluation centers on BCA, the mean of per-class sensitivity and specificity averaged over CN, MCI, and AD, computed separately on stable and converter subsets.

What would settle it

Re-run the 10-fold cross-validation with the model-filling imputer trained separately on each training fold and applied to the held-out test fold, then compare the Transformer's converter accuracy against the RNN baselines; if the advantage shrinks to below statistical significance, the reported converter gains were artifacts of test-set leakage through imputed values.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that Transformer-based architectures allocate their predictive effort differently from RNNs: the RNNs in the study achieve high accuracy mostly by predicting that stable patients remain stable, with converter accuracy on the raw dataset as low as 0.129 for minimalRNN and 0.217 for LSTM, while the Transformer reaches 0.648 and the time-series Transformer 0.694. On a class-balanced dataset the Transformer's converter accuracy rises to 0.792 versus 0.439 for the best RNN, and the differences in accuracy, BCA, sensitivity, and specificity are statistically significant. The advantage is concentrated in the earliest visit groups and fades after roughly four prior visits. Ablations show that removing cognitive scores cuts BCA by about 16 percent relative, while removing volumetric MRI or biomarkers changes little, and training on only the last one or four visits degrades BCA by 6.6 and 5.0 percent respectively. The authors interpret the overall pattern as evidence that attention-based models are better suited for early detection, even though they underperform RNNs on stable patients.

Load-bearing premise

The paper assumes the missing-value imputation step fills in realistic measurements for features missing in 29 to 83 percent of visits without leaking the diagnosis being predicted, and it does not state whether that imputer was trained only on the training folds.

Editorial extensions

If this is right

  • If attention-based models are genuinely better at converter detection, early-warning systems for Alzheimer's disease should prefer them over RNN baselines, since catching a CN-to-MCI or MCI-to-AD transition is the clinically important event.
  • The converter advantage appears for both the raw and class-balanced datasets, so it is not merely a side effect of majority-class bias.
  • The largest gains come from short visit histories (groups 1-3), meaning a model trained on two to four prior visits could flag high-risk patients before long-term longitudinal data accumulate.
  • Because cognitive scores dominate the predictions, data-collection protocols could in principle be streamlined around cognitive testing without sacrificing most of the model's accuracy.
  • Longer visit histories do not automatically yield better predictions, so future model design needs to address data sparsity and irregular spacing in later visit groups rather than assuming more data always helps.

Reading between the lines

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

  • A critical open question the paper leaves unresolved is whether the minimalRNN imputer was fit on training folds only; if it saw the full dataset, the filled test-visit features could encode the target diagnosis and inflate all reported converter gains.
  • The study excludes 'reverter' subjects and any visits after a second conversion, so the model's real-world utility for patients with fluctuating or improving diagnoses is untested.
  • If the imputation-leakage concern is resolved and the effect persists, an immediate next experiment is to test the same architecture on external cohorts with different visit schedules and feature missingness patterns to see whether the short-history advantage generalizes.
  • The feature-ablation result implies that expensive and invasive biomarker measurements contribute little beyond cognitive scores in this model, which, if confirmed, could reduce the burden and cost of longitudinal Alzheimer's monitoring.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The manuscript proposes a Transformer-based model for predicting the next-visit clinical diagnosis (CN, MCI, AD) from a subject's longitudinal visit history in the TADPOLE/ADNI data. The authors construct stable and converter sequences, handle missing values by a 'model filling' imputation step using a minimalRNN, and compare the Transformer to LSTM, GRU, minimalRNN, and a Time Series Transformer under 10-fold cross-validation. They report that Transformer-based models significantly outperform RNNs on converter sequences, particularly for short histories, at the cost of somewhat lower stable-sequence accuracy, and they include ablations on visit-history length and feature categories.

Significance. If the reported effects are valid, the paper provides a practically useful comparison showing that attention-based models can identify AD progression converters better than RNN baselines, with public data, clearly defined stable-versus-converter sequence construction, and code availability. The per-group evaluation and the feature/ablation analyses are also strengths. However, the validity of the central claim depends on preprocessing details that are currently underspecified, especially the imputation protocol and the temporal auxiliary feature, so the contribution is conditional on those being resolved.

