Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI
Open accessThe DOI matches its Crossref record; 70 of the 71 references it cites were found in a registry or an index; 19 of the 26 sources its claims cite were checked for support.
Abstract
Background: Frontotemporal dementia (FTD) represents a collection of neurobehavioral and neurocognitive syndromes that are associated with a significant degree of clinical, pathological, and genetic heterogeneity. Such heterogeneity hinders the identification of effective biomarkers, preventing effective targeted recruitment of participants in clinical trials for developing potential interventions and treatments. In the present study, we aim to automatically differentiate patients with three clinical phenotypes of FTD, behavioral-variant FTD (bvFTD), semantic variant PPA (svPPA), and nonfluent variant PPA (nfvPPA), based on their structural MRI by training a deep neural network (DNN). Methods: Data from 277 FTD patients (173 bvFTD, 63 nfvPPA, and 41 svPPA) recruited from two multi-site neuroimaging datasets: the Frontotemporal Lobar Degeneration Neuroimaging Initiative and the ARTFL-LEFFTDS Longitudinal Frontotemporal Lobar Degeneration databases. Raw T1-weighted MRI data were preprocessed and parcellated into patch-based ROIs, with cortical thickness and volume features extracted and harmonized to control the confounding effects of sex, age, total intracranial volume, cohort, and scanner difference. A multi-type parallel feature embedding framework was trained to classify three FTD subtypes with a weighted cross-entropy loss function used to account for unbalanced sample sizes. Feature visualization was achieved through post-hoc analysis using an integrated gradient approach. Results: The proposed differential diagnosis framework achieved a mean balanced accuracy of 0.80 for bvFTD, 0.82 for nfvPPA, 0.89 for svPPA, and an overall balanced accuracy of 0.84. Feature importance maps showed more localized differential patterns among different FTD subtypes compared to groupwise statistical mapping. Conclusion: In this study, we demonstrated the efficiency and effectiveness of using explainable deep-learning-based parallel feature embedding and visualization framework on MRI-derived multi-type structural patterns to differentiate three clinically defined subphenotypes of FTD: bvFTD, nfvPPA, and svPPA, which could help with the identification of at-risk populations for early and precise diagnosis for intervention planning.
Overview
CoChat Summary
## TL;DR
This study trained an explainable deep neural network on structural T1-weighted MRI to differentiate three clinical frontotemporal dementia subtypes: bvFTD, nfvPPA, and svPPA. Using 277 patients from two multi-site cohorts, the model achieved overall balanced accuracy of 0.84 and produced feature-importance maps that aligned with known subtype-specific atrophy patterns.
## Key findings
- The proposed framework achieved mean balanced accuracy of 0.80 for bvFTD, 0.82 for nfvPPA, 0.89 for svPPA, and 0.84 overall.
- In the more detailed results, the best model achieved 79.7% balanced accuracy for bvFTD, 81.9% for nfvPPA, 89.2% for svPPA, and 83.6% overall.
- In the ablation study, thickness-only features outperformed volume-only features for bvFTD, svPPA, and overall performance; simple concatenation of thickness and volume reduced performance relative to thickness-only, while the proposed multi-level parallel embedding improved performance for bvFTD, nfvPPA, and overall.
- Integrated Gradient maps showed subtype-relevant regions: left temporal patches for nfvPPA and svPPA, inferior frontal/frontal operculum regions for nfvPPA, and left anterior temporal, frontal opercular/insular, and bilateral frontal pole regions for bvFTD.
## Methods
- Design: Retrospective multi-class classification study using an explainable deep-learning pipeline for differential diagnosis of FTD subtypes from structural MRI.
- Sample/data: 277 FTD patients total: 173 bvFTD, 63 nfvPPA, 41 svPPA, drawn from the ALLFTD and FTLDNI/NIFD multi-site cohorts; cognitively normal controls were excluded because only 27 were available.
- Main analysis: Raw T1 MRI were parcellated into 360 cortical patches plus 15 subcortical structures for 735 features total; features were harmonized with a regression-based w-score approach controlling for cohort, scanner/coil, sex, age, and TIV, then classified with a two-level MLP-based parallel feature embedding DNN evaluated by 10-fold nested cross-validation and interpreted with Integrated Gradients.
