Classification of Subcortical Vascular Cognitive Impairment Using Single MRI Sequence and Deep Learning Convolutional Neural Networks
Open accessAbstract
Deep learning has great potential for imaging classification by extracting low to high-level features. Our aim was to train a convolutional neural network (CNN) with single T2-weighted FLAIR sequence to classify different cognitive performances in patients with subcortical ischemic vascular disease (SIVD). In total, 217 patients with SIVD [including 52 with vascular dementia (VaD), 82 with vascular mild cognitive impairment (VaMCI), and 83 with non-cognitive impairment (NCI)] and 46 matched healthy controls (HCs) underwent MRI scans and neuropsychological assessments. 2D and 3D CNNs were trained to classify VaD, VaMCI, NCI, and HCs based on FLAIR data. For 3D-based model, the loss curves of the training set approached 0.017 after about 20 epochs, while the curves of the testing set maintained at about 0.114. The accuracy of training set and testing set reached 99.7 and 96.9% after about 30 and 35 epochs, respectively. However, the accuracy of the 2D-based model was only around 70%, which performed significantly worse than 3D-based model. This experiment suggests that deep learning is a powerful and convenient method to classify different cognitive performances in SIVD by extracting the shift and scale invariant features of neuroimaging data with single FLAIR sequence. 3D-CNN is superior to 2D-CNN which involves clinical evaluation with MRI multiplanar reformation or volume scanning.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Numerous groups have taken on the challenge of nADD diagnosis using deep learning, but even these tasks were constructed as simple binary classifications between disease subtypes.
“Though numerous groups have taken on the challenge of nADD diagnosis using deep learning18–20,25,26, even these tasks were constructed as simple binary classifications between disease subtypes.”
From Multimodal deep learning for Alzheimer's disease dementia assessment (2022) · cites this paper as [25]
The Introduction says most earlier SVCI studies used binary classification, but this paper explicitly presents its own task as multiclass. The Materials and Methods (Statistical Analysis; no page number in our copy) says its deep-learning model distinguishes four categories, not two. (LLM, conf=0.90)
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredCNN classifiers using T2-FLAIR can distinguish VaD, vascular MCI, and nonimpaired cohorts in subcortical ischemic vascular disease with 96.9% accuracy.
“CNN classifiers using T2-FLAIR can distinguish VaD, vascular MCI, and nonimpaired cohorts in subcortical ischemic vascular disease with 96.9% accuracy”
From Multimodal neuroimaging and AI integration in cognitive disorders: advances, challenges, and future directions for precision medicine (2026) · cites this paper as Yao Wang et al., 2019
The Abstract and Results > Loss and Accuracy report 96.9% testing accuracy using a single T2-FLAIR sequence, but that result is for the 3D CNN classifying four groups: VaD, VaMCI, non-cognitively impaired SIVD patients, and healthy controls. The cited result is not a three-group accuracy, and the 2D CNN achieved only about 70%. (LLM, conf=0.88)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Gorelick, P. B., Scuteri, A., Black, S. E., DeCarli, C., Greenberg, S. M., Iadecola, C., Launer, L. J., Laurent, S., Lopez, O. L., Nyenhuis, D., et al. (2011) · Stroke · doi:10.1161/str.0b013e3182299496 · Cited by 3,811 · cited 3× · see citing works
“According to the AHA Statement on Vascular Contributions to Cognitive Impairment and Dementia (Gorelick et al., 2011), VaD diagnosis was based on a decline in cognitive function from a prior baseline and a deficit in performance in ≥ 2 cognitive domains that were of sufficient severity to affect the subject’s ability to p…”
Pantoni, L. (2010) · The Lancet Neurology · doi:10.1016/s1474-4422(10)70104-6 · Cited by 3,405 · cited 1× · see citing works
“These MRI signs have been recognized as the reflection of major pathologies underlying vascular dementia (VaD) and important causes of age-related cognitive decline (Pantoni, 2010; Roh and Lee, 2014).”