major comments (5)
  1. [§2.2, Table 1] The imputation protocol is underspecified in a way that can directly invalidate the central claim. Table 1 shows per-feature missingness between 29.40% and 83.38%, so most values in a typical input vector are filled. The text only says that 'model filling' uses predictions from a minimalRNN model; it does not state whether this imputer was trained on each training fold, on the full dataset, or with access to the target diagnosis. If the minimalRNN was fit on the full dataset before the 10-fold construction, the filled test values can encode the label at the target visit, and because Diagnosis is itself one of the 23 features, leakage would preferentially inflate converter detection and the reported advantage over RNNs. Please specify the exact protocol, verify in code that no test fold contributes to imputer training, and consider an ablation without model filling.
  2. [§2.2, Definitions 1-3, §2.4] The auxiliary feature 'number of months until the final visit' may constitute target leakage. Each generated sequence ends at the target visit (the conversion visit for converters and the last stable visit for stable sequences), so for input visits the feature equals the time remaining until the visit whose diagnosis is being predicted. In a prospective deployment, the date of the next visit is sometimes known, but the feature also encodes follow-up duration, which may differ systematically between converters and stable subjects; the manuscript does not show that this feature is available at inference time or that results are robust to its removal. Please add an ablation without this feature and discuss its clinical availability.
  3. [§2.5] The description of the cross-validation procedure is internally inconsistent. The text first says each fold is divided so that 10% of each group is in the test split and 90% is in the train split, then says 'We split the data temporally, ensuring that each fold has an equal proportion of sequences from each group.' A temporal split by visit date or sequence length will not, in general, preserve equal group proportions, and a stratified split is not purely temporal. Please specify the exact fold-generation algorithm, including how ties in dates and multiple sequences per subject are handled, and state whether the imputer is fit inside each fold. Without this, the standard errors and p-values across folds are not well defined.
  4. [§3, first paragraph] The statistical tests are not appropriate for the data structure. All significance claims use a two-sample Welch's t-test with n1=n2=10 on fold-level metrics. Fold metrics are paired across models because the same fold's test sequences are scored by every model, and they are not independent across folds because the training sets for different folds overlap heavily. A paired test or a mixed-effects model with fold and model as factors is the appropriate analysis; moreover, many metrics and subgroup comparisons are tested without multiple-comparison control. The very small p-values for converter accuracy would likely survive a more valid test, but the full set of claimed differences in Tables 3-6 should be re-evaluated.
  5. [§2.4, Table 2] The comparison does not control for model capacity. The Transformer is reported to have 6,603,267 parameters, but no parameter counts or compute budgets are given for LSTM, GRU, minimalRNN, or TST. If the RNN baselines have far fewer parameters, the converter advantage could be due to capacity rather than the attention mechanism. Please report parameter counts for every model and, ideally, include a capacity-matched comparison (e.g., larger RNNs or a smaller Transformer).
minor comments (5)
  1. [§2.3] Definition 1 includes CN-to-AD conversions, but the text later says converter sequences are generated only from CN-to-MCI and MCI-to-AD conversions; please clarify whether CN-to-AD sequences exist in the dataset and how they are labeled.
  2. [§2.3] The balanced dataset is created by randomly discarding stable sequences; please report the random seed or run multiple resamplings to ensure results are not artifacts of one draw.
  3. [§2.4] The description of the Transformer architecture is incomplete; state whether the model is encoder-only or encoder-decoder, how visits are embedded, and what positional encoding is used.
  4. [Tables 2 and 5] F1 scores are reported only for overall performance; reporting F1 for stable and converter subsets would aid interpretation of the trade-off.
  5. [Figures 2, 5-7] The figures are hard to read in the provided version; please enlarge axis labels and define group numbers in the captions.

Circularity Check

0 steps flagged · score 0.0 of 10

No derived construction is circular; empirical comparison rests on standard supervised learning; imputation leakage is an unsubstantiated validation risk, not a demonstrated circularity.

full rationale

The paper's central claim is an empirical performance comparison, not a derivation from first principles. The Transformer, TST, LSTM, GRU, and minimalRNN are trained with 10-fold cross-validation on fixed TADPOLE-derived sequences, and the reported converter/stable metrics are ordinary test-set evaluations. The stable/converter labels are defined by the target diagnosis, but the model input is the preceding visits only (v_1,...,v_n), so the label is not used to construct the predictor's input features. No equation in the paper is shown to be equivalent by construction to its own input; the authors do not cite themselves for any load-bearing premise, and the only potentially circular element—the 'model filling' imputation with a minimalRNN—is not specified in enough detail to establish that test labels entered the imputation. Underspecification of the imputation's train/test discipline is a legitimate validation-risk flag (missingness is 29–83% per feature), but without evidence that the imputer was fit on test folds or test targets, it is not an exhibited reduction, so it does not constitute a demonstrated circularity under the standard applied here.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central empirical claim rests on the imputation protocol, the monotonic-progression preprocessing, and the fold-based statistics; none of these is independently verified in the paper.