## Limitations
- The paper states that it did not include cognitively normal controls because the sample was too small; it therefore used bvFTD as the reference group for harmonization, which the authors note means the feature-importance maps mainly reflect differences among the three FTD subtypes rather than differences versus healthy aging.
- The authors state that they treated demographic variables as confounders, but future work could incorporate demographic information directly into the model; this indicates the current framework may miss useful demographic signal.
- Evident from the text: the study used only structural T1-weighted MRI and only three clinical subtypes, so generalization to other modalities or broader dementia groups was not tested in this paper.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures






Verification
Claims the paper makes
14 Valid · 0 Not valid · 5 Not checked · 32 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“svPPA is commonly associated with striking asymmetric atrophy of the dominant hemisphere temporal pole (Rogalski et al., 2011).”
§ Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI > Discussion
What the sources say
- Supportedcites Rogalski et al., 2011 source
The PPA-S results identify peak atrophy in the anterior temporal lobe, including the temporal pole, and say it was greater in the left hemisphere than the right. The paper identifies PPA-S as the semantic variant, so this supports the cited description of asymmetric temporal-pole atrophy in svPPA. (LLM, conf=0.92)
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“This distinctive atrophy pattern is usually due to the presence of TDP-43 Type C neuropathology in these regions (Kawles et al., 2022; Keszycki et al., 2022).”
§ Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI > Discussion
What the sources say
- Partly supportedcites Kawles et al., 2022 sourceExperimental — accuracy not yet measured
The Introduction states that 'TDP-type C is commonly associated with severe anterior temporal atrophy, invariably resulting in poor word comprehension as the principal feature in the semantic variant of the PPA syndrome', and the Abstract/Figure 1 confirm that PPA cases showed leftward asymmetric cortical atrophy with 'selective and severe targeting of the temporal poles' — so the clinicopathological link between TDP-43 type C and the distinctive svPPA atrophy pattern is in our copy. However, the paper frames this as an association ('putative correlates of neurodegeneration') and its central finding runs counter to the causal wording of the claim: in the regions with predilection for severe atrophy the relationship was inverse — 'the anterior temporal lobe displayed profound neuronal loss but sparse TDP inclusions', with inclusions thought to wane/disappear as their neurons are lost (Discussion; Inverse relationship section; Abstract). The part the source does not support is the assertion that the atrophy in these regions is 'due to the presence' of TDP-43 type C inclusions there; at autopsy the most atrophic regions had the lowest inclusion density. (LLM, conf=0.75)
- Partly supportedcites Keszycki et al., 2022 sourceExperimental — accuracy not yet measured
The ‘Clinical syndromes associated with TDP-43 proteinopathy’ section (page unavailable) says type C is the most common cause of semantic PPA, supporting the usual neuropathological association. The paper does not identify the distinctive atrophy pattern or establish that type C pathology in the specified regions causes that pattern. (LLM, conf=0.84)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Indeed, bvFTD can be due to underlying FTLD-Tau, FTLD-TDP, or less commonly, AD neuropathology (Peet et al., 2021).”
§ Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI > Discussion
What the sources say
- Supportedcites Peet et al., 2021 source
The FTLD Neuropathology and Genetics section identifies tau and TDP-43 pathology among the substrates of bvFTD and explicitly includes Alzheimer’s disease. It says Alzheimer’s disease accounts for a small percentage of cases in the core FTD syndromes, supporting “less commonly.” (LLM, conf=0.94)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Additionally, our classification of interest was clinical diagnosis, as clinical syndromes are known to correspond more closely to neuroanatomical lesions as compared to neuropathology (Seeley et al., 2009).”
§ Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI > Discussion > Limitations and future directions
What the sources say
- Supportedcites Seeley et al., 2009 source
The Results describe distinct anatomical atrophy patterns for five clinical syndromes. The Experimental Procedures (Subjects) distinguish clinical syndromes from histopathological entities and state that most degenerative syndromes can arise from several underlying pathologies. Together, these passages support the cited comparison, although the paper does not quantify how much closer the correspondence is. (LLM, conf=0.89)
boostedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“the significant degree of clinical, pathological and genetic heterogeneity observed in FTD hinders the development of sensitive and specific biomarkers that would allow for targeted recruitment of groups at highest risk for clinical/cognitive decline (Katzeff et al., 2022).”
§ Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI > Introduction
What the sources say
- Partly supportedcites Katzeff et al., 2022 sourceExperimental — accuracy not yet measured
Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
22 more claims
Create a free account to see all 51 claims and to verify the 2 references still open.
- Probably not a substantial claim. 12
- It may state two separate claims as one. 14
- The quote only partly states this claim. 1
Citations
Cites
A Unified Approach to Interpreting Model Predictions
Lundberg, S., & Lee, S.-I. (2017) · arXiv · doi:10.48550/arxiv.1705.07874 · Cited by 23,707
Deep learning in neural networks: an overview
Schmidhuber, J. (2015) · Neural Networks · doi:10.1016/j.neunet.2014.09.003 · Cited by 18,570 · cited 1×
“DNN classifiers have shown superior performance compared to traditional machine learning methods, such as support vector machines (SVM) and random forest for accurate classification of disease groups using neuroimaging data (Schmidhuber, 2015; Eslami and Saeed, 2019; Amini et al., 2021).”
MethodFischl, B. (2012) · NeuroImage · doi:10.1016/j.neuroimage.2012.01.021 · Cited by 9,526 · cited 1×
“The manifold of cerebral cortical surface data was first derived through brain tissue segmentation (gray matter, white matter, and cerebral spinal fluid – CSF), followed by cortical surface reconstruction using FreeSurfer 5.3 (Fischl, 2012).”
MethodInfluentialSensitivity of revised diagnostic criteria for the behavioural variant of frontotemporal dementia
Rascovsky, K., Hodges, J. R., Knopman, D., Mendez, M. F., Kramer, J. H., Neuhaus, J., van Swieten, J. C., Seelaar, H., Dopper, E. G. P., Onyike, C. U., et al. (2011) · Brain · doi:10.1093/brain/awr179 · Cited by 5,221
Self-citationClassification of primary progressive aphasia and its variants
Gorno-Tempini, M. L., Hillis, A. E., Weintraub, S., Kertesz, A., Mendez, M., Cappa, S. F., Ogar, J. M., Rohrer, J. D., Black, S., Boeve, B. F., et al. (2011) · Neurology · doi:10.1212/wnl.0b013e31821103e6 · Cited by 5,141
A multi-modal parcellation of human cerebral cortex
Glasser, M. F., Coalson, T. S., Robinson, E. C., Hacker, C. D., Harwell, J., Yacoub, E., Ugurbil, K., Andersson, J., Beckmann, C. F., Jenkinson, M., et al. (2016) · Nature · doi:10.1038/nature18933 · Cited by 5,111
Axiomatic attribution for deep networks
Sundararajan, M., Taly, A., & Yan, Q. (2017) · arXiv · doi:10.48550/arxiv.1703.01365 · Cited by 4,087 · cited 7×
“The Integrated Gradient based FTD subtype feature attribute visualization patterns are shown in Figure 5 for both cortical thickness and volume features.”
MethodInfluentialAppelhof, B., Bakker, C., Van Duinen-van den Ijssel, J. C. L., Zwijsen, S. A., Smalbrugge, M., Verhey, F. R. J., de Vugt, M. E., Zuidema, S. U., & Koopmans, R. T. C. M. (2017) · Dementia and Geriatric Cognitive Disorders · doi:10.1159/000477087 · Cited by 2,760
Neurodegenerative Diseases Target Large-Scale Human Brain Networks
Seeley, W. W., Crawford, R. K., Zhou, J., Miller, B. L., & Greicius, M. D. (2009) · Neuron · doi:10.1016/j.neuron.2009.03.024 · Cited by 2,350
Qualitative Assessment of Verbal Fluency Performance in Frontotemporal Dementia
van den Berg, E., Jiskoot, L. C., Grosveld, M. J. H., van Swieten, J. C., & Papma, J. M. (2017) · Dementia and Geriatric Cognitive Disorders · doi:10.1159/000477538 · Cited by 2,046
Cited by
Wang, H., Toumaj, S., Heidari, A., Souri, A., Jafari, N., & Jiang, Y. (2025) · Engineering applications of artificial intelligence · doi:10.1016/j.engappai.2025.110752 · Cited by 33
Dattola, S., Ielo, A., Varone, G., Cacciola, A., Quartarone, A., & Bonanno, L. (2025) · Frontiers in Aging Neuroscience · doi:10.3389/fnagi.2025.1547727 · Cited by 17
Poonam, K., Kotra, V. S., Guha, R., & Chakrabarti, P. P. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-78195-7_26 · Cited by 3
Lifespan Tree of Brain Anatomy: Diagnostic Values for Motor and Cognitive Neurodegenerative Diseases
Coupé, P., Mansencal, B., Manjón, J. V., Péran, P., Meissner, W. G., Tourdias, T., & Planche, V. (2025) · Human Brain Mapping · doi:10.1002/hbm.70336 · Cited by 3
“…recognized that incorporating additional variables—such as other imaging modalities, clinical scores, or fluid biomark-ers—can significantly improve classification performance (Talai et al. 2021; Ma et al. 2024; Nguyen et al. 2024a; Tong et al. 2017; Koikkalainen et al. 2016; Tolonen et al. 2018).”
