A Comprehensive Review on Synergy of Multi-Modal Data and AI Technologies in Medical Diagnosis
Open accessAbstract
Disease diagnosis represents a critical and arduous endeavor within the medical field. Artificial intelligence (AI) techniques, spanning from machine learning and deep learning to large model paradigms, stand poised to significantly augment physicians in rendering more evidence-based decisions, thus presenting a pioneering solution for clinical practice. Traditionally, the amalgamation of diverse medical data modalities (e.g., image, text, speech, genetic data, physiological signals) is imperative to facilitate a comprehensive disease analysis, a topic of burgeoning interest among both researchers and clinicians in recent times. Hence, there exists a pressing need to synthesize the latest strides in multi-modal data and AI technologies in the realm of medical diagnosis. In this paper, we narrow our focus to five specific disorders (Alzheimer’s disease, breast cancer, depression, heart disease, epilepsy), elucidating advanced endeavors in their diagnosis and treatment through the lens of artificial intelligence. Our survey not only delineates detailed diagnostic methodologies across varying modalities but also underscores commonly utilized public datasets, the intricacies of feature engineering, prevalent classification models, and envisaged challenges for future endeavors. In essence, our research endeavors to contribute to the advancement of diagnostic methodologies, furnishing invaluable insights for clinical decision making.
AI snapshot
Written from the paper’s full text.
TL;DR
This review surveys AI and multi-modal data for diagnosing Alzheimer’s disease, breast cancer, depression, heart disease, and epilepsy, covering datasets, preprocessing, feature engineering, models, and evaluation. It reports that deep learning, attention, transfer learning, and multi-modal fusion are improving diagnostic performance, while data imbalance, weak generalization, limited interpretability, privacy, and fairness remain key challenges.
Key findings
- Across the five diseases, modality preferences differ: Alzheimer’s disease relies mainly on MRI/PET plus speech; breast cancer on X-ray, ultrasound, MRI, text, and genetic data; depression on text, speech, and EEG; heart disease on echocardiography, ECG, text, and genetic data; epilepsy on video, MRI, text, and EEG.
- Reported performance examples include Mayo Clinic’s CNN on 12-lead ECG for hypertrophic cardiomyopathy with AUC 0.96, sensitivity 87%, and specificity 90% in adults, and AUC 0.98, sensitivity 92%, and specificity 95% in children; Dai et al.’s CNN on 12-lead ECG reached accuracy 99.84%, sensitivity 99.52%, and specificity 99.95% for 3-s signals; Jabeen et al. reported 99.1% accuracy on BUSI ultrasound images; Thara et al. reported 99.89% training and 99.08% test accuracy for stacked bidirectional LSTM epilepsy detection.
- Multi-modal fusion and advanced architectures—attention mechanisms, transfer learning, transformers, and large models—are highlighted as major trends; large models discussed include Med-PaLM, XrayGPT, and GeneCompass, and GPT-3’s heart disease diagnosis/triage performance was reported to nearly approach that of professional medical practitioners.
- The review identifies data imbalance, weak model generalization, lack of interpretability (only 28 included studies directly or indirectly addressed it), privacy/security, and ethical bias as major challenges, with proposed solutions including transfer learning, synthetic data, ensemble methods, federated/swarm learning, balanced sampling, and interpretable models/heatmaps.
Methods
- Design: comprehensive narrative review/survey of AI-based diagnosis for five disorders; it is not a primary experimental study.
- Data or sample: synthesizes public datasets and reported works, including ADNI, OASIS-3, UKB, TCGA, BUSI, GEO, DAIC-WOZ, MODMA, CHB-MIT, Bonn EEG, and Temple University EEG corpus; no new patient data were collected.
- Main analysis: organizes the AI pipeline into preprocessing, feature engineering, model selection, and performance evaluation, then summarizes modality-specific methods and results in tables; no meta-analysis or systematic search protocol is reported.
Limitations
- Stated by the paper: field-wide challenges include medical multimodality data imbalance, weak model generalization, lack of interpretability (only 28 included studies addressed it), data privacy/security, and ethical/moral bias and fairness concerns.
- Stated by the paper: some reviewed depression EEG studies have small sample sizes that limit clinical generalizability.
- Evident from the text: as a narrative review, it does not report a systematic search strategy, inclusion/exclusion criteria, or risk-of-bias assessment; the summary tables also contain missing results (“-”) and inconsistent reporting across studies.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
0 of 171 cited sources checked · 0 supported · 0 flagged
- Findingnone given
In most studies reviewed, the persistence of differences often stemmed from not having enough data to achieve mitigation.
“In most studies reviewed, the persistence of differences often stemmed from not having enough data to achieve mitigation.”
Quote in the paperNot a substantial claim - Findingown data
This review found that only 28 of the included studies directly or indirectly tackled the crucial aspect of interpretability.
“This review found that only 28 of the included studies directly or indirectly tackled the crucial aspect of interpretability.”
Quote in the paperSubstantial claim - Findingown data
The majority of studies did not adequately consider interpretability.
“However, regrettably, the majority of studies did not adequately consider this crucial factor.”
Quote in the paperSubstantial claim - Interpretationreasoning
Large models currently fall short in being meticulously designed for specific tasks and guided by medical knowledge.
“In contrast, LMs currently fall short in this aspect.”
Quote in the paperSubstantial claim - Premisecitation
Baydargil et al.'s unsupervised adversarial parallel model for AD anomaly analysis exhibits classification accuracy of 96.03% and AUC of 75.21% in distinguishing AD, MCI, and normal control groups.
“The model exhibits robust classification with rates and area under the curve (AUC) scores reaching 96.03% and 75.21%, respectively, underscoring its effective discriminative performance.”
- Also stated in: § 3.1. Diagnosis of Alzheimer’s Disease · plain
“| Baydargil et al. [71] | Original PET slices | PET | ADNI | 0.9603 (ACC: AD vs. NC vs. MCI) 0.7521 (AUC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [71] Baydargil, H.B., Park, J.S., and Kang, D.Y. (2021). Anomaly analysis of Alzheimer’s disease in PET images using an unsupervised adversarial deep learning model. Appl. Sci., 11. source
Not checked in the free run.
- Also stated in: § 3.1. Diagnosis of Alzheimer’s Disease · plain
- Interpretationreasoning
Large models have yet to reach their full potential in specific medical domain tasks.
“While these LMs have proven effective in various general domain tasks, they have yet to reach their full potential in specific medical domain tasks.”
Quote in the paperSubstantial claim - Premisecitation
Transformer is valuable for processing and modeling diverse modalities of medical data, encompassing clinical texts, medical images, and time series data.
“In the realm of medical diagnosis, A transformer proves valuable for processing and modeling diverse modalities of medical data, encompassing clinical texts, medical images, and time series data [59,60,61].”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites [59] Zhu, 2021, Dual attention multi-instance deep learning for Alzheimer’s disease diagnosis with structural MRI, IEEE Trans. Med. Imaging source
Not checked in the free run.
- cites [60] Chen, 2022, Contrastive Learning for Prediction of Alzheimer’s Disease Using Brain 18F-FDG PET, IEEE J. Biomed. Health Inform. source
Not checked in the free run.
- cites [61] Zhu, 2023, A dictionary-guided attention network for biomedical named entity recognition in Chinese electronic medical records, Expert Syst. Appl. source
Not checked in the free run.
- Interpretationreasoning
Deep learning methods demonstrate a compelling advantage over traditional machine learning methods, especially when abundant data are available.
“Particularly in tasks such as AI diagnosis and prediction, deep learning methods demonstrate a compelling advantage over traditional machine learning methods, especially when abundant data are available.”
Quote in the paperNot a substantial claim - Premisecitation
Li et al. achieved 0.939 AUC on ADNI using hippocampal morphology features from MRI.
“| Li et al. [52] | Hippocampal morphology feature | MRI | ADNI | 0.939 (AUC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [52] Li, H., Habes, M., and Fan, Y. (2017). Deep ordinal ranking for multi-category diagnosis of Alzheimer’s disease using hippocampal MRI data. arXiv.
Not checked in the free run.
- Premisecitation
Several large models focused on medical text and speech, including the Med-PaLM series, HuaTuo Algorithm, ChatDoctor, DoctorGLM, BianQue, and BioGPT, have demonstrated significant potential in providing valuable assistance across various healthcare-related domains.
“Several LMs focused on medical text and speech have also emerged, including the Med-PaLM series (Med-PaLM [19] and PaLM 2 [20]), HuaTuo Algorithm [21], ChatDoctor [22], DoctorGLM [23], BianQue [24], and BioGPT [25], which have demonstrated significant potential in providing valuable assistance across various healthcare-related domains.”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites [19] Singhal, K., Azizi, S., Tu, T., Mahdavi, S.S., Wei, J., Chung, H.W., Scales, N., Tanwani, A., Cole-Lewis, H., and Pfohl, S. (2022). Large language models encode clinical knowledge. arXiv.
Not checked in the free run.
- cites [20] Singhal, K., Tu, T., Gottweis, J., Sayres, R., Wulczyn, E., Hou, L., Clark, K., Pfohl, S., Cole-Lewis, H., and Neal, D. (2023). Towards expert-level medical question answering with large language models. arXiv.
Not checked in the free run.
- cites [21] Wang, H., Liu, C., Xi, N., Qiang, Z., Zhao, S., Qin, B., and Liu, T. (2023). Huatuo: Tuning llama model with chinese medical knowledge. arXiv.
Not checked in the free run.
- Premisereasoning
There has been extensive research on large models tailored for specific medical domains.
“Consequently, there has been extensive research on LMs tailored for specific medical domains to better fulfill the requirements.”
Quote in the paperNot a substantial claim - Premisenone given
Artificial intelligence techniques in breast cancer diagnosis research based on genomics primarily focus on RNA-seq data, single nucleotide polymorphisms, protein sequences, and the integration of multi-omics data.
“Presently, artificial intelligence techniques in breast cancer diagnosis research based on genomics primarily focus on RNA-seq data, single nucleotide polymorphisms (SNPs), protein sequences, and the integration of multi-omics data.”
Quote in the paperNot a substantial claim - Premisecitation
Shi et al. achieved 0.9713 ± 0.0444 accuracy for AD vs. NC on ADNI using high-level features of MRI and PET.
“| Shi et al. [79] | high-level features of MRI and PET | MRI, PET | ADNI | 0.9713 ± 0.0444 (ACC: AD vs. NC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [79] Shi, 2017, Multimodal neuroimaging feature learning with multimodal stacked deep polynomial networks for diagnosis of Alzheimer’s disease, IEEE J. Biomed. Health Inform. source
Not checked in the free run.
- Premisecitation
Sherafatian et al.'s Rpart algorithm achieved the best classification performance for breast cancer subtypes using miRNA data from TCGA.
“The results showed that Rpart achieved the best classification performance.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [50] Sherafatian, 2018, Tree-based machine learning algorithms identified minimal set of miRNA biomarkers for breast cancer diagnosis and molecular subtyping, Gene source
Not checked in the free run.
- Premisecitation
Al-Antari et al. [94] achieved an accuracy of 0.985 on CBIS-DDSM and DDSM using original X-ray mammographic data.
“| Al-Antari et al. [94] | Original X-ray mammographic data | X-ray | CBIS-DDSM and DDSM | 0.985 (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [94] Al-Antari, M.A., Han, S.M., and Kim, T.S. (2020). Evaluation of deep learning detection and classification towards computer-aided diagnosis of breast lesions in digital X-ray mammograms. Comput. Methods Programs Biomed., 196. source
Not checked in the free run.
- Premisecitation
Chen et al. integrated semantic trees into BERT to construct an interpretable neural network.
“Chen et al. [104] took the capabilities of BERT further by integrating semantic trees into the model, thus constructing an interpretable neural network.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [104] Chen, D., Zhong, K., and He, J. (2021, January 13–15). BDCN: Semantic Embedding Self-Explanatory Breast Diagnostic Capsules Network. Proceedings of the China National Conference on Chinese Computational Linguistics, Hohhot, China. source
Not checked in the free run.
- Interpretationreasoning
These studies illustrate a growing trend in leveraging sophisticated computational models and algorithms to enhance the accuracy and reliability of Alzheimer's disease diagnostics by capitalizing on complementary information from multiple imaging modalities.
“These studies illustrate a growing trend in leveraging sophisticated computational models and algorithms to enhance the accuracy and reliability of Alzheimer’s disease diagnostics by capitalizing on complementary information from multiple imaging modalities.”
Quote in the paperSubstantial claim- Rests on an assumption: The integration of MRI and PET imaging modalities yields a synergistic approach that offers a more holistic representation of the brain’s structure and function.