free parameters (2)
  • Transformer architecture hyperparameters = 4 encoder layers, 8 decoder layers, 4 attention heads, input dimensionality 256 (6,603,267 parameters)
    Selected by grid search on the dataset; the paper does not describe nested validation, so their effect on the reported cross-validation numbers is unknown.
  • minimalRNN imputation model configuration = not reported
    The model-filling imputer determines all filled feature values for sequences with 30% to 83% missingness, but its hyperparameters and fitting protocol are not given.
assumptions (5)
  • domain assumption Alzheimer's disease progression is monotonic for the analyzed population; subjects who revert to an earlier stage are excluded.
    Preprocessing drops reverter subjects and all visits after the first conversion (Section 2.2, steps 2-3). The converter and stable split and the central converter-detection claim only apply to this monotonic subset.
  • domain assumption The minimalRNN model-filling imputation produces valid feature values without leaking test-fold information.
    Imputation is applied before train and test splits are described, and the paper does not state whether the imputer was fitted on training folds only (Section 2.2, 'model filling').
  • domain assumption Clinical diagnosis labels in ADNI and TADPOLE are correct ground truth.
    All targets and conversion definitions rely on the recorded diagnosis values (Sections 2.1 and 2.3).
  • domain assumption Ten-fold cross-validation metrics can be treated as independent samples for Welch t-tests.
    The paper uses n1 = n2 = 10 t-tests across folds although folds overlap in training data (Section 3, first paragraph), which violates the independence assumption.
  • domain assumption The 23 TADPOLE-recommended features are sufficient for next-visit diagnosis prediction.
    The model uses only structured features and no raw imaging; the paper acknowledges this in Section 5 as a limitation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Transformer Model for Alzheimer's Disease Progression Prediction Using Longitudinal Visit Sequences." pith.science (2026). https://pith.science/paper/RUPJRNPG

@misc{pith2026250703899,
  author       = {Pith},
  title        = {Pith review of: Transformer Model for Alzheimer's Disease Progression Prediction Using Longitudinal Visit Sequences},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RUPJRNPG}},
  note         = {Machine review of arXiv:2507.03899}
}
read the original abstract

Alzheimer's disease (AD) is a neurodegenerative disorder with no known cure that affects tens of millions of people worldwide. Early detection of AD is critical for timely intervention to halt or slow the progression of the disease. In this study, we propose a Transformer model for predicting the stage of AD progression at a subject's next clinical visit using features from a sequence of visits extracted from the subject's visit history. We also rigorously compare our model to recurrent neural networks (RNNs) such as long short-term memory (LSTM), gated recurrent unit (GRU), and minimalRNN and assess their performances based on factors such as the length of prior visits and data imbalance. We test the importance of different feature categories and visit history, as well as compare the model to a newer Transformer-based model optimized for time series. Our model demonstrates strong predictive performance despite missing visits and missing features in available visits, particularly in identifying converter subjects -- individuals transitioning to more severe disease stages -- an area that has posed significant challenges in longitudinal prediction. The results highlight the model's potential in enhancing early diagnosis and patient outcomes.

Figures

Figures reproduced from arXiv: 2507.03899 by the authors.

Figure 1
Figure 1. Distribution of visit sequences by group number [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Distribution of Future DX across groups hidden state from the prior time point ht−1. The equations for updating the hidden state and calculat￾ing the output vary between LSTM, GRU, and min￾RNN [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. Prediction process for the Transformer To compare our proposed model to a newer archi￾tecture, we also consider the Time Series Transformer (TST) (Zerveas et al., 2021). All models are implemented using the PyTorch Python Library (Paszke et al., 2019). Some hyper￾parameter values worth mentioning are the number of encoder layers (4), the number of decoder layers (8), the number of attention heads (4), and input di￾m… view at source ↗
Figures from the paper (6 more)
Figure 3
Figure 3. Figure 3: Prediction process for RNN-based models Following Nguyen et al. (2019), we train these mod￾els by predicting clinical diagnosis for every month into the future up to six years, starting with only the first visit in the sequence v1. If the actual feature val￾ues for tim…
Figure 5
Figure 5. Figure 5: Model BCA on all sequences (a) Raw stable results (b) Balanced stable results [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Model BCA on stable sequences (a) Raw converter results (b) Balanced converter results [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Model BCA on converter sequences 9 [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Regression slope of ADAS13 scores across visit sequences, short vs. long history Despite these challenges, the model still shows ev￾idence of learning from longer visit histories, as seen in [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: ADAS13 distribution: All AD visits vs. penultimate visits for AD sequences incor￾rectly classified as MCI. Overall, these findings highlight the importance of model selection in predicting disease progression. A model that performs well on the majority class but fails …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 21 canonical work pages