Structural MRI in frontotemporal dementia and Alzheimer’s disease: stage-dependent atrophy patterns
Khalili, E., Sodaei, F., Noroozıan, M., Michel, T. M., & Vafaee, M. S. (2026) · Journal of Neural Transmission · doi:10.1007/s00702-026-03158-4 · Cited by 2
“…regional neurodegenerative heterogeneous syndromes including behavioral-variant FTD and primary progressive aphasia variants (nfvPPA, svPPA) that demonstrate variable frontal and temporal network vulnerability (Ma et al. 2024; Mandelli et al. 2016; Ranasinghe et al. 2016; Ossenkoppele et al. 2015).”
Metadata
- Authors
- Ma, D., Stocks, J., Rosen, H., Kantarci, K., Lockhart, S. N., Bateman, J. R., Craft, S., Gurcan, M. N., Popuri, K., Beg, M. F., Wang, L., & on behalf of the ALLFTD consortium
- Affiliations
- Wake Forest University Health Sciences (US)
- Northwestern University (US)
- University of California, San Francisco (US)
- Mayo Clinic (US)
- Memorial University of Newfoundland (CA)
- Simon Fraser University (CA)
- Corresponding author
- Ma, Da
- Stocks, Jane
- Wang, Lei
- ORCID
- Venue
- Frontiers in Neuroscience
- Publisher
- Frontiers Media SA
- Published
- 2024-02-07
- Volume
- 18
- Pages
- 1331677-1331677
- ISSN
- 1662-453X (ISSN-L)
- PMC
- PMC10879283
- OpenAlex
- W4391609079
- Semantic Scholar
- b215fff6643db4399bc84244cad1609484d3b321
- Cited by
- 9
- References
- 71
- Citations by year
- 2026: 5 · 2025: 3 · 2024: 1
- FWCI
- 2.63
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Frontiers in Neuroscience (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- eScholarship (California Digital Library) (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- National Institute on Aging (R01AG055121, R56AG055121, T32AG020506, P30AG072977, P30AG072947)
- Topics
- Health Sciences › Medicine › Psychiatry and Mental health › Dementia and Cognitive Impairment Research
- Life Sciences › Neuroscience › Cellular and Molecular Neuroscience › Alzheimer's disease research and treatments
- Health Sciences › Medicine › Clinical Neurology › Amyotrophic Lateral Sclerosis Research
- Keywords
- frontotemporal dementia, differential diagnosis, behavioral variant FTD, semantic variant PPA, structural MRI, explainable deep learning, cortical thickness, cortical volume, integrated gradients, feature visualization, biomarkers
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI. https://app.cochat.ai/papers/10.3389/fnins.2024.1331677
- The paper
- Ma, D., Stocks, J., Rosen, H., Kantarci, K., Lockhart, S. N., Bateman, J. R., Craft, S., Gurcan, M. N., Popuri, K., Beg, M. F., Wang, L., & on behalf of the ALLFTD consortium. (2024). Differential diagnosis of frontotemporal dementia subtypes with explainable deep learning on structural MRI. Frontiers in Neuroscience. https://doi.org/10.3389/fnins.2024.1331677
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.