- Premisecitation
Xu et al.'s gcForest classifier achieved a higher accuracy score of 92% for predicting four subtypes of breast cancer compared with KNN, SVM, and MLP.
“The results showed that gcForest showed a higher accuracy score of 92%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [112] Xu, J., Wu, P., Chen, Y., and Zhang, L. (2018, January 14–17). Comparison of different classification methods for breast cancer subtypes prediction. Proceedings of the 2018 International Conference on Security, Pattern Analysis, and Cybernetics (SPAC), Jinan, China. source
Not checked in the free run.
- Premisecitation
Numerous research efforts are focused on utilizing genetic data and artificial intelligence algorithms to develop diagnostic models to enhance the clinical efficiency and accuracy of breast cancer diagnosis.
“Currently, numerous research efforts are focused on utilizing genetic data and artificial intelligence algorithms to develop diagnostic models to enhance the clinical efficiency and accuracy of breast cancer diagnosis [109,110,111].”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites [109] Witten, 2010, A framework for feature selection in clustering, J. Am. Stat. Assoc. source
Not checked in the free run.
- cites [110] Shen, 2009, Integrative clustering of multiple genomic data types using a joint latent variable model with application to breast and lung cancer subtype analysis, Bioinformatics source
Not checked in the free run.
- cites [111] Curtis, 2012, The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups, Nature source
Not checked in the free run.
- Premisecitation
Golovanevsky et al. achieved 0.9688 accuracy on GEO using imaging, genetic, and clinical data.
“| Golovanevsky et al. [89] | imaging, genetic, and clinical data | Gene-based | GEO | 0.9688 (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [89] Golovanevsky, 2022, Multimodal attention-based deep learning for Alzheimer’s disease diagnosis, J. Am. Med Inform. Assoc. source
Not checked in the free run.
- Premisecitation
Cheng et al. achieved 0.9528 AUC for AD vs. NC on ADNI using a sequence of 2D slice groups from 3D PET.
“| Cheng et al. [53] | a sequence of 2D slice groups from 3D PET | PET | ADNI | 0.9528 (AUC: AD vs. NC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [53] Cheng, D., and Liu, M. (2017, January 18–20). Combining convolutional and recurrent neural networks for Alzheimer’s disease diagnosis using PET images. Proceedings of the 2017 IEEE International Conference on Imaging Systems and Techniques (IST), Beijing, China. source
Not checked in the free run.
- Premisecitation
Jabeen et al.'s classification framework for ultrasound images achieved a highest accuracy of 99.1% on the Augmented Breast Ultrasound Images (BUSI) dataset.
“The implementation of this framework on the Augmented Breast Ultrasound Images (BUSI) dataset resulted in an impressive highest accuracy of 99.1%”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [100] Jabeen, K., Khan, M.A., Alhaisoni, M., Tariq, U., Zhang, Y.D., Hamza, A., Mickus, A., and Damaševičius, R. (2022). Breast cancer classification from ultrasound images using probability-based optimal deep learning feature fusion. Sensors, 22. source
Not checked in the free run.
- Premisecitation
Martinc et al. achieved accuracy 0.9167 on Dementia-Bank Pit using bag-of-n-grams text features and eGeMAPS voice features.
“| Martinc et al. [43] | bag-of-n-grams features (text) eGeMAPS feature set (voice) | speech, text | Dementia- Bank Pit | Accuracy: 0.9167 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [43] Martinc, 2021, Temporal integration of text transcripts and acoustic features for Alzheimer’s diagnosis based on spontaneous speech, Front. Aging Neurosci. source
Not checked in the free run.
- Premisecitation
Yu et al. achieved accuracy 0.93 on a private dataset using phoneme characteristics, pronunciation coordination characteristics, and pitch variance.
“| Yu et al. [74] | Based on phoneme characteristics, pronunciation coordination characteristics, and pitch variance | speech | private dataset | Accuracy: 0.93 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [74] Yu, B., Quatieri, T.F., Williamson, J.R., and Mundt, J.C. (2015, January 6–10). Cognitive impairment prediction in the elderly based on vocal biomarkers. Proceedings of the Sixteenth Annual Conference of the International Speech Communication Association, Dresden, Germany. source
Not checked in the free run.
- Premisecitation
Madani et al. [139]'s semi-supervised GAN data augmentation attained a test accuracy of 92.3% with limited data.
“GANs demonstrated superior performance than traditional CNNs with limited data, attaining a test accuracy of 92.3%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [139] Madani, 2018, Deep echocardiography: Data-efficient supervised and semi-supervised deep learning towards automated diagnosis of cardiac disease, NPJ Digit. Med. source
Not checked in the free run.
- Premisecitation
Zhang et al. [44] achieved an AUC over 90% for classifying acute myocardial infarction and stable coronary artery disease.
“This algorithm was applied to classify patients with Acute Myocardial Infarction (AMI) and stable coronary artery disease (SCAD), using GEO datasets GSE60993, GSE62646, GSE48060, and GSE59867, achieving an AUC over 90%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [44] Zhang, 2023, Integration of machine learning to identify diagnostic genes in leukocytes for acute myocardial infarction patients, J. Transl. Med. source
Not checked in the free run.
- Premisecitation
Single-channel EEG analysis can effectively differentiate depression patients.
“By refining feature selection and classification models via genetic algorithms, it has been discovered that single-channel analysis can effectively differentiate depression patients, underscoring the potential for employing portable EEG devices in preliminary depression screening despite a noted limitation in clinical generalizability due to small sample sizes.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [130] Wan, 2019, Single-channel EEG-based machine learning method for prescreening major depressive disorder, Int. J. Inf. Technol. Decis. Mak. source
Not checked in the free run.
- Premisecitation
Higuchi's fractal dimension (HFD) and sample entropy (SampEn) are efficacious in segregating depression patients from healthy individuals with high accuracy across machine learning classifiers.
“These nonlinear metrics have proven efficacious in segregating depression patients from healthy individuals, with high accuracy figures reported across a range of machine learning classifiers.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [128] Pokrajac, 2020, The successful discrimination of depression from EEG could be attributed to proper feature extraction and not to a particular classification method, Cogn. Neurodyn. source
Not checked in the free run.
- Premisecitation
Sherafatian et al. [50] achieved an accuracy of 92% on TCGA.
“| Sherafatian et al. [50] | miRNA | Gene-based | TCGA | 92% (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [50] Sherafatian, 2018, Tree-based machine learning algorithms identified minimal set of miRNA biomarkers for breast cancer diagnosis and molecular subtyping, Gene source
Not checked in the free run.
- Premisecitation
Tison et al. [152]'s vectorized processing approach better retained meaningful features in ECGs, enhancing interpretability and accuracy compared to traditional neural network models.
“Compared to traditional neural network models, this vectorized processing approach better retained meaningful features in ECGs, thus enhancing the interpretability and accuracy of diagnostic results.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [152] Tison, 2019, Automated and interpretable patient ECG profiles for disease detection, tracking, and discovery, Circ. Cardiovasc. Qual. Outcomes source
Not checked in the free run.
- Premisecitation
Xu et al. [115] found that their Linguistic Landscape Model performed on par with leading task-specific models across psychological stress classification, depression severity assessment, suicide ideation detection, and suicide risk evaluation.
“The empirical results underscored the LLM’s robust performance, placing it on par with the leading task-specific models in the field.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [115] Xu, X., Yao, B., Dong, Y., Yu, H., Hendler, J., Dey, A.K., and Wang, D. (2023). Leveraging large language models for mental health prediction via online text data. arXiv.
Not checked in the free run.
- Premisecitation
Asymmetries in the alpha2 and theta bands, particularly when analyzed with a Support Vector Machine (SVM), achieved an accuracy rate of 88.33% for depression diagnosis.
“Notably, findings suggest that asymmetries in the alpha2 and theta bands, particularly when analyzed with a Support Vector Machine (SVM), lead to higher diagnostic precision, with an accuracy rate of 88.33%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [129] Mahato, 2020, Classification of depression patients and normal subjects based on electroencephalogram (EEG) signal using alpha power and theta asymmetry, J. Med. Syst. source
Not checked in the free run.
- Premisecitation
Nasimova et al. [140] achieved a test accuracy of 98.2% for classifying dilated cardiomyopathy versus hypertrophic cardiomyopathy.
“These images and extracted features were input into a six-layer CNN for classification, achieving a test accuracy of 98.2%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [140] Nasimova, N., Muminov, B., Nasimov, R., Abdurashidova, K., and Abdullaev, M. (2021, January 3–5). Comparative analysis of the results of algorithms for dilated cardiomyopathy and hypertrophic cardiomyopathy using deep learning. Proceedings of the 2021 International Conference on Information Science and Communications Technologies (ICISCT), Tashkent, Uzbekistan. source
Not checked in the free run.
- Premisecitation
Mishra et al. [144] achieved an F1 score of 97% for heart disease risk estimation.
“Through comprehensive simulations and evaluations, the Rec-CONVnet demonstrated remarkable performance, achieving an impressive F1 score of 97%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [144] Mishra, J., Tiwari, M., Singh, S.T., and Goswami, S. (2022, January 11–12). Detection of heart disease employing Recurrent CONVoluted neural networks (Rec-CONVnet) for effectual classification process in smart medical application. Proceedings of the 2021 4th International Conference on Recent Trends in Computer Science and Technology (ICRTCST), Jamshedpur, India. source
Not checked in the free run.
- Premisecitation
Madani et al. [139] achieved an accuracy of 91.2% for binary left ventricular hypertrophy diagnosis using echocardiography images.
“The core framework of their model included a U-Net for eliminating auxiliary information from image and a series of convolutional neural networks, resulting in an accuracy of 91.2%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [139] Madani, 2018, Deep echocardiography: Data-efficient supervised and semi-supervised deep learning towards automated diagnosis of cardiac disease, NPJ Digit. Med. source
Not checked in the free run.
- Premisecitation
Liu et al. [148] achieved an AUC of 97.1% in the training set and 98.9% in the testing set for coronary artery disease classification.
“They used data from the GEO dataset GSE113079, achieving an AUC of 97.1% in the training set and 98.9% in the testing set.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [148] Liu, 2021, Classifiers for predicting coronary artery disease based on gene expression profiles in peripheral blood mononuclear cells, Int. J. Gen. Med. source
Not checked in the free run.
- Premisecitation
Hou et al. [149] achieved an AUC of 0.944.
“The study achieved an AUC of 0.944.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [149] Hou, 2023, Role of serum cytokines in the prediction of heart failure in patients with coronary artery disease, ESC Heart Fail. source
Not checked in the free run.
- Premisecitation
Ghosh et al. [114] found that a multi-task learning framework significantly boosts the efficacy of emotion recognition and depression detection when learned concurrently.
“Their findings suggest that the multi-tasking framework significantly boosts the efficacy of both tasks when learned concurrently.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [114] Ghosh, 2022, A multitask framework to detect depression, sentiment and multi-label emotion from suicide notes, Cogn. Comput. source
Not checked in the free run.
- Premisecitation
Multi-modal fusion of EEG and eye movement signals in hidden layers yields improved recognition accuracy over single-feature methods for mild depression.
“The application of deep learning to fuse these multi-modal data sets enables real-time monitoring and detection of mild depression, with the fusion approach in the hidden layers yielding improved recognition accuracy over single-feature methods, and showcasing the benefits of combining diverse physiological signals.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [47] Zhu, 2019, Multimodal mild depression recognition based on EEG-EM synchronization acquisition network, IEEE Access source
Not checked in the free run.
- Premisecitation
The data preparation phase is critical for achieving optimal classification accuracy in EEG-based depression diagnosis.
“Through rigorous data preprocessing and feature extraction—identifying noise types and harnessing both linear and nonlinear features—the critical role of the data preparation phase is emphasized for achieving optimal classification accuracy.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [131] Cai, 2018, Study on feature selection methods for depression detection using three-electrode EEG data, Interdiscip. Sci. Comput. Life Sci. source
Not checked in the free run.
- Premisecitation
The Asymmetry Matrix Image (AMI) outperforms traditional methods for classifying depression patients versus healthy controls.
“AMI has been shown to outperform traditional methods, delivering superior classification accuracy and enhancing the distinction between depression patients and healthy controls.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [127] Kang, M., Kwon, H., Park, J.H., Kang, S., and Lee, Y. (2020). Deep-asymmetry: Asymmetry matrix image for deep learning method in pre-screening depression. Sensors, 20. source
Not checked in the free run.
- Premisecitation
Achilles et al. [158] achieved sensitivity of 87% and specificity of 81% for generalized tonic-clonic seizures.