  1. [1]

    Alzheimer's & Dementia, 20: 0 3708 -- 3821, 2024

    2024 alzheimer's disease facts and figures. Alzheimer's & Dementia, 20: 0 3708 -- 3821, 2024. doi:https://doi.org/10.1002/alz.13809

    doi:10.1002/alz.13809 2024

  2. [2]

    Kim, Goo-Rak Kwon, Kun Ho Lee, and Ho Yub Jung

    Samsuddin Ahmed, Kyu Yeong Choi, Jang Jae Lee, Byeong C. Kim, Goo-Rak Kwon, Kun Ho Lee, and Ho Yub Jung. Ensembles of patch-based classifiers for diagnosis of alzheimer diseases. IEEE Access, 7: 0 73373--73383, 2019

    work page 2019

  3. [3]

    Shenuarin Bhuiyan, Elizabeth A

    Sait Alp, Taymaz Akan, Md. Shenuarin Bhuiyan, Elizabeth A. Disbrow, Steven A. Conrad, John A. Vanchiere, Christopher G. Kevil, and Mohammad Alfrad Nobel Bhuiyan. Joint transformer architecture in brain 3d mri classification: its application in alzheimer's disease classification. Scientific Reports, 14, 2024

    work page 2024

  4. [4]

    Functional brain network classification for alzheimer's disease detection with deep features and extreme learning machine

    Xin Bi, Xiangguo Zhao, Hong Huang, Deyang Chen, and Yuliang Ma. Functional brain network classification for alzheimer's disease detection with deep features and extreme learning machine. Cognitive Computation, 12: 0 513 -- 527, 2019

    work page 2019

  5. [5]

    Minimalrnn: Toward more interpretable and trainable recurrent neural networks, 2018

    Minmin Chen. Minimalrnn: Toward more interpretable and trainable recurrent neural networks, 2018. URL https://arxiv.org/abs/1711.06788

    arXiv 2018

  6. [6]

    Longformer: Longitudinal Transformer for Alzheimer's Disease Classification with Structural MRIs

    Qiuhui Chen and Yi Hong. Longformer: Longitudinal transformer for alzheimer's disease classification with structural mris, 2023. URL https://arxiv.org/abs/2302.00901

    work page Pith review arXiv 2023

  7. [7]

    Learning phrase representations using rnn encoder-decoder for statistical machine translation

    Kyunghyun Cho, Bart van Merri \"e nboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. Learning phrase representations using rnn encoder-decoder for statistical machine translation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1724--1734, 2014

    work page 2014

  8. [8]

    Hippocampus analysis by combination of 3-d densenet and shapes for alzheimer's disease diagnosis

    Ruoxuan Cui and Manhua Liu. Hippocampus analysis by combination of 3-d densenet and shapes for alzheimer's disease diagnosis. IEEE Journal of Biomedical and Health Informatics, 23: 0 2099--2107, 2019

    work page 2019

Show all 26 references
  1. [9]

    Rnn-based longitudinal analysis for diagnosis of alzheimer's disease

    Ruoxuan Cui, Manhua Liu, and Alzheimer's Disease Neuroimaging Initiative. Rnn-based longitudinal analysis for diagnosis of alzheimer's disease. Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society, 73: 0 1--10, 2019

    2019

  2. [10]

    Hand and Robert J

    David J. Hand and Robert J. Till. A simple generalisation of the area under the roc curve for multiple class classification problems. Machine Learning, 45: 0 171--186, 2001

    2001

  3. [11]

    Vision transformers for the prediction of mild cognitive impairment to alzheimer’s disease progression using mid-sagittal smri

    Gia-Minh Hoang, Ue-Hwan Kim, and Jae Gwan Kim. Vision transformers for the prediction of mild cognitive impairment to alzheimer’s disease progression using mid-sagittal smri. Frontiers in Aging Neuroscience, 15, 2023. doi:https://doi.org/10.3389/fnagi.2023.1102869

    2023

  4. [12]

    Long short-term memory

    Sepp Hochreiter and J \"u rgen Schmidhuber. Long short-term memory. Neural computation, 9 0 (8): 0 1735--1780, 1997

    1997

  5. [13]

    Jack, Matt A

    Clifford R. Jack, Matt A. Bernstein, Nick C Fox, Paul M. Thompson, Gene E. Alexander, Danielle J. Harvey, Bret J. Borowski, Paula J. Britson, Jennifer L Whitwell, Chadwick P. Ward, Anders M. Dale, Joel P. Felmlee, Jeffrey L. Gunter, Derek L. G. Hill, Ronald J. Killiany, Norber...