BackgroundCole, J. H., Poudel, R. P. K., Tsagkrasoulis, D., Caan, M. W. A., Steves, C., Spector, T. D., & Montana, G. (2017) · NeuroImage · doi:10.1016/j.neuroimage.2017.07.059 · Cited by 872 · see citing works
Diagnostic Criteria for Vascular Cognitive Disorders: A VASCOG Statement
Sachdev, P., Kalaria, R., O’Brien, J., Skoog, I., Alladi, S., Black, S. E., Blacker, D., Blazer, D. G., Chen, C., Chui, H., et al. (2014) · Alzheimer Disease & Associated Disorders · doi:10.1097/wad.0000000000000034 · Cited by 852 · cited 1× · see citing works
“Recent criteria from the International Society for Vascular Behavior and Cognitive Disorders proposed that one large infarct or hemorrhage, a strategic infarct or hemorrhage, multiple lacunes, or extensive and confluent WMH of vascular origin may be sufficient (Sachdev et al., 2014).”
BackgroundAutomatic Detection of Cerebral Microbleeds From MR Images via 3D Convolutional Neural Networks
Dou, Q., Chen, H., Yu, L., Zhao, L., Qin, J., Wang, D., Mok, V. C., Shi, L., & Heng, P.-A. (2016) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2016.2528129 · Cited by 679 · cited 2× · see citing works
“Deep learning models showed excellent performances in organ and substructure segmentation, disease detection and classification in some areas of pathology, brain, lung, abdomen, cardiac, breast, bone, retina, etc (Ji et al., 2013; Dou et al., 2016; Hamidian et al., 2017; Lao et al., 2017).”
BackgroundMethodA deep learning-based radiomics model for prediction of survival in glioblastoma multiforme
Lao, J., Chen, Y., Li, Z.-C., Li, Q., Zhang, J., Liu, J., & Zhai, G. (2017) · Scientific Reports · doi:10.1038/s41598-017-10649-8 · Cited by 630 · see citing works
Self-citationPredicting Alzheimer's disease: a neuroimaging study with 3D convolutional neural networksPreprint
Payan, A., & Montana, G. (2015) · International Conference on Pattern Recognition Applications and Methods · doi:10.48550/arxiv.1502.02506 · Cited by 546 · cited 1× · see citing works
“Payan and Montana (2015) trained sparse autoencoders and 3D CNN models for AD diagnosis.”
BackgroundIslam, J., & Zhang, Y. (2018) · Brain Informatics · doi:10.1186/s40708-018-0080-3 · Cited by 491 · see citing works
Deep-Learning Convolutional Neural Networks Accurately Classify Genetic Mutations in Gliomas
Chang, P. D., Grinband, J., Weinberg, B. D., Bardis, M. D., Khy, M., Cadena, G., Su, M., Cha, S., Filippi, C. G., Bota, D. A., et al. (2018) · American Journal of Neuroradiology · doi:10.3174/ajnr.a5667 · Cited by 466 · cited 1× · see citing works
“With the rapid progress of this field, convolutional neural network (CNN) has been utilized in many medical fields, such as in the classification of Alzheimer’s brains and healthy brains and the prediction of underlying molecular genetic mutation status in gliomas (Sarraf and Tofighi, 2016; Chang et al., 2018), which…”
MethodSarraf, S., & Tofighi, G. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1603.08631 · Cited by 204 · see citing works
Hachinski, V. C., Bowler, J. V., & Loeb, C. (1993) · Neurology · doi:10.1212/wnl.43.10.2159-a · Cited by 171 · see citing works
Ng, S., Allen, W., Brust, J., & Susser, M. (1993) · Neurology · doi:10.1212/wnl.43.10.2153 · Cited by 162 · see citing works
Sun, Y.-W., Qin, L.-D., Zhou, Y., Xu, Q., Qian, L.-J., Tao, J., & Xu, J.-R. (2011) · Behavioural Brain Research · doi:10.1016/j.bbr.2011.05.006 · Cited by 117 · see citing works
Self-citationDistinctive Clinical Features of Mild Cognitive Impairment with Subcortical Cerebrovascular Disease
Galluzzi, S., Sheu, C.-F., Zanetti, O., & Frisoni, G. (2005) · Dementia and Geriatric Cognitive Disorders · doi:10.1159/000083499 · Cited by 103 · cited 1× · see citing works
“SIVD can be defined as subcortical WM hyperintensity on T2-weighted images with at least one lacunar infarct, in accordance with the criteria suggested by Galluzzi et al. (2005).”
ResultRecent Updates on Subcortical Ischemic Vascular Dementia
Roh, J. H., & Lee, J.-H. (2014) · Journal of Stroke · doi:10.5853/jos.2014.16.1.18 · Cited by 93 · cited 1× · see citing works
“These MRI signs have been recognized as the reflection of major pathologies underlying vascular dementia (VaD) and important causes of age-related cognitive decline (Pantoni, 2010; Roh and Lee, 2014).”