“Combining outputs from both networks, they achieved the prediction of ictal or interictal epilepsy phases, with their method demonstrating high sensitivity (87%) and specificity (81%) for generalized tonic-clonic seizures.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [158] Achilles, 2015, Deep convolutional neural networks for automatic identification of epileptic seizures in infrared and depth images, J. Neurol. Sci. source
Not checked in the free run.
- Premisecitation
The Mayo Clinic algorithm generally performed better in the adolescent group than in the pediatric group.
“The study found that the algorithm generally performed better in the adolescent group than in the pediatric group.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [154] Ko, 2020, Detection of hypertrophic cardiomyopathy using a convolutional neural network-enabled electrocardiogram, J. Am. Coll. Cardiol. source
Not checked in the free run.
- Interpretationcitation
Deep learning methods like mDNN models have potential in TLE classification when combined with multiple datasets.
“These findings affirm the potential of deep learning methods, like mDNN models, in TLE classification when combined with multiple datasets.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [174] Torres-Velázquez, M., Hwang, G., Cook, C.J., Hermann, B.P., Prabhakaran, V., Meyerand, M.E., and McMillan, A.B. (2020, January 4). Multi-Channel Deep Neural Network For Temporal Lobe Epilepsy Classification Using Multimodal Mri Data. Proceedings of the 2020 IEEE 17th International Symposium on Biomedical Imaging Workshops (ISBI Workshops), Iowa City, IA, USA. source
Not checked in the free run.
- Premisecitation
Hamid et al. [167]'s NLP tool was proficient in detecting concepts indicative of PNES and those negating its presence.
“This algorithm was proficient in detecting concepts indicative of PNES and those negating its presence.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [167] Hamid, 2013, Validating a natural language processing tool to exclude psychogenic nonepileptic seizures in electronic medical record-based epilepsy research, Epilepsy Behav. source
Not checked in the free run.
- Premisecitation
Sahebzamani et al. [163] achieved 94% accuracy for epilepsy diagnosis.
“This comprehensive segmentation and whole-brain analysis methodology yielded a 94% accuracy rate.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [163] Sahebzamani, G., Saffar, M., and Soltanian-Zadeh, H. (2019, January 6–7). Machine learning based analysis of structural MRI for epilepsy diagnosis. Proceedings of the 2019 4th International Conference on Pattern Recognition and Image Analysis (IPRIA), Tehran, Iran. source
Not checked in the free run.
- Interpretationreasoning
The selection of modalities for automated diagnosis is intricately tied to the unique characteristics and pathological features of each disease.
“In conclusion, the selection of modalities for automated diagnosis is intricately tied to the unique characteristics and pathological features of each disease.”
Quote in the paperSubstantial claim - Premisecitation
Ahmedt-Aristizabal [159] achieved an average classification accuracy of 89% for differentiating two types of epileptic seizures.
“The developed network demonstrated its capability to differentiate between two types of epileptic seizures, achieving an average classification accuracy of 89%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [159] Ahmedt-Aristizabal, D., Nguyen, K., Denman, S., Sarfraz, M.S., Sridharan, S., Dionisio, S., and Fookes, C. (2019, January 23–27). Vision-based mouth motion analysis in epilepsy: A 3d perspective. Proceedings of the 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Berlin, Germany. source
Not checked in the free run.
- Interpretationreasoning
For Alzheimer's disease diagnosis, MRI and PET images are the predominant modalities, supplemented by voice data.
“For Alzheimer’s disease diagnosis, Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) images emerge as the predominant modalities, supplemented by the inclusion of voice data.”
Quote in the paperNot a substantial claim - Interpretationreasoning
In depression diagnosis, the emphasis shifts toward textual data and EEG.
“In the context of depression diagnosis, the emphasis shifts toward textual data and Electroencephalogram (EEG).”
Quote in the paperNot a substantial claim - Premisecitation
Pevy et al. [168]'s algorithm distinguished epileptic seizures from PNES based on verbal expression patterns.
“The algorithm then trained machine learning classifiers with these features, enabling it to distinguish patients based on their verbal expression patterns.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [168] Pevy, 2021, Feasibility of using an automated analysis of formulation effort in patients’ spoken seizure descriptions in the differential diagnosis of epileptic and nonepileptic seizures, Seizure source
Not checked in the free run.
- Premisecitation
Maia et al. [157] achieved 65% AUC for classifying epileptic seizures.
“Maia et al. [157] applied a threshold-based approach to first detect regions of interest (beds) in video data, aligning them vertically for consistency, then utilized Convolutional Neural Networks and Multilayer Perceptrons to classify epileptic seizures, achieving 65% AUC.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [157] Maia, P., Hartl, E., Vollmar, C., Noachtar, S., and Cunha, J.P.S. (2019, January 22–23). Epileptic seizure classification using the NeuroMov database. Proceedings of the 2019 IEEE 6th Portuguese Meeting on Bioengineering (ENBENG), Lisbon, Portugal. source
Not checked in the free run.
- Interpretationreasoning
There are at least two common scenarios where ethical issues arise in medical data.
“There are at least two common scenarios where ethical issues arise in medical data.”
Quote in the paperNot a substantial claim - Premisecitation
Even desensitized data may still be re-identifiable through sophisticated analysis methods.
“However, studies indicate that even desensitized data may still be re-identifiable through sophisticated analysis methods [178].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [178] Chia, P.H., Desfontaines, D., Perera, I.M., Simmons-Marengo, D., Li, C., Day, W.Y., Wang, Q., and Guevara, M. (2019, January 19–23). KHyperLogLog: Estimating reidentifiability and joinability of large data at scale. Proceedings of the 2019 IEEE Symposium on Security and Privacy (SP), San Francisco, CA, USA. source
Not checked in the free run.
- Interpretationreasoning
Deploying AI models developed for specific diseases to other disease predictions often demonstrates limited generalization ability.
“However, deploying AI models developed for specific diseases to other disease predictions often demonstrates limited generalization ability.”
Quote in the paperSubstantial claim - Premisecitation
Connolly et al. [169]'s NLP algorithm successfully classified partial epilepsy, generalized epilepsy, and unclassified epilepsy with remarkable accuracy.
“By analyzing text features extracted from electronic medical records, their algorithm successfully classified different subtypes of epilepsy with remarkable accuracy.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [169] Connolly, 2014, Assessing the similarity of surface linguistic features related to epilepsy across pediatric hospitals, J. Am. Med. Inform. Assoc. source
Not checked in the free run.
- Interpretationcitation
Algorithms trained on datasets with these characteristics may exacerbate health disparities if adopted in healthcare.
“If algorithms trained on datasets with these characteristics are adopted in healthcare, they may exacerbate health disparities [185].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [185] Braveman, 2006, Health disparities and health equity: Concepts and measurement, Annu. Rev. Public Health source
Not checked in the free run.
- Interpretationreasoning
Distinctive preferences for specific modalities exist across different diseases.
“Distinctive preferences for specific modalities exist across different diseases.”
Quote in the paperNot a substantial claim - Premisecitation
Yao et al. [173] achieved the best accuracy with a 31-layer Dense IndRNN with attention among ten IndRNN architectures.
“Yao et al. [173] experimented with ten different and independently improved RNN (IndRNN) architectures, achieving the best accuracy with a 31-layer Dense IndRNN with attention (DIndRNN).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [173] Yao, X., Cheng, Q., and Zhang, G.Q. (2019). Automated classification of seizures against nonseizures: A deep learning approach. arXiv.
Not checked in the free run.
- Premisecitation
Garner et al. [162] achieved 69% accuracy in predicting epilepsy outcomes.
“The model achieved a 69% accuracy rate in predicting epilepsy outcomes on the test set after 100 stratified cross-validation rounds, using 70% of resting-state fMRI scans for training and 30% for testing.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [162] Garner, R., La Rocca, M., Barisano, G., Toga, A.W., Duncan, D., and Vespa, P. (May, January 29). A machine learning model to predict seizure susceptibility from resting-state fMRI connectivity. Proceedings of the 2019 Spring Simulation Conference (SpringSim), Tucson, AZ, USA. source
Not checked in the free run.
- Classified as probably not a substantial claim. 103
- May combine two claims. 61
- The quote only partly states this claim. 1
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationOASIS-3 primarily investigates the progression of Alzheimer's disease in 1378 individuals.
“This dataset primarily investigates the progression of AD in 1378 individuals.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [2] Marcus, 2007, Open Access Series of Imaging Studies (OASIS): Cross-sectional MRI data in young, middle aged, nondemented, and demented older adults, J. Cogn. Neurosci. source
Not checked in the free run.
Probably not a substantial claim: may not pass S4 (based).
Premisenone givenThe clinical diagnostic process is inherently intricate, involving the generation and analysis of diverse data types encompassing images, speech, text, and genetic information.
“The clinical diagnostic process is inherently intricate, involving the generation and analysis of diverse data types encompassing images, speech, text, and genetic information (as depicted in Figure 1).”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationThe Bonn EEG time series database involves EEG data obtained from a 128-channel acquisition system, featuring recordings from 5 patients identified as A, B, C, D, and E.
“The Bonn EEG time series database [36] involves EEG data obtained from a 128-channel acquisition system, featuring recordings from 5 patients identified as A, B, C, D, and E.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [36] Andrzejak, 2001, Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state, Phys. Rev. E source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationThe Cancer Genome Atlas (TCGA) is a widely utilized dataset for predicting breast cancer, involving MRI and CT scans, clinical records, and genetic information.
“The Cancer Genome Atlas (TCGA) [26] is a widely utilized dataset for predicting breast cancer. It involves MRI and CT scans, clinical records and genetic information.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [26] Tomczak, 2015, Review The Cancer Genome Atlas (TCGA): An immeasurable source of knowledge, Contemp. Oncol. Onkol.
Not checked in the free run.
Probably not a substantial claim: may not pass S4 (based). May combine two claims.
InterpretationreasoningEach modality (image, text, speech, genetic, physiological signals) furnishes unique and valuable insights that collectively contribute to a holistic understanding of patients' physiological states.
“Each modality furnishes unique and valuable insights that collectively contribute to a holistic understanding of patients’ physiological states.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationThe CHB-MIT Database comprises EEG recordings collected from 22 pediatric subjects with intractable seizures and was established in 2010.
“The CHB-MIT [35] Database comprises EEG recordings collected from 22 pediatric subjects with intractable seizures and was established in 2010.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [35] Goldberger, 2000, PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals, Circulation source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationThe Multi-modal Open Dataset for Mental Disorder Analysis (MODMA) is a multi-modal dataset tailored for mental disorders, featuring both clinically depressed patients and individuals from the normal population.
“The Multi-modal Open Dataset for Mental Disorder Analysis (MODMA) [34] is a multi-modal dataset tailored for mental disorders, featuring both clinically depressed patients and individuals from the normal population.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [34] Cai, 2022, A multi-modal open dataset for mental-disorder analysis, Sci. Data source
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.
Interpretationnone givenAI techniques spanning machine learning, deep learning, and large model paradigms can significantly augment physicians in rendering more evidence-based decisions, presenting a pioneering solution for clinical practice.
“Artificial intelligence (AI) techniques, spanning from machine learning and deep learning to large model paradigms, stand poised to significantly augment physicians in rendering more evidence-based decisions, thus presenting a pioneering solution for clinical practice.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationIn the Bonn database, set E contains intracranial EEG data captured during epileptic seizures.
“Set E contains intracranial EEG data captured during epileptic seizures.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [36] Andrzejak, 2001, Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state, Phys. Rev. E source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationIn the Bonn database, each set consists of 100 text files, each containing a single EEG time series represented in ASCII code and comprising 4097 samples.
“Each set consists of 100 text files, each containing a single EEG time series represented in ASCII code and comprising 4097 samples.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [36] Andrzejak, 2001, Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state, Phys. Rev. E source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationPatients in the ADNI database are categorized into stages such as AD, MCI (Mild Cognitive Impairment), and NC (Normal Cognition).
“The patients in the ADNI database are categorized into different stages such as AD, MCI (Mild Cognitive Impairment), and NC (Normal Cognition).”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [7] Mueller, 2005, The Alzheimer’s disease neuroimaging initiative, Neuroimaging Clin. source
Not checked in the free run.
- cites [8] Schmand, 2016, Advanced Neuropsychological Diagnostics Infrastructure (ANDI): A normative database created from control datasets, Front. Psychol.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe Alzheimer's Disease Neuroimaging Initiative (ADNI) database is widely recognized as one of the most prominent datasets for predicting Alzheimer's disease.