    2008

  6. [14]

    Early diagnosis of alzheimer's disease based on selective kernel network with spatial attention

    Huanhuan Ji, Zhenbing Liu, Wei Qi Yan, and Reinhard Klette. Early diagnosis of alzheimer's disease based on selective kernel network with spatial attention. In Asian Conference on Pattern Recognition, 2019

    2019

  7. [15]

    Lin, and Dinggang Shen

    Yang Li, Yaping Wang, Guorong Wu, Feng Shi, Luping Zhou, W. Lin, and Dinggang Shen. Discriminant analysis of longitudinal cortical thickness changes in alzheimer's disease using dynamic and network features. Neurobiology of Aging, 33: 0 427.e15--427.e30, 2012

    2012

  8. [16]

    Dementia prevention, intervention, and care: 2020 report of the lancet commission

    Gill Livingston, Jonathan Huntley, Andrew Sommerlad, David Ames, Clive Ballard, Sube Banerjee, Carol Brayne, Alistair Burns, Jiska Cohen-Mansfield, Claudia Cooper, Sergi G Costafreda, Amit Dias, Nick Fox, Laura N Gitlin, Robert Howard, Helen C Kales, Mika Kivimäki, Eric B Lars...

    2020

  9. [17]

    Marinescu, Neil P

    Razvan V. Marinescu, Neil P. Oxtoby, Alexandra L. Young, Esther E. Bron, Arthur W. Toga, Michael W. Weiner, Frederik Barkhof, Nick C. Fox, Stefan Klein, Daniel C. Alexander, and the EuroPOND Consortium. Tadpole challenge: Prediction of longitudinal evolution in alzheimer's dis...

    2018 arXiv

  10. [18]

    Mueller, Michael W

    Susanne G. Mueller, Michael W. Weiner, Leon Thal, Ronald C. Petersen, Clifford R. Jack, William J. Jagust, John Q. Trojanowski, Arthur W. Toga, and Laurel A Beckett. Ways toward an early diagnosis in alzheimer’s disease: The alzheimer’s disease neuroimaging initiative (adni). ...

    2005

  11. [19]

    Alexander, Jiashi Feng, and B

    Minh-Quan Nguyen, Tong He, Lijun An, Daniel C. Alexander, Jiashi Feng, and B. T. Thomas Yeo. Predicting alzheimer’s disease progression using deep recurrent neural networks. NeuroImage, 222: 0 117203 -- 117203, 2019

    2019

  12. [20]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Kopf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu ...

    2019

  13. [21]

    Khojaste Sarakhsi, Seyedhamidreza Shahabi Haghighi, Seyyed M

    M. Khojaste Sarakhsi, Seyedhamidreza Shahabi Haghighi, Seyyed M. T. Fatemi Ghomi, and Elena Marchiori. Deep learning for alzheimer's disease diagnosis: A survey. Artificial intelligence in medicine, 130: 0 102332, 2022. doi:https://doi.org/10.1016/j.artmed.2022.102332

    2022

  14. [22]

    Alzheimer's disease

    Philip Scheltens, Kaj Blennow, Monique M B Breteler, Bart de Strooper , Giovanni B Frisoni, Stephen Salloway, and Wiesje Maria Van der Flier . Alzheimer's disease. The Lancet, 388 0 (10043): 0 505--517, 2016. ISSN 0140-6736. doi:https://doi.org/10.1016/S0140-6736(15)01124-1. U...

    2016 doi

  15. [23]

    Spasov, Luca Passamonti, Andrea Duggento, Pietro Lio’, and Nicola Toschi

    Simeon E. Spasov, Luca Passamonti, Andrea Duggento, Pietro Lio’, and Nicola Toschi. A parameter-efficient deep learning approach to predict conversion from mild cognitive impairment to alzheimer's disease. NeuroImage, 189: 0 276--287, 2018

    2018

  16. [24]

    Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N

    Ashish Vaswani, Noam M. Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Neural Information Processing Systems, 2017

    2017

  17. [25]

    A transformer-based framework for multivariate time series representation learning

    George Zerveas, Srideepika Jayaraman, Dhaval Patel, Anuradha Bhamidipaty, and Carsten Eickhoff. A transformer-based framework for multivariate time series representation learning. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, KDD '21, p...

    2021

  18. [26]

    Predicting future clinical changes of mci patients using longitudinal and multimodal biomarkers

    Daoqiang Zhang and Dinggang Shen. Predicting future clinical changes of mci patients using longitudinal and multimodal biomarkers. PLoS ONE, 7, 2012

    2012

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.