BackgroundNew insights into cerebral small vessel disease and vascular cognitive impairment from MRI
Smith, E. E., & Beaudin, A. E. (2017) · Current Opinion in Neurology · doi:10.1097/wco.0000000000000513 · Cited by 90 · cited 1× · see citing works
“So finding an efficient way to perform early diagnosis of VaD is essential for disease prevention and modifying therapies (Smith and Beaudin, 2018).”
Method3D Convolutional Neural Network for Automatic Detection of Lung Nodules in Chest CT
Hamidian, S., Sahiner, B., Petrick, N., & Pezeshk, A. (2017) · Medical Imaging · doi:10.1117/12.2255795 · Cited by 79 · cited 2× · see citing works
“Deep learning technologies have been proved to be efficient in various medical image analysis tasks such as MRI, CT, X-ray, mammography, ultrasound, and microscopy (Hamidian et al., 2017; Jyoti and Yanqing, 2018).”
BackgroundLee, A. Y. (2011) · Chonnam Medical Journal · doi:10.4068/cmj.2011.47.2.66 · Cited by 58 · cited 1× · see citing works
“Vascular cognitive impairment (VCI) is a broad term including mild to severe cognitive impairment associated with or caused by cerebrovascular disease (Hachinski and Bowler, 1993).”
BackgroundIngles, J., Boulton, D. C., Fisk, J., & Rockwood, K. (2007) · Stroke · doi:10.1161/01.str.0000259716.04739.60 · Cited by 48 · cited 1× · see citing works
“Previous research emphasized that cognitive impairment may be the earliest, most common and sensitive manifestation of cerebrovascular disease, and proposed it as a prognostic indicator (Ingles et al., 2007).”
BackgroundCiulli, S., Citi, L., Salvadori, E., Valenti, R., Poggesi, A., Inzitari, D., Mascalchi, M., Toschi, N., Pantoni, L., & Diciotti, S. (2016) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2016.2537808 · Cited by 38 · cited 1× · see citing works
“Another study implemented SVM-based machine learning strategy for discrimination between SVCI patients with different cognitive performances through predefined feature vectors extracted from diffusion tensor imaging (DTI) data alone (Ciulli et al., 2016).”
MethodWang, Y., Cao, W., Sun, Y.-W., Chen, X., Ding, W.-N., Xu, Q., Zhou, Y., Xu, J.-R., & Suo, S. (2017) · Current Alzheimer Research · doi:10.2174/1567205014666170329113707 · Cited by 13
Self-citationMultimodal MRI classification in vascular mild cognitive impairment
Diciotti, S., Ciulli, S., Ginestroni, A., Salvadori, E., Poggesi, A., Pantoni, L., Inzitari, D., Mascalchi, M., & Toschi, N. (2015) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society · doi:10.1109/embc.2015.7319340 · Cited by 12 · cited 2× · see citing works
“Previous studies made efforts to address this challenging task by traditional machine learning using multimodal MRI data, including diffusion and morphometry features, and eventually proved the usefulness of machine learning techniques in discriminating HCs from patients suffering from VaD and its prodromal stage: vas…”
Background
Flagged first, then by citations
2 of the 24 references couldn’t be matched to a known paper.