“The Alzheimer’s Disease Neuroimaging Initiative (ADNI) database [7,8], established in 2003, is widely recognized as one of the most prominent datasets for predicting AD.”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [7] Mueller, 2005, The Alzheimer’s disease neuroimaging initiative, Neuroimaging Clin. source
Not checked in the free run.
- cites [8] Schmand, 2016, Advanced Neuropsychological Diagnostics Infrastructure (ANDI): A normative database created from control datasets, Front. Psychol.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationMODMA comprises speech data and ECG data.
“It comprises speech data and ECG data.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [34] Cai, 2022, A multi-modal open dataset for mental-disorder analysis, Sci. Data source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationThe Bonn database is devoid of artifacts, obviating the necessity for preprocessing prior to classifying the signals as healthy (non-epileptic) or unhealthy (epileptic).
“This database is devoid of artifacts, obviating the necessity for preprocessing prior to classifying the signals as healthy (non-epileptic) or unhealthy (epileptic).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [36] Andrzejak, 2001, Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state, Phys. Rev. E source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationIn the TCGA dataset, breast cancer is categorized into subtypes including Luminal A, Luminal B, HER2+, Basel, etc.
“In the TCGA dataset, breast cancer is categorized into different subtypes, including Luminal A, Luminal B, HER2+, Basel, etc.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [26] Tomczak, 2015, Review The Cancer Genome Atlas (TCGA): An immeasurable source of knowledge, Contemp. Oncol. Onkol.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationOASIS-3 is a longitudinal dataset that integrates multiple modalities including neuroimaging, clinical biomarkers, and cognitive assessment.
“Another typical database is the longitudinal dataset called OASIS-3, which integrates multiple modalities [2], including neuroimaging, clinical biomarkers, and cognitive assessment.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [2] Marcus, 2007, Open Access Series of Imaging Studies (OASIS): Cross-sectional MRI data in young, middle aged, nondemented, and demented older adults, J. Cogn. Neurosci. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe Distress Analysis Interview Corpus-Wizard of OZ (DAIC-WOZ) stands as one of the most popular speech datasets utilized for depression prediction.
“The Distress Analysis Interview Corpus-Wizard of OZ (DAIC-WOZ) [33] stands as one of the most popular speech datasets utilized for depression prediction.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [33] Gratch, J., Artstein, R., Lucas, G.M., Stratou, G., Scherer, S., Nazarian, A., Wood, R., Boberg, J., DeVault, D., and Marsella, S. (2014, January 26–31). The distress analysis interview corpus of human and computer interviews. Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), Reykjavik, Iceland.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationSince 2006, the UK Biobank has amassed substantial data from participants encompassing environmental factors, lifestyle choices, sociodemographic information, overall health and well-being, and cognitive and physical assessments.
“Additionally, since 2006, the UK Biobank (UKB) [3,4,5] has amassed a substantial amount of data from participants, encompassing various fields such as environmental factors, lifestyle choices, sociodemographic information, overall health and well-being, as well as cognitive and physical assessments [6].”
Quote in the paperNot a substantial claim0 of 4 sources checked- cites [3] Allen, 2012, UK Biobank: Current status and what it means for epidemiology, Health Policy Technol. source
Not checked in the free run.
- cites [4] Littlejohns, 2019, UK Biobank: Opportunities for cardiovascular research, Eur. Heart J. source
Not checked in the free run.
- cites [5] Sudlow, C., Gallacher, J., Allen, N., Beral, V., Burton, P., Danesh, J., Downey, P., Elliott, P., Green, J., and Landray, M. (2015). UK biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med., 12. source
Not checked in the free run.
- cites [6] Wilkinson, 2019, Identifying dementia outcomes in UK Biobank: A validation study of primary care, hospital admissions and mortality data, Eur. J. Epidemiol. source
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable), S4 (based).
Interpretationnone givenThe amalgamation of diverse medical data modalities (image, text, speech, genetic data, physiological signals) is imperative to facilitate a comprehensive disease analysis.
“Traditionally, the amalgamation of diverse medical data modalities (e.g., image, text, speech, genetic data, physiological signals) is imperative to facilitate a comprehensive disease analysis, a topic of burgeoning interest among both researchers and clinicians in recent times.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe DAIC-WOZ corpus encompasses a broad spectrum of psychological disorders, including depression, anxiety, and post-traumatic stress disorder.
“The corpus encompasses a broad spectrum of psychological disorders, including depression, anxiety, and post-traumatic stress disorder.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [33] Gratch, J., Artstein, R., Lucas, G.M., Stratou, G., Scherer, S., Nazarian, A., Wood, R., Boberg, J., DeVault, D., and Marsella, S. (2014, January 26–31). The distress analysis interview corpus of human and computer interviews. Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), Reykjavik, Iceland.
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
InterpretationcitationAI-driven healthcare represents a promising avenue for addressing critical disease diagnosis challenges in regions characterized by disparities in medical resources.
“It represents a promising avenue for addressing critical disease diagnosis challenges in regions characterized by disparities in medical resources, garnering significant attention from both scholars and practitioners [1].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [1] Anto, 2011, Supervised machine learning approaches for medical data set classification—A review, Int. J. Comput. Sci. Trends Technol.
Not checked in the free run.
- Rests on an assumption: Achieving timely diagnosis remains challenging, particularly in developing areas and regions with limited medical resources.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationIn the Bonn database, sets C and D encompass intracranial EEG recordings taken during seizure-free intervals, with set C recorded from within the seizure-generating area and set D from outside the seizure-generating area of epileptic patients.
“Sets C and D encompass intracranial EEG recordings taken during seizure-free intervals, with set C recorded from within the seizure-generating area and set D from outside the seizure-generating area of epileptic patients.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [36] Andrzejak, 2001, Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state, Phys. Rev. E source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe ADNI database encompasses brain imaging data such as MRI and PET scans, clinical data, biospecimen information, and genetic data.
“It encompasses various types of data, including brain imaging data such as MRI and PET scans, clinical data, biospecimen information, and genetic data.”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [7] Mueller, 2005, The Alzheimer’s disease neuroimaging initiative, Neuroimaging Clin. source
Not checked in the free run.
- cites [8] Schmand, 2016, Advanced Neuropsychological Diagnostics Infrastructure (ANDI): A normative database created from control datasets, Front. Psychol.
Not checked in the free run.
May combine two claims.
PremisecitationCNN has been widely applied in various medical imaging, laboratory reports, pathology reports, etc., and has achieved remarkable success in the field of AI-based diagnosis.
“Recently, the CNN has been widely applied in various medical imaging, laboratory reports, pathology reports, etc., and has achieved remarkable success in the field of AI-based diagnosis [51,52,53,54,55,56,57,58].”
Quote in the paperNot a substantial claim0 of 8 sources checked- cites [51] Dong, 2022, An unsupervised domain adaptation brain CT segmentation method across image modalities and diseases, Expert Syst. Appl. source
Not checked in the free run.
- cites [52] Li, H., Habes, M., and Fan, Y. (2017). Deep ordinal ranking for multi-category diagnosis of Alzheimer’s disease using hippocampal MRI data. arXiv.
Not checked in the free run.
- cites [53] Cheng, D., and Liu, M. (2017, January 18–20). Combining convolutional and recurrent neural networks for Alzheimer’s disease diagnosis using PET images. Proceedings of the 2017 IEEE International Conference on Imaging Systems and Techniques (IST), Beijing, China. source
Not checked in the free run.
- cites [54] Liu, 2021, A computer-aided system for ocular myasthenia gravis diagnosis, Tsinghua Sci. Technol. source
Not checked in the free run.
- cites [55] Calvo, 2018, Advances on automatic speech analysis for early detection of Alzheimer disease: A non-linear multi-task approach, Curr. Alzheimer Res. source
Not checked in the free run.
- cites [56] Guan, 2023, Deep Learning Blockchain Integration Framework for Ureteropelvic Junction Obstruction Diagnosis Using Ultrasound Images, Tsinghua Sci. Technol. source
Not checked in the free run.
- cites [57] Zhang, J., Liu, B., Wu, J., Wang, Z., and Li, J. (2023). DeepCAC: A deep learning approach on DNA transcription factors classification based on multi-head self-attention and concatenate convolutional neural network. BMC Bioinform., 24. source
Not checked in the free run.
- cites [58] Islam, 2020, An integrative deep learning framework for classifying molecular subtypes of breast cancer, Comput. Struct. Biotechnol. J. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningIn the medical domain, where high precision is paramount, traditional machine learning methods are progressively being substituted by deep learning methods.
“In the medical domain, where high precision is paramount, traditional machine learning methods are progressively being substituted by deep learning methods.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe Temple University EEG corpus database represents an extensive collection of EEG data acquired between 2000 and 2013.
“The Temple University EEG corpus database [37] represents an extensive collection of EEG data acquired between 2000 and 2013.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [37] Obeid, 2016, The temple university hospital EEG data corpus, Front. Neurosci. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationIn disease diagnosis tasks using artificial intelligence technology, performance evaluation metrics are commonly calculated based on the confusion matrix for binary classification tasks.
“In disease diagnosis tasks using artificial intelligence technology, performance evaluation metrics are commonly calculated based on the confusion matrix for binary classification tasks [69], which include four types of classifications: True Positive (TP), False Positive (FP), True Negative (TN), and False Negative (FN).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [69] Tharwat, 2020, Classification assessment methods, Appl. Comput. Inform. source
Not checked in the free run.
May combine two claims.
PremisecitationZhu et al. achieved 0.9193 accuracy for AD vs. NC vs. MCI and 0.9287 AUC on ADNI and AIBL using patch proposals selected from MRI scans.
“| Zhu et al. [59] | Patch proposals selected from the MRI scans | MRI | ADNI, AIBL | 0.9193 (ACC: AD vs. NC vs. MCI) 0.9287 (AUC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [59] Zhu, 2021, Dual attention multi-instance deep learning for Alzheimer’s disease diagnosis with structural MRI, IEEE Trans. Med. Imaging source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationA large universal foundational model only requires one training session to acquire a wide range of general knowledge and can subsequently adapt to various downstream tasks through prompts.
“Unlike traditional specialized models trained for specific problems, a large universal foundational model only requires one training session to acquire a wide range of general knowledge and can subsequently adapt to various downstream tasks through prompts.”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites [62] Bommasani, R., Hudson, D.A., Adeli, E., Altman, R., Arora, S., von Arx, S., Bernstein, M.S., Bohg, J., Bosselut, A., and Brunskill, E. (2021). On the opportunities and risks of foundation models. arXiv.
Not checked in the free run.
- cites [63] Zhou, C., Li, Q., Li, C., Yu, J., Liu, Y., Wang, G., Zhang, K., Ji, C., Yan, Q., and He, L. (2023). A comprehensive survey on pretrained foundation models: A history from bert to chatgpt. arXiv.
Not checked in the free run.
- cites [64] Brown, 2020, Language models are few-shot learners, Adv. Neural Inf. Process. Syst.
Not checked in the free run.
Probably not a substantial claim: may not pass S4 (based). May combine two claims.
Premisenone givenSpecialized models are meticulously designed for specific tasks in architecture and are guided by medical knowledge to better understand and capture subtle differences and semantic features in the data.
“In comparison to specialized models, LMs still exhibit certain gaps because specialized models are not only meticulously designed for specific tasks in terms of architecture but also guided by medical knowledge to better understand and capture subtle differences and semantic features in the data.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.
Premisenone givenTraditional machine learning algorithms continue to be favored in the development of AI diagnostic models because they require fewer data points and offer better interpretability.
“In the era of rapid advancements in deep learning algorithms, traditional machine learning algorithms continue to be favored in the development of AI diagnostic models due to their unique advantages. They require fewer data points and offer better interpretability.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationLian et al. achieved 0.9 accuracy and 0.95 AUC for AD vs. NC on ADNI using original MRI scan features.
“| Lian et al. [70] | Original MRI scan feature | MRI | ADNI | 0.9 (ACC); 0.95 (AUC:AD vs. NC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [70] Lian, 2018, Hierarchical fully convolutional network for joint atrophy localization and Alzheimer’s disease diagnosis using structural MRI, IEEE Trans. Pattern Anal. Mach. Intell. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningPreprocessing using machine learning and deep learning technologies is a crucial step for disease diagnosis.
“Pre-processing using machine learning and deep learning technologies is a crucial step for disease diagnosis.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.
Premisenone givenDeep learning methods can leverage all the information present in the data as features for training models, eliminating the need for predefined features.
“Unlike traditional machine learning approaches, deep learning methods can leverage all the information present in the data as features for training models, eliminating the need for predefined features.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationYang et al. introduced GeneCompass, the first knowledge-based cross-species milestone foundational model, surpassing competitive state-of-the-art models in multiple tasks within a single species.
“In the realm of genetic data, Yang et al. [49] introduced GeneCompass, the first knowledge-based cross-species milestone foundational model, surpassing competitive state-of-the-art models in multiple tasks within a single species.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [49] Yang, X., Liu, G., Feng, G., Bu, D., Wang, P., Jiang, J., Chen, S., Yang, Q., Zhang, Y., and Man, Z. (2023). GeneCompass: Deciphering Universal Gene Regulatory Mechanisms with Knowledge-Informed Cross-Species Foundation Model. bioRxiv, 1–8. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
Premisenone givenOther classification metrics such as the Area Under the ROC Curve (AUC-ROC) are commonly adopted.
“In addition, other classification metrics such as the Area Under the ROC Curve (AUC-ROC) are also commonly adopted.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationNaive Bayes (NB) is a probabilistic algorithm that does not rely on networks and performs well with high-dimensional features.
“NB [46] is a probabilistic algorithm that does not rely on networks and performs well with high-dimensional features.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [46] Zeng, Q.T., Goryachev, S., Weiss, S., Sordo, M., Murphy, S.N., and Lazarus, R. (2006). Extracting principal diagnosis, co-morbidity and smoking status for asthma research: Evaluation of a natural language processing system. BMC Med. Inform. Decis. Mak., 6. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
Premisenone givenTraditional machine learning algorithms often require domain experts to pre-define features before model training, resulting in additional manual costs and increased resource expenses.
“They often require domain experts to pre-define the features to be learned before model training, resulting in additional manual costs and increased resource expenses.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Interpretationnone givenFeature engineering plays a crucial role in disease diagnosis using artificial intelligence technologies.
“Feature engineering plays a crucial role in disease diagnosis using artificial intelligence technologies.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationChen et al. achieved 0.9193 accuracy for AD vs. NC vs. MCI and 0.9287 AUC on ADNI using optimized anchor data from brain 18F-FDG PET slices.
“| Chen et al. [60] | optimized anchor data from brain 18F-FDG PET slices | PET | ADNI | 0.9193 (ACC: AD vs. NC vs. MCI) 0.9287 (AUC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [60] Chen, 2022, Contrastive Learning for Prediction of Alzheimer’s Disease Using Brain 18F-FDG PET, IEEE J. Biomed. Health Inform. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationLSTM excels in capturing contextual relationships and predicting subsequent data based on the preceding sequence.
“Notably, LSTM excels in capturing contextual relationships and predicting subsequent data based on the preceding sequence.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [12] Hochreiter, 1997, Long short-term memory, Neural Comput. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Premisenone givenLarge models have proven effective in various general domain tasks.
“While these LMs have proven effective in various general domain tasks, they have yet to reach their full potential in specific medical domain tasks.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningIn disease diagnosis tasks, feature enhancement helps to more accurately distinguish different disease categories, thereby improving the accuracy and robustness of the model.
“In disease diagnosis tasks, feature enhancement helps to more accurately distinguish different disease categories, thereby improving the accuracy and robustness of the model.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe Temple University EEG corpus encompasses diverse EEG clinical settings from approximately 10,874 patients.
“This repository encompasses diverse EEG clinical settings from approximately 10,874 patients.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [37] Obeid, 2016, The temple university hospital EEG data corpus, Front. Neurosci. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationXrayGPT is a conversational medical visual language model capable of analyzing and answering open-ended questions regarding chest X-rays.
“XrayGPT is an innovative conversational medical visual language model capable of analyzing and answering open-ended questions regarding chest X-rays.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [68] Thawkar, O., Shaker, A., Mullappilly, S.S., Cholakkal, H., Anwer, R.M., Khan, S., Laaksonen, J., and Khan, F.S. (2023). Xraygpt: Chest radiographs summarization using medical vision-language models. arXiv.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationMei et al. achieved an accuracy of 87% and an RMSE of 3.727 for AD detection.
“They meticulously fine-tuned the wav2vec2.0 model on speech from assorted frequency bands, culminating in a remarkable accuracy of 87% and an RSME of 3.727.”
- Also stated in: § 3.1. Diagnosis of Alzheimer’s Disease · plain
“| Mei et al. [83] | Silent characteristics ComParE 2016 feature set, eGeMAPS feature set wav2vec2 pre-trained model feature vector Hubert pre-trained model feature vector | Speech, text | AADReSS-M | Accuracy: 0.87, RMSE: 3.727 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [83] Mei, K., Ding, X., Liu, Y., Guo, Z., Xu, F., Li, X., Naren, T., Yuan, J., and Ling, Z. (2023, January 4–6). The Ustc System for Adress-m Challenge. Proceedings of the ICASSP 2023—2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Rhode Island, Greece. source
Not checked in the free run.
- Also stated in: § 3.1. Diagnosis of Alzheimer’s Disease · plain
Probably not a substantial claim: may not pass S3 (load-bearing).
InterpretationcitationLarge model approaches are expected to further improve the ability to capture and generalize complex features in medical data.
“In the field of medical data, large model approaches are expected to further improve the ability to capture and generalize complex features [19,20,21,22,23,24,25].”
Quote in the paperNot a substantial claim0 of 7 sources checked- cites [19] Singhal, K., Azizi, S., Tu, T., Mahdavi, S.S., Wei, J., Chung, H.W., Scales, N., Tanwani, A., Cole-Lewis, H., and Pfohl, S. (2022). Large language models encode clinical knowledge. arXiv.
Not checked in the free run.
- cites [20] Singhal, K., Tu, T., Gottweis, J., Sayres, R., Wulczyn, E., Hou, L., Clark, K., Pfohl, S., Cole-Lewis, H., and Neal, D. (2023). Towards expert-level medical question answering with large language models. arXiv.
Not checked in the free run.
- cites [21] Wang, H., Liu, C., Xi, N., Qiang, Z., Zhao, S., Qin, B., and Liu, T. (2023). Huatuo: Tuning llama model with chinese medical knowledge. arXiv.
Not checked in the free run.
- cites [22] Yunxiang, L., Zihan, L., Kai, Z., Ruilong, D., and You, Z. (2023). Chatdoctor: A medical chat model fine-tuned on llama model using medical domain knowledge. arXiv.
Not checked in the free run.
- cites [23] Xiong, H., Wang, S., Zhu, Y., Zhao, Z., Liu, Y., Wang, Q., and Shen, D. (2023). Doctorglm: Fine-tuning your chinese doctor is not a herculean task. arXiv.
Not checked in the free run.
- cites [24] Chen, Y., Wang, Z., Xing, X., Xu, Z., Fang, K., Wang, J., Li, S., Wu, J., Liu, Q., and Xu, X. (2023). BianQue: Balancing the Questioning and Suggestion Ability of Health LLMs with Multi-turn Health Conversations Polished by ChatGPT. arXiv.
Not checked in the free run.
- cites [25] Luo, R., Sun, L., Xia, Y., Qin, T., Zhang, S., Poon, H., and Liu, T.Y. (2022). BioGPT: Generative pre-trained transformer for biomedical text generation and mining. Brief. Bioinform., 23. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationLogistic regression (LR) has been widely applied in disease diagnosis.
“LR has been widely applied in disease diagnosis.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [45] Ho, T.K. (1995, January 14–16). Random decision forests. Proceedings of the 3rd International Conference on Document Analysis and Recognition, Montreal, QC, Canada.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
Premisenone givenXrayGLM showcases remarkable potential in medical image diagnosis and multi-turn interactive dialogues.
“XrayGLM aims to become the first Chinese multi-modal medical LM proficient in interpreting chest X-ray images, showcasing remarkable potential in medical image diagnosis and multi-turn interactive dialogues.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.
Interpretationnone givenBy preprocessing raw data, inaccurate or irrelevant information can be removed and key features relevant to disease diagnosis are extracted.
“By preprocessing raw data, inaccurate or irrelevant information can be removed and key features relevant to disease diagnosis are extracted.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Interpretationnone givenFeature fusion can enhance the efficiency of classifiers in detection tasks.
“Feature fusion can enhance the efficiency of classifiers in detection tasks.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.
Premisenone givenRedundant features can confuse machine learning models, while few features might not effectively and correctly classify data.
“The redundant features can confuse machine learning models, while few features might not effectively and correctly classify data.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationLarge models targeting different data types, such as images and time series, have been developed and demonstrate powerful performance.
“Alongside technological advancements, large models targeting different data types, such as images (SAM [66]) and time series (TimeGPT-1 [67]), have also been developed, demonstrating their powerful performance.”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [66] Zhang, C., Liu, L., Cui, Y., Huang, G., Lin, W., Yang, Y., and Hu, Y. (2023). A Comprehensive Survey on Segment Anything Model for Vision and Beyond. arXiv.
Not checked in the free run.
- cites [67] Garza, A., and Mergenthaler-Canseco, M. (2023). TimeGPT-1. arXiv.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationSupport vector machine (SVM) is a commonly employed algorithm in disease diagnosis.
“The SVM [9] is another commonly employed algorithm in disease diagnosis [42,43,44].”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites [42] Shah, Z., Qi, S.A., Wang, F., Farrokh, M., Tasnim, M., Stroulia, E., Greiner, R., Plitsis, M., and Katsamanis, A. (2023, January 4–6). Exploring Language-Agnostic Speech Representations Using Domain Knowledge for Detecting Alzheimer’s Dementia. Proceedings of the ICASSP 2023—2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Rhode Island, Greece. source
Not checked in the free run.
- cites [43] Martinc, 2021, Temporal integration of text transcripts and acoustic features for Alzheimer’s diagnosis based on spontaneous speech, Front. Aging Neurosci. source
Not checked in the free run.
- cites [44] Zhang, 2023, Integration of machine learning to identify diagnostic genes in leukocytes for acute myocardial infarction patients, J. Transl. Med. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationAgbavor et al. achieved F1 0.728 and RMSE 3.493 on ADReSSo using data2vec and wav2vec2.
“| Agbavor et al. [84] | data2vec, wav2vec2 | Speech, text | ADReSSo | F1: 0.728, RMSE: 3.493 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [84] Agbavor, F., and Liang, H. (2022). Artificial Intelligence-Enabled End-To-End Detection and Assessment of Alzheimer’s Disease Using Voice. Brain Sci., 13. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationPark et al. achieved 0.823 accuracy on GEO using multi-omics data.
“| Park et al. [88] | Multi-omics data | Gene-based | GEO | 0.823 (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [88] Park, 2020, Prediction of Alzheimer’s disease based on deep neural network by integrating gene expression and DNA methylation dataset, Expert Syst. Appl. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationLopez et al. achieved accuracy 0.89 on a private dataset using linear and nonlinear speech features.
“| Lopez et al. [55] | Linear features include spectral domain features and time domain features, such as harmonicity, spectrum centroid, formants, etc. Nonlinear characteristics include fractal dimension, permutation entropy, multi-scale permutation entropy, etc. | speech | private dataset | Accuracy: 0.89 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [55] Calvo, 2018, Advances on automatic speech analysis for early detection of Alzheimer disease: A non-linear multi-task approach, Curr. Alzheimer Res. source
Not checked in the free run.
Classified as probably not a substantial claim. May combine two claims.
InterpretationcitationGPT-4 could enhance the accuracy and efficiency of diagnostic procedures by automating extraction of key data points, supporting healthcare professionals in making informed decisions.
“By automating the extraction of key data points, GPT-4 could enhance the accuracy and efficiency of diagnostic procedures, supporting healthcare professionals in making informed decisions.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationHuman cancer is a heterogeneous disease caused by stochastic cellular mutations and driven by various genomic alterations.
“Human cancer is a heterogeneous disease caused by stochastic cellular mutations and driven by various genomic alterations [107,108].”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [107] Hanahan, 2011, Hallmarks of cancer: The next generation, Cell source
Not checked in the free run.
- cites [108] Sun, 2017, Computational approach for deriving cancer progression roadmaps from static sample data, Nucleic Acids Res.
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationCai et al. achieved accuracy 0.8484 ± 0.0544 on Dementia-Bank Pit using GNN text features and WavLM voice features.
“| Cai et al. [82] | GNN (text features) WavLM (voice features) | Speech, text | Dementia- Bank Pit | Accuracy: 0.8484 ± 0.0544 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [82] Cai, H., Huang, X., Liu, Z., Liao, W., Dai, H., Wu, Z., Zhu, D., Ren, H., Li, Q., and Liu, T. (2023). Exploring Multimodal Approaches for Alzheimer’s Disease Detection Using Patient Speech Transcript and Audio Data. arXiv. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationLiu et al. achieved F1 0.7802 on Dementia-Bank Pitt using bottleneck feature vector.
“| Liu et al. [75] | Bottleneck feature vector (depth representation feature) | speech | Dementia- Bank Pitt | F1: 0.7802 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [75] Liu, Z., Guo, Z., Ling, Z., and Li, Y. (2021, January 6–11). Detecting Alzheimer’s disease from speech using neural networks with bottleneck features and data augmentation. Proceedings of the ICASSP 2021—2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Toronto, ON, Canada. source
Not checked in the free run.
The quote only partly states this claim.
PremisecitationZhou et al. found that pre-training BERT on a cancer-specific dataset substantially improved model performance for extracting breast cancer phenotypes from pathology reports and clinical records.
“Their findings underscore the significance of domain-specific pre-training, as it substantially improved the performance of the model, making it more attuned to the nuances of cancer-related data.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [105] Zhou, 2022, CancerBERT: A cancer domain-specific language model for extracting breast cancer phenotypes from electronic health records, J. Am. Med Inform. Assoc. source
Not checked in the free run.
May combine two claims.
PremisecitationMohaiminul et al. [58] achieved an accuracy of 76.7% and an AUC of 83.8% for classifying molecular subtypes of breast cancer using copy number alteration and gene expression data from METABRIC.
“The results achieved an accuracy of 76.7% and an AUC of 83.8%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [58] Islam, 2020, An integrative deep learning framework for classifying molecular subtypes of breast cancer, Comput. Struct. Biotechnol. J. source
Not checked in the free run.
May combine two claims.
PremisecitationSherafatian et al.'s Rpart achieved accuracy 90.2%, sensitivity 93.9%, and specificity 86.4% for the HER2-enriched subtype.
“For the HER2-enriched subtype, the accuracy, sensitivity, and specificity were 90.2%, 93.9%, and 86.4%, respectively.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [50] Sherafatian, 2018, Tree-based machine learning algorithms identified minimal set of miRNA biomarkers for breast cancer diagnosis and molecular subtyping, Gene source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationKim et al. achieved 0.75 accuracy on a private dataset using EEG segment with respect to RP.
“| Kim et al. [92] | EEG segment with respect to RP(Absolute power of EEG signals in three different frequency bands) | EEG | Private dataset | 0.75 (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [92] Kim, D., and Kim, K. (2018, January 18–21). Detection of early stage Alzheimer’s disease using EEG relative power with deep neural network. Proceedings of the 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Honolulu, HI, USA. source
Not checked in the free run.
May combine two claims.
PremisecitationSharma et al. achieved 0.9603 accuracy for AD vs. NC vs. MCI and 0.7521 AUC on ADNI using fused image by wavelet packet transform (WPT) from MRI and PET.
“| Sharma et al. [80] | Fused image by wavelet packet transform (WPT) | MRI, PET | ADNI | 0.9603 (ACC: AD vs. NC vs. MCI) 0.7521 (AUC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [80] Sharma, 2022, Conv-ERVFL: Convolutional Neural Network Based Ensemble RVFL Classifier for Alzheimer’s Disease Diagnosis, IEEE J. Biomed. Health Inform. source
Not checked in the free run.
May combine two claims.
PremisecitationBertini et al. achieved accuracy 0.933 and F1 0.885 on Dementia-Bank Pitt using spectrogram.
“| Bertini et al. [76] | spectrogram | specch | Dementia- Bank Pitt | Accuracy is 0.933, F1 score is 0.885 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [76] Bertini, 2022, An automatic Alzheimer’s disease classifier based on spontaneous spoken English, Comput. Speech Lang. source
Not checked in the free run.
May combine two claims.
PremisecitationKumar et al.'s BERT-based model achieved F1 scores of 0.8625 on the development set and 0.8501 on the test set for classifying breast cancer-related posts on Twitter.
“ultimately resulted in the model achieving F1 scores of 0.8625 on the development set and 0.8501 on the test set”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [102] Kumar, 2021, Phoenix@ SMM4H Task-8: Adversities Make Ordinary Models Do Extraordinary Things, NAACL-HLT
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationLi et al. achieved 0.859 AUC and 0.781 accuracy on GEO using RNA-seq gene-based data.
“| Li et al. [84] | RNA-seq | Gene-based | GEO | 0.859 (AUC), 0.781 (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [84] Agbavor, F., and Liang, H. (2022). Artificial Intelligence-Enabled End-To-End Detection and Assessment of Alzheimer’s Disease Using Voice. Brain Sci., 13. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationTaeho et al. achieved 0.82 AAUC on ADNI using SNP gene-based data.
“| Taeho et al. [85] | SNP | Gene-based | ADNI | 0.82 (AAUC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [85] Jo, T., Nho, K., Bice, P., Saykin, A.J., and Initiative, A.D.N. (2022). Deep learning-based identification of genetic variants: Application to Alzheimer’s disease classification. Brief. Bioinform., 23. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationShah et al. achieved accuracy 0.696 and RMSE 4.8 on ADReSS-M using word-level duration, pause rate, and speech intelligibility features from speech and text.
“| Shah et al. [42] | Word-level duration feature set, pause rate data set, speech intelligibility feature set | speech, text | ADReSS-M | Accuracy: 0.696, RMSE: 4.8 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [42] Shah, Z., Qi, S.A., Wang, F., Farrokh, M., Tasnim, M., Stroulia, E., Greiner, R., Plitsis, M., and Katsamanis, A. (2023, January 4–6). Exploring Language-Agnostic Speech Representations Using Domain Knowledge for Detecting Alzheimer’s Dementia. Proceedings of the ICASSP 2023—2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Rhode Island, Greece. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationJavier et al. achieved 0.719 accuracy on ADNI using genetic variation data.
“| Javier et al. [87] | genetic variation data | Gene-based | ADNI | 0.719 (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [87] De Velasco Oriol, J., Vallejo, E.E., Estrada, K., Taméz Peña, J.G., and Disease Neuroimaging Initiative, T.A. (2019). Benchmarking machine learning models for late-onset alzheimer’s disease prediction from genomic data. BMC Bioinform., 20. source
Not checked in the free run.
May combine two claims.
PremisecitationNing et al. achieved 0.976 AUC and 0.969 accuracy for AD vs. NC on ADNI using MRI and PET.
“| Ning et al. [72] | magnetic resonance imaging (MRI) and positron emission tomography (PET) | MRI, PET | ADNI | 0.976 (AUC: AD vs. NC) 0.969 (ACC: AD vs. NC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [72] Ning, 2021, Relation-induced multi-modal shared representation learning for Alzheimer’s disease diagnosis, IEEE Trans. Med. Imaging source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationHernández et al. achieved accuracy 0.8 on a private dataset using speech duration and descriptive statistical variables.
“| Hernández et al. [73] | Speech duration, descriptive statistical variables | specch | private dataset | Accuracy: 0.8 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [73] Pulido, 2022, Speech evaluation of patients with alzheimer’s disease using an automatic interviewer, Expert Syst. Appl. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationXu et al. achieved 0.857 accuracy on UniProt using protein sequence gene-based data.
“| Xu et al. [86] | protein | sequence Gene-based | UniProt | 0.857 (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [86] Xu, L., Liang, G., Liao, C., Chen, G.D., and Chang, C.C. (2018). An efficient classifier for Alzheimer’s disease genes identification. Molecules, 23. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationHason et al. achieved accuracy 0.822 on ADReSS using MFCC speech features.
“| Hason et al. [72] | MFCC | speech | ADReSS | Accuracy: 0.822 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [72] Ning, 2021, Relation-induced multi-modal shared representation learning for Alzheimer’s disease diagnosis, IEEE Trans. Med. Imaging source
Not checked in the free run.
May combine two claims.
PremisecitationRiyadh et al.'s mixed deep learning CAD system achieved 100% accuracy for binary classification and 95.80% for multi-class prediction of breast lesions.
“This innovative model achieved a perfect 100% accuracy rate for binary classification and an impressive 95.80% for multi-class prediction tasks”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [97] Al-Tam, R.M., Al-Hejri, A.M., Narangale, S.M., Samee, N.A., Mahmoud, N.F., Al-Masni, M.A., and Al-Antari, M.A. (2022). A hybrid workflow of residual convolutional transformer encoder for breast cancer classification using digital X-ray mammograms. Biomedicines, 10. source
Not checked in the free run.
May combine two claims.
InterpretationreasoningAmalgamating speech and text data captures a more extensive spectrum of AD-related features and bolsters diagnostic accuracy for Alzheimer's disease.
“By amalgamating speech and text data, a more extensive spectrum of AD-related features is captured, bolstering the diagnostic accuracy for this condition.”
Quote in the paperSubstantial claim- Rests on an assumption: Changes in language and communication frequently serve as precursors to cognitive deterioration, making them significant markers for early detection of Alzheimer’s disease.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationDeng et al. emphasized GPT-4's capability to rapidly mine crucial information from extensive medical records, which could potentially influence the diagnosis of breast cancer.
“The authors emphasized GPT-4’s capability to rapidly mine crucial information from extensive medical records, which could potentially influence the diagnosis of breast cancer.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [106] Deng, 2023, GPT-4 in breast cancer combat: A dazzling leap forward or merely a whim?, Int. J. Surg. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationYeman et al.'s parallel deep CNN for bilateral breast X-ray images achieved a prediction accuracy rate of 93.33%.
“The concatenated deep features from this twin-CNN setup were introduced into a Neural Network (NN) classifier, which achieved a remarkable prediction accuracy rate of 93.33%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [96] Brhane Hagos, Y., Gubern Mérida, A., and Teuwen, J. (2018, January 16). Improving breast cancer detection using symmetry information with deep learning. Proceedings of the International Workshop on Reconstruction and Analysis of Moving Body Organs, Québec City, QC, Canada. source
Not checked in the free run.
May combine two claims.
PremisecitationSherafatian et al.'s Rpart achieved accuracy 88.9%, sensitivity 82.4%, and specificity 95.4% for the Luminal subtype.
“For the Luminal subtype, the accuracy, sensitivity, and specificity were 88.9%, 82.4%, and 95.4%, respectively.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [50] Sherafatian, 2018, Tree-based machine learning algorithms identified minimal set of miRNA biomarkers for breast cancer diagnosis and molecular subtyping, Gene source
Not checked in the free run.
May combine two claims.
PremisecitationAbunasser et al. trained six deep learning models for classifying eight types of breast cancer on MRI, with accuracy rates of 97.54%, 95.33%, 98.14%, 97.67%, 93.98%, and 98.28% respectively.
“These models demonstrated remarkable accuracy in their classification tasks, with rates of 97.54%, 95.33%, 98.14%, 97.67%, 93.98%, and 98.28% respectively”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [98] Abunasser, 2023, Convolution Neural Network for Breast Cancer Detection and Classification Using Deep Learning, Asian Pac. J. Cancer Prev. APJCP source
Not checked in the free run.
May combine two claims.
PremisecitationDai et al. [151] achieved an accuracy of 99.84%, sensitivity of 99.52%, and specificity of 99.95% for three-second ECG signals.
“For three-second signal detection, the accuracies were 99.84%, sensitivity 99.52%, and specificity 99.95%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [151] Dai, H., Hwang, H.G., and Tseng, V.S. (2021). Convolutional neural network based automatic screening tool for cardiovascular diseases using different intervals of ECG signals. Comput. Methods Programs Biomed., 203. source
Not checked in the free run.
May combine two claims.
PremisecitationHCM Dynamic Echo achieved an accuracy of 93.13%, an F1 score of 92.98%, a PPV of 94.64%, a specificity of 94.87%, and an AUC of 93.13% for binary classification of echocardiography videos into HCM or normal.
“In rigorous evaluations, HCM Dynamic Echo outperformed state-of-the-art baselines, with an accuracy of 93.13%, an F1 score of 92.98%, a Positive Predictive Value (PPV) of 94.64%, a specificity of 94.87%, and an Area Under the Curve (AUC) of 93.13%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [138] Almadani, A., Agu, E., Sarwar, A., Ahluwalia, M., and Kpodonu, J. (2023, January 7–13). HCM-Dynamic-Echo: A Framework for Detecting Hypertrophic Cardiomyopathy (HCM) in Echocardiograms. Proceedings of the 2023 IEEE International Conference on Digital Health (ICDH), Shenzhen, China. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationGhorbani et al. [142] achieved an AUC-ROC of 0.75 and an F1 score of 0.57.