Research works printed but not linked: 1
- Ji S. Xu W. Yang M. Yu K. (2013). 3d convolutional neural networks for human action recognition. IEEE Trans. Pattern Anal. Mach. Intell. 35 221– 231. 10.1109/tpami.2012.59 22392705 · DOI
Cited by
Sorted by citations
Multimodal deep learning for Alzheimer's disease dementia assessment
Qiu, S., Miller, M. I., Joshi, P. S., Lee, J. C., Xue, C., Ni, Y., Wang, Y., De Anda-Duran, I., Hwang, P. H., Cramer, J. A., et al. (2022) · Nature Communications · doi:10.1038/s41467-022-31037-5 · Cited by 420 · see citing works
Self-citationArtificial intelligence in medicine: Technical basis and clinical applications
Erickson, B. J. (2020) · Elsevier eBooks · doi:10.1016/b978-0-12-821259-2.00002-8 · Cited by 71 · see citing works
Liu, M., Wang, Y., Zhang, H., Yang, Q., Shi, F., Zhou, Y. F., & Shen, D. G. (2022) · Cerebral Cortex · doi:10.1093/cercor/bhab507 · Cited by 25 · see citing works
Self-citationGoncharov, N. V., Popova, P. I., Kudryavtsev, I. V., Golovkin, A., Savitskaya, I., Avdonin, P. P., Korf, E., Voitenko, N., Belinskaia, D., Serebryakova, M., et al. (2024) · International Journal of Molecular Sciences · doi:10.3390/ijms25031888 · Cited by 22 · see citing works
Sun, K., Zhang, Y.-W., Liu, J., Yu, L., Zhou, Y., Xie, F., Guo, Q.-H., Zhang, H., Wang, Q., & Shen, D.-G. (2024) · Communications Engineer · doi:10.1038/s44172-024-00245-w · Cited by 21 · see citing works
Self-citationSerotonergic Regulation in Alzheimer’s Disease
Долгачева, Л. П., Зинченко, В. П., Nadeev, A. D., & Goncharov, N. V. (2025) · International Journal of Molecular Sciences · doi:10.3390/ijms26115218 · Cited by 21 · see citing works
“The 5-HT 1 R and 5-HT 5 R subtypes couple with the G i/o protein, leading to the inhibition of adenylyl cyclase (AC) activity and a reduction in cyclic adenosine monophosphate (cAMP) levels [136].”
Clinically applicable deep learning-based decision aids for treatment of neovascular AMD
Gutfleisch, M., Ester, O., Aydin, S., Quassowski, M., Spital, G., Lommatzsch, A., Rothaus, K., Dubis, A., & Pauleikhoff, D. (2022) · Graefe's Archive for Clinical and Experimental Ophthalmology · doi:10.1007/s00417-022-05565-1 · Cited by 16 · see citing works
“They have been established in subcortical vascular cognitive impairment [10] and glaucoma [11].”
BackgroundAslan, N., Sengul, M. K., & Koca, G. (2023) · Engineering applications of artificial intelligence · doi:10.1016/j.engappai.2023.106794 · Cited by 15
Alzheimer's and vascular disease classification using regional texture biomarkers in FLAIR MRI.
Chan, K., Fischer, C. E., Maralani, P., Black, S., Moody, A., & Khademi, A. (2023) · NeuroImage: Clinical · doi:10.1016/j.nicl.2023.103385 · Cited by 14 · see citing works
“3D CNN [14] FLAIR FLAIR volumes SIVD subtypes classification 96.”
BackgroundMethodInfluentialChen, Q., Wang, Y., Qiu, Y., Wu, X., Zhou, Y., & Zhai, G. (2020) · Frontiers in Neuroscience · doi:10.3389/fnins.2020.00557 · Cited by 16 · see citing works
Self-citationDeep learning applications in vascular dementia using neuroimaging
Dong, C., & Hayashi, S. (2024) · Current Opinion in Psychiatry · doi:10.1097/yco.0000000000000920 · Cited by 13 · see citing works
Huang, L.-L., Li, Z., Zhu, X.-L., Zhao, H., Mao, C.-L., Ke, Z.-H., Mo, Y.-T., Yang, D., Cheng, Y., Qin, R.-M., et al. (2025) · npj Digital Medicine · doi:10.1038/s41746-025-01813-w · Cited by 13 · see citing works
Liu, B., Meng, S., Cheng, J., Zeng, Y., Zhou, D., Deng, X.-J., Kuang, L.-Q., Wu, X.-J., Tang, L., Wang, H., et al. (2022) · Frontiers in Oncology · doi:10.3389/fonc.2022.852726 · Cited by 13
Zhang, L.-F., Li, Y., Bian, L., Luo, Q., Zhang, X., & Zhao, B. (2022) · Frontiers in Public Health · doi:10.3389/fpubh.2021.813641 · Cited by 9
“With the aging of current society, the incidence of cerebrovascular disease and VCI has increased significantly, ranking second only to Alzheimer’s disease (AD), and is an important cause of the cognitive decline in the elderly (3).”
BackgroundDang, M., Liu, B., Chen, Y., & Zhang, Z. (2026) · Psychoradiology · doi:10.1093/psyrad/kkag007 · Cited by 7 · see citing works
Özkan, D., KATAR, O., Ak, M., Al–antari, M. A., Ak, N. Y., Yıldırım, Ö., Mir, H., Tan, R., & Acharya, U. R. (2024) · IEEE Access · doi:10.1109/access.2024.3454709 · Cited by 7 · see citing works
“Wang et al. [96] compared the performance of 2D and 3D CNNs in categorizing individuals with subcortical ischemic vascular disease (SIVD) based on their cognitive performance using a single T2-weighted FLAIR sequence.”