“This method achieved an AUC-ROC of 0.75 and an F1 score of 0.57.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [142] Ghorbani, 2020, Deep learning interpretation of echocardiograms, NPJ Digit. Med. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationZhang et al. [141] achieved an AUC of 93% and a p-value of 0.23 for HCM detection.
“The images were processed through the VGG-16 model with a fully connected layer featuring two output units, achieving an AUC of 93% and p-value of 0.23 for HCM detection.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [141] Zhang, 2018, Fully automated echocardiogram interpretation in clinical practice: Feasibility and diagnostic accuracy, Circulation source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationAbunasser et al. [98] achieved an F1-score of 98.28.
“| Abunasser et al. [98] | Original MRI data | MRI | Kaggle depository | 98.28 (F1-score) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [98] Abunasser, 2023, Convolution Neural Network for Breast Cancer Detection and Classification Using Deep Learning, Asian Pac. J. Cancer Prev. APJCP source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationHuang et al. [99] achieved an AUC of 0.907 and an accuracy of 85.8% with a Multilayer Perceptron (MLP).
“| Huang et al. [99] | multi-parametric MRI | MRI | Private dataset | Multilayer Perceptron (MLP): 0.907 (AUC) and 85.8% (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [99] Huang, 2021, Multi-parametric MRI-based radiomics models for predicting molecular subtype and androgen receptor expression in breast cancer, Front. Oncol. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationSun et al. [108] achieved an AUC of 82% on METABRIC.
“| Sun et al. [108] | Clinical, CNV, RNA-seq | Gene-based | METABRIC | 82% (AUC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [108] Sun, 2017, Computational approach for deriving cancer progression roadmaps from static sample data, Nucleic Acids Res.
Not checked in the free run.
Classified as probably not a substantial claim.
PremisecitationAuditory anomalies, repetitive speech patterns, word retrieval struggles, coherence degradation, and variance in lexical density and richness are pivotal in distinguishing dementia and depression.
“They pinpointed a spectrum of differentiators such as auditory anomalies, repetitive speech patterns, word retrieval struggles, coherence degradation, and variance in lexical density and richness—all of which are pivotal in distinguishing these disorders.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [132] Ehghaghi, M., Rudzicz, F., and Novikova, J. (2022). Data-driven Approach to Differentiating between Depression and Dementia from Noisy Speech and Language Data. arXiv.
Not checked in the free run.
May combine two claims.
PremisecitationDai et al. [151] achieved an accuracy of 99.59%, sensitivity of 99.04%, and specificity of 99.87% for one-second ECG signals.
“The proposed CNN model achieved an accuracy, sensitivity, and specificity of 99.59%, 99.04%, and 99.87%, respectively, for one-second signals, demonstrating superior performance.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [151] Dai, H., Hwang, H.G., and Tseng, V.S. (2021). Convolutional neural network based automatic screening tool for cardiovascular diseases using different intervals of ECG signals. Comput. Methods Programs Biomed., 203. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationSundaram et al. [143] achieved an accuracy of 86% with 608 features and 85% with 30 features.
“The Random Forest (RF) model attained an accuracy of 86% using 608 features and achieved 85% accuracy with 30 features.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [143] Sundaram, 2021, Natural language processing based machine learning model using cardiac MRI reports to identify hypertrophic cardiomyopathy patients, Frontiers in Biomedical Devices
Not checked in the free run.
May combine two claims.
PremisecitationJayasudha et al. [145] achieved specificity of 95.84%, accuracy of 94.80%, and sensitivity of 95.36%.
“The classifier showed outstanding performance across specificity, accuracy, and sensitivity, reaching 95.84%, 94.80 and 95.36% in each metric.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [145] Jayasudha, 2023, Hybrid optimization enabled deep learning-based ensemble classification for heart disease detection, Signal Image Video Process. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationAI-ECG can extract more physiologically and pathophysiologically relevant information related to obstructive HCM from ECGs than traditional manual interpretation methods.
“The study noted that AI-ECG could extract more physiologically and pathophysiologically relevant information related to obstructive HCM from ECGs, surpassing traditional manual interpretation methods.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [153] Tison, 2022, Assessment of disease status and treatment response with artificial Intelligence- Enhanced electrocardiography in obstructive hypertrophic cardiomyopathy, J. Am. Coll. Cardiol. source
Not checked in the free run.
May combine two claims.
PremisecitationSamadishadlou et al. [150] achieved an AUC-ROC of 96% and an accuracy of 94%.
“Their model demonstrated an AUC-ROC of 96% and an accuracy of 94%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [150] Samadishadlou, 2023, Unlocking the potential of microRNAs: Machine learning identifies key biomarkers for myocardial infarction diagnosis, Cardiovasc. Diabetol. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationVerma et al. [113] found that the Random Forest model achieved a peak accuracy of 78% for depression detection from tweets.
“Of these, the Random Forest model demonstrated superior performance, achieving an impressive accuracy peak of 78%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [113] Verma, P., Sharma, K., and Walia, G.S. (2021, January 20–22). Depression detection among social media users using machine learning. Proceedings of the International Conference on Innovative Computing and Communications: Proceedings of ICICC 2020, Delhi, India. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationRagab et al. [101] achieved an accuracy of 96.92%.
“| Ragab et al. [101] | Segmented regions from original | ultrasound images Ultrasound Images | - | 96.92% (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [101] Ragab, M., Albukhari, A., Alyami, J., and Mansour, R.F. (2022). Ensemble deep-learning-enabled clinical decision support system for breast cancer diagnosis and classification on ultrasound images. Biology, 11. source
Not checked in the free run.
May combine two claims.
PremisecitationDai et al. [151] achieved an overall accuracy of 99.80%, sensitivity of 99.48%, and specificity of 99.93% for two-second ECG signals.
“For two-second signals using pre-trained models, the system achieved an overall accuracy, sensitivity, and specificity of 99.80%, 99.48%, and 99.93%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [151] Dai, H., Hwang, H.G., and Tseng, V.S. (2021). Convolutional neural network based automatic screening tool for cardiovascular diseases using different intervals of ECG signals. Comput. Methods Programs Biomed., 203. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationRiyadh et al. [97] achieved an AUC of 0.933.
“| Riyadh et al. [97] | Extracted patches centered on the points from the original X-ray | X-ray | General Electric, Siemens, and Hologic | 0.933 (AUC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [97] Al-Tam, R.M., Al-Hejri, A.M., Narangale, S.M., Samee, N.A., Mahmoud, N.F., Al-Masni, M.A., and Al-Antari, M.A. (2022). A hybrid workflow of residual convolutional transformer encoder for breast cancer classification using digital X-ray mammograms. Biomedicines, 10. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationZhou et al. [105] achieved macro-F1 of 0.876 for exact match and 0.904 for lenient match.
“| Zhou et al. [105] | mutil feature | Text | private dataset | exact match and lenient match, macro-F1: 0.876, 0.904 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [105] Zhou, 2022, CancerBERT: A cancer domain-specific language model for extracting breast cancer phenotypes from electronic health records, J. Am. Med Inform. Assoc. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationChen et al. [104] achieved Mi-P, Mi-R, and Mi-F1 of 91.58% and Ma-P 75.95%, Ma-R 79.73%, and Ma-F1 77.14%.
“| Chen et al. [104] | Word embedding, syntactic structure | Text | Shanghai Ruijin Hospital Molybdenum Mammography X-ray Report | Mi-P(%) = 91.58 Mi-R(%) = 91.58 Mi-F1(%) = 91.58 Ma-P(%) = 75.95 Ma-R(%) = 79.73 Ma-F1(%) = 77.14 |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [104] Chen, D., Zhong, K., and He, J. (2021, January 13–15). BDCN: Semantic Embedding Self-Explanatory Breast Diagnostic Capsules Network. Proceedings of the China National Conference on Chinese Computational Linguistics, Hohhot, China. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationJabeen et al. [100] achieved an accuracy of 99.1% on the BUSI dataset.
“| Jabeen et al. [100] | Original ultrasound images data | Ultrasound Images | BUSI dataset | 99.1% (ACC) |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [100] Jabeen, K., Khan, M.A., Alhaisoni, M., Tariq, U., Zhang, Y.D., Hamza, A., Mickus, A., and Damaševičius, R. (2022). Breast cancer classification from ultrasound images using probability-based optimal deep learning feature fusion. Sensors, 22. source
Not checked in the free run.
May combine two claims.
PremisecitationPeng et al. [147] achieved AUCs of 81.3% for SVM, 72.7% for RF, and 78.3% for LR for coronary atherosclerosis heart disease classification.
“Their respective areas under the curve were 81.3% (95% CI 0.761–0.866, p < 0.0001), 72.7% (95% CI 0.665–0.788, p < 0.0001), and 78.3% (95% CI 0.725–0.841, p < 0.0001).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [147] Peng, 2022, Construction of genetic classification model for coronary atherosclerosis heart disease using three machine learning methods, BMC Cardiovasc. Disord. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationKumar et al. [102] and Peng et al. [103] achieved an F1 of 0.8501.
“| Kumar et al. [102], Peng et al. [103] | Word embedding | Text | witter self-report | F1: 0.8501 |”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [102] Kumar, 2021, Phoenix@ SMM4H Task-8: Adversities Make Ordinary Models Do Extraordinary Things, NAACL-HLT
Not checked in the free run.
- cites [103] Peng, Y., Yan, S., and Lu, Z. (2019). Transfer learning in biomedical natural language processing: An evaluation of BERT and ELMo on ten benchmarking datasets. arXiv. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationYeman et al. [96] achieved accuracies of 94.50%, 95.83%, and 97.50% for three models on DDSM and INbreast.
“| Yeman et al. [96] | Breast lesion detection from entire mammograms by object detection model | X-ray | DDSM and INbreast | ACC of three models: 94.50%, 95.83%, and 97.50% |”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [96] Brhane Hagos, Y., Gubern Mérida, A., and Teuwen, J. (2018, January 16). Improving breast cancer detection using symmetry information with deep learning. Proceedings of the International Workshop on Reconstruction and Analysis of Moving Body Organs, Québec City, QC, Canada. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationLevine et al. [146] found that GPT-3's performance in diagnosing and triaging heart disease nearly approached that of professional medical practitioners.
“The findings indicated that GPT-3’s performance nearly approached that of professional medical practitioners.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [146] Levine, D.M., Tuwani, R., Kompa, B., Varma, A., Finlayson, S.G., Mehrotra, A., and Beam, A. (2023). The diagnostic and triage accuracy of the GPT-3 artificial intelligence model. medRxiv, 1–22. source
Not checked in the free run.
May combine two claims.
InterpretationcitationCentralized solutions come with inherent drawbacks, including concerns about data ownership, confidentiality, privacy, and security, as well as the potential for data monopolies biased towards data aggregators.
“However, centralized solutions come with inherent drawbacks, including concerns about data ownership, confidentiality, privacy, and security, as well as the potential for data monopolies biased towards data aggregators [177].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [177] Kaissis, 2020, Secure, privacy-preserving and federated machine learning in medical imaging, Nat. Mach. Intell. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningFor diagnosing heart disease, it is common to combine ultrasound imaging with medical texts.
“For diagnosing heart disease, it’s common to combine ultrasound imaging with medical texts.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningTransfer learning has proven effective in scenarios where acquiring a large, labeled medical dataset is impractical.
“Transfer learning has proven effective in scenarios where acquiring a large, labeled medical dataset is impractical, thus facilitating the development of robust diagnostic models.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationThe Mayo Clinic algorithm achieved an AUC of 0.96, sensitivity of 87%, and specificity of 90% for adult HCM detection.
“The algorithm performed impressively in adult HCM patient ECG detection, with an AUC of 0.96, sensitivity of 87%, and specificity of 90%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [154] Ko, 2020, Detection of hypertrophic cardiomyopathy using a convolutional neural network-enabled electrocardiogram, J. Am. Coll. Cardiol. source
Not checked in the free run.
May combine two claims.
PremisecitationThe Mayo Clinic algorithm achieved an AUC of 0.98, sensitivity of 92%, specificity of 95%, PPV of 22%, and NPV of 99% in a test of 300 children and over 18,000 controls.
“The algorithm’s performance in a test of 300 children and over 18,000 age and gender-matched controls was equally impressive: the HCM detection model achieved an AUC of 0.98, sensitivity of 92%, specificity of 95%, Positive Predictive Value (PPV) of 22%, and Negative Predictive Value (NPV) of 99%.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [154] Ko, 2020, Detection of hypertrophic cardiomyopathy using a convolutional neural network-enabled electrocardiogram, J. Am. Coll. Cardiol. source
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable).
Interpretationnone givenThe multi-modal paradigm provides a more comprehensive and accurate understanding of complex phenomena by integrating diverse data modalities.
“The advantages of the multi-modal paradigm lie in its ability to provide a more comprehensive and accurate understanding of complex phenomena by integrating diverse data modalities.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningThe progression has transitioned from machine learning dominance to primary reliance on deep learning, complemented by attention mechanisms and transfer learning.
“This progression has transitioned from machine learning dominance to primary reliance on deep learning, complemented by innovative techniques such as attention mechanisms and transfer learning.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationThara et al. [172] achieved an accuracy of 99.89% on the training set and 99.08% on the test set.
“The researchers built a model with two LSTM layers, dropout and dense layers, and trained and optimized it using activation functions such as sigmoid and softmax, achieving good results with an accuracy of 99.89% on the training set and 99.08% on the test set.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [172] Thara, 2019, Epileptic seizure detection and prediction using stacked bidirectional long short term memory, Pattern Recognit. Lett. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningFor heart disease diagnosis, the prevalent modalities include echocardiography, electrocardiography, and medical texts.
“For heart disease diagnosis, the prevalent modalities include echocardiography, electrocardiography, and medical texts.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationFürbass et al. [171]'s DeepSpike algorithm attained sensitivity of 89%, specificity of 70%, and overall accuracy of 80%.
“The experimental results indicated that the DeepSpike algorithm attained a sensitivity of 89%, a specificity of 70%, and an overall accuracy rate of 80%, showcasing its high effectiveness in identifying EEG discharges.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [171] Kural, 2020, An artificial intelligence-based EEG algorithm for detection of epileptiform EEG discharges: Validation against the diagnostic gold standard, Clin. Neurophysiol. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Interpretationnone givenCurrently, most large-scale models in healthcare focus on text, analyzing medical records, discharge summaries, and other types of written data.
“Currently, most large-scale models in healthcare focus on text, analyzing medical records, discharge summaries, and other types of written data.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningAttention mechanisms have shown effectiveness in different medical imaging tasks, leading to more precise and context-aware diagnoses.
“Attention mechanisms have shown effectiveness in different medical imaging tasks, leading to diagnoses that are more precise and aware of the context.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Interpretationnone givenThe application of AI in the health and medical field involves three types of parties: healthcare professionals, researchers, and AI experts.
“The application of AI in the health and medical field involves three types of parties, i.e., healthcare professionals, researchers, and AI experts.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningThe evolution of research in automated disease diagnosis is accompanied by the continual improvement of performance.
“The evolution of research in automated disease diagnosis is accompanied by the continual improvement of performance.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
Interpretationnone givenThe medical data produced from different disease diagnosis processes has commonalities, mainly encompassing image, text, genetic, signal, and voice modalities.
“The medical data produced from different disease diagnosis processes has commonalities, mainly encompassing image, text, genetic, signal, and voice modalities.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
InterpretationreasoningFacilitating collaboration among these three parties contributes to the advancement of AI in the health and medical domain.
“Facilitating collaboration among these three parties contributes to the advancement of AI in the health and medical domain.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningIn breast cancer diagnostics, a multifaceted approach involves genetic data, X-ray imaging, ultrasound, and a notable amount of textual information.
“Contrastingly, in breast cancer diagnostics, a multifaceted approach involves genetic data, X-ray imaging, ultrasound, and a notable amount of textual information.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.
InterpretationreasoningDepression diagnosis predominantly relies on speech modalities, with supplementary integration of text or video data.
“Depression diagnosis predominantly relies on speech modalities, with supplementary integration of text or video data.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based). May combine two claims.
InterpretationreasoningThe multi-modal approach enhances robustness, improves interpretability, and allows for personalized and optimized solutions across various domains.
“This approach enhances robustness, improves interpretability, and allows for personalized and optimized solutions across various domains.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based). May combine two claims.
Interpretationnone givenDeep learning has improved the accuracy and efficiency of disease detection.
“Deep learning has improved the accuracy and efficiency of disease detection.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationKarácsony et al. [156] employed clinical Motion Capture to quantitatively analyze seizure-related symptoms and differentiate epilepsy syndromes, providing clinical localization and lateralization information.
“Karácsony et al. [156] employed clinical Motion Capture (MoCap) to quantitatively analyze seizure-related symptoms such as ictal head turning and upper limb automatisms, marking a pioneering discovery in differentiating epilepsy syndromes, providing clinical localization and lateralization information.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [156] Karácsony, T., Loesch-Biffar, A.M., Vollmar, C., Noachtar, S., and Cunha, J.P.S. (2021, January 27–30). DeepEpil: Towards an epileptologist-friendly AI enabled seizure classification cloud system based on deep learning analysis of 3D videos. Proceedings of the 2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI), Athens, Greece. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.
Premisenone givenData imbalance encompasses two dimensions: imbalance within classes in a single modality, and distributional imbalances across different modalities.
“Typically, data imbalance encompasses two dimensions: the imbalance within classes in a single modality, and the distributional imbalances across different modalities.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim. May combine two claims.
InterpretationreasoningWhile AI technology has made certain achievements in the medical field, significant limitations remain in disease diagnosis applications.
“Finally, we acknowledge that while AI technology has made certain achievements in the medical field, significant limitations remain in disease diagnosis applications.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
InterpretationreasoningFor epilepsy diagnostics, a comprehensive strategy incorporates MRI, video data capturing patient movements, EEG, and relevant textual information.
“For epilepsy diagnostics, a comprehensive strategy incorporates Magnetic Resonance Imaging (MRI), video data capturing patient movements, Electroencephalogram (EEG), and relevant textual information.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningIn cancer research, there is a significant emphasis on combining imaging and genetic data.
“In cancer research, there is a significant emphasis on combining imaging and genetic data.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationClarke et al. [170] achieved an average sensitivity of 95% and an average false positive rate of one per minute for detecting epileptic seizures from clinical EEGs.
“The system achieved an impressive average sensitivity of 95% and kept the average false positive rate to just one per minute.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [170] Clarke, 2021, Computer-assisted EEG diagnostic review for idiopathic generalized epilepsy, Epilepsy Behav. source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
InterpretationreasoningEpilepsy diagnosis currently benefits from the mutual utilization of various imaging modalities, such as MRI and PET images.
“Epilepsy diagnosis currently benefits from the mutual utilization of various imaging modalities, such as Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) images.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationPDC alone provided the most accurate TLE classification, followed by the combination of PDC with MRI-based brain structural indices.
“Results indicated that PDC alone provided the most accurate TLE classification, followed by the combination of PDC with MRI-based brain structural indices.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [174] Torres-Velázquez, M., Hwang, G., Cook, C.J., Hermann, B.P., Prabhakaran, V., Meyerand, M.E., and McMillan, A.B. (2020, January 4). Multi-Channel Deep Neural Network For Temporal Lobe Epilepsy Classification Using Multimodal Mri Data. Proceedings of the 2020 IEEE 17th International Symposium on Biomedical Imaging Workshops (ISBI Workshops), Iowa City, IA, USA. source
Not checked in the free run.
The paper relies on this without support.
Achieving timely diagnosis remains challenging, particularly in developing areas and regions with limited medical resources.
“Nonetheless, achieving this objective remains challenging, particularly in developing areas and regions with limited medical resources.”
The paper relies on this without support.
The challenge of disease diagnosis primarily stems from the low specialist-to-patient ratio and the time-consuming, labor-intensive nature of manual diagnosis, which heavily relies on specialized expertise.
“This challenge primarily stems from two key factors: firstly, the low specialist-to-patient ratio, and secondly, the time-consuming and labor-intensive nature of the manual diagnosis, which heavily relies on specialized expertise.”
The paper relies on this without support.
These issues often result in delayed treatment, exacerbating illness severity, and escalating medical costs.
“These issues often result in delayed treatment, exacerbating illness severity, and escalating medical costs.”
The paper relies on this without support.
Existing reviews predominantly concentrate on a singular modality or a single disease, whether focusing on a specific disease within multi-modal contexts, various disorders within a specific modality, or a single disease with exclusive reliance on a particular data type.
“However, these reviews predominantly concentrate on a singular modality or a single disease, whether focusing on a specific disease within multi-modal contexts, various disorders within a specific modality, or a single disease with exclusive reliance on a particular data type.”
The paper relies on this without support.
Although medical datasets generated by different disease diagnosis processes exhibit commonalities, distinct preferences for specific modalities prevail across different diseases.
“Although medical datasets generated by different disease diagnosis processes exhibit commonalities, distinct preferences for specific modalities prevail across different diseases.”
The paper relies on this without support.
Most diseases are typically only recognized by patients themselves after they manifest, and continuous data collection and monitoring can assist patients in achieving effective disease prevention.
“Most diseases are typically only recognized by patients themselves after they manifest, and continuous data collection and monitoring can assist patients in achieving effective disease prevention.”
The paper relies on this without support.
The advent of artificial intelligence has rendered the process of data accumulation more intelligent and efficient, thereby holding significant implications for disease prevention and control.
“The advent of artificial intelligence has rendered the process of data accumulation more intelligent and efficient, thereby holding significant implications for disease prevention and control.”
The paper relies on this without support.
MRI and PET are indispensable tools for Alzheimer’s disease diagnosis.
“In the realm of AI-driven diagnostic investigations for Alzheimer’s disease, medical imaging modalities such as MRI and PET are universally recognized as indispensable tools.”
The paper relies on this without support.
Changes in language and communication frequently serve as precursors to cognitive deterioration, making them significant markers for early detection of Alzheimer’s disease.
“Changes in language and communication frequently serve as precursors to cognitive deterioration, making them significant markers for early detection.”
The paper relies on this without support.
The integration of MRI and PET imaging modalities yields a synergistic approach that offers a more holistic representation of the brain’s structure and function.
“The integration of MRI and PET imaging modalities has yielded a synergistic approach in medical diagnostics, particularly for disorders such as Alzheimer’s disease (AD). This technique of image fusion leverages the unique strengths of each imaging method to offer a more holistic representation of the brain’s structure and function.”
The paper relies on this without support.
Early detection and an integrated approach to treatment improve survival and quality of life for breast cancer patients.
“The key to enhancing survival and ensuring a better quality of life for those impacted by this disease lies in early detection and an integrated approach to treatment”
The paper relies on this without support.
Specific genetic mutations may present as unique visual patterns.
“Besides, specific genetic mutations may present as unique visual patterns.”
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Reference 72 is attributed to two different studies
MinorThe text attributes reference 72 to Hason et al. for a speech study, while the reference list identifies it as Ning et al.'s multi-modal representation-learning study.
“Studies by Hason et al. [72], Hernández et al. [73], and Yu et al. [74] are examples of such research efforts.”
“72. Ning Z. Xiao Q. Feng Q. Chen W. Zhang Y. Relation-induced multi-modal shared representation learning for Alzheimer’s disease diagnosis”
Reference 84 is assigned to an RNA-seq study but describes a voice study
MinorTable 3 attributes an RNA-seq result to Li et al. [84]. Reference 84 is an Alzheimer's voice-analysis paper by Agbavor and Liang.
“| Li et al. [84] | RNA-seq | Gene-based | GEO | 0.859 (AUC), 0.781 (ACC) |”
“84. Agbavor F. Liang H. Artificial Intelligence-Enabled End-To-End Detection and Assessment of Alzheimer’s Disease Using Voice”
Reference 58 is attributed to a depression study but describes breast cancer
MinorThe depression section attributes reference 58 to Aragon et al.'s depression model. Reference 58 instead identifies Islam et al.'s breast-cancer-subtyping framework.
“Aragon et al. [58] introduced a sophisticated deep emotional attention model tailored for the detection of anorexia and depression.”
“58. Islam M.M. Huang S. Ajwad R. Chi C. Wang Y. Hu P. An integrative deep learning framework for classifying molecular subtypes of breast cancer”
Peng et al.'s AUCs are labeled as accuracies in Table 6
MinorTable 6 reports 81.3%, 72.7%, and 78.3% as accuracies. The discussion identifies those same three values as areas under the curve, not accuracies.
“| Peng et al. [147] | Gene-based | RNA-seq | GEO | SVM: 81.3% (ACC); RF: 72.7% (ACC); LR: 78.3% (ACC) |”
“Their respective areas under the curve were 81.3% (95% CI 0.761–0.866, p < 0.0001), 72.7% (95% CI 0.665–0.788, p < 0.0001), and 78.3% (95% CI 0.725–0.841, p < 0.0001).”
SAFHS description conflicts with reference 