MethodDeep Learning for Classifying and Cognitive Profiling of Subcortical Vascular Cognitive Impairment
He, M., Yin, Y., Qu, J., Wang, Y., Que, X., Xia, X., Zhang, T., Li, J., Shen, J., Song, W., et al. (2026) · Cyborg and Bionic Systems · doi:10.34133/cbsystems.0561 · Cited by 1 · see citing works
“Existing studies relied on T1-weighted or T2-FLAIR images for SVCI identification [ 38 – 40 ].”
Self-citationEndothelium, Aging, and Vascular Diseases
Goncharov, N. V., Popova, P. I., Nadeev, D., Belinskaia, D., Korf, E., & Avdonin, P. (2024) · Journal of Evolutionary Biochemistry and Physiology · doi:10.1134/s0022093024060048 · Cited by 1 · see citing works
“Damaged ECs can cause persistent blood flow dysregulation, lead(cid:2) ing to the development of subcortical ischemic vas(cid:2) cular disease (SIVD) [188].”
Jiang, X., Wang, G.-Y., Zhang, L.-M., Xi, X., Leone, R. D., & Qiao, L. (2024) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2024.106766 · Cited by 1 · see citing works
Cross-modal structural residuals reveal cognitive impairment in cerebral small vessel disease
He, M., Liu, Q.-Y., Que, X.-W., Xia, X.-Y., Li, J., He, M.-J., Liu, Y.-C., Qin, Q., Li, C., & Tang, Y. (2026) · Journal of Physics, Conference Series · doi:10.1088/1742-6596/3291/1/012013
Chen, X., Fu, P., Luo, X., Kang, C., Huang, G., Wu, Y., Zhang, M., Yin, X., & Chen, Z. (2026) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-026-10668-7
“Wang et al. [96] employed a 3D-CNN for preprocessing and model construction using T2-FLAIR images of SID, VCIND, non-cognitive impairment (NCI), and 46 NC cases.”
Wu, Q., Zhang, J., Peng, L., Zhang, Z., Zhu, X., & Huang, C. (2025) · Neuroscience · doi:10.1016/j.neuroscience.2025.09.021 · Cited by 1 · see citing works
Liu, M.-X., Zhang, J., Wang, Y., Zhou, Y., Xie, F., Guo, Q.-H., Shi, F., Zhang, H., Wang, Q., & Shen, D. (2023) · arXiv.org · doi:10.48550/arxiv.2302.11871
Self-citationNeelaveni, J., Devasena, M. S. G., & Gopu, G. (2021) · Elsevier eBooks · doi:10.1016/b978-0-12-822271-3.00002-5
- Authors
- Wang, Y., Tu, D., Du, J., Han, X., Sun, Y., Xu, Q., Zhai, G., & Zhou, Y.
- Affiliations
- Shanghai Jiao Tong University (CN)
- Corresponding author
- Zhai, Guangtao
- Zhou, Yan
- Venue
- Frontiers in Neuroscience
- Publisher
- Frontiers Media SA
- Published
- 2019-06-19
- Volume
- 13
- Pages
- 627-627
- ISSN
- 1662-453X (ISSN-L)
- PMC
- PMC6593093
- OpenAlex
- W2952091205
- Cited by
- 25
- References
- 24
- Citations by year
- 2026: 3 · 2025: 3 · 2024: 6 · 2023: 5 · 2022: 5 · 2021: 1 · 2020: 2
- FWCI
- 1.59
- 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)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- Shanghai Municipal Education Commission (20172013)
- Science and Technology Commission of Shanghai Municipality (17411964300)
- Shanghai Jiao Tong University
- Renji Hospital
- Topics
- Health Sciences › Medicine › Psychiatry and Mental health › Dementia and Cognitive Impairment Research
- Life Sciences › Neuroscience › Neurology › Neurological Disease Mechanisms and Treatments
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Retinal Imaging and Analysis
- Keywords
- subcortical ischemic vascular disease, vascular cognitive impairment, vascular dementia, vascular mild cognitive impairment, FLAIR MRI, convolutional neural networks, 3D convolutional neural networks, 2D convolutional neural network, neuroimaging classification
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.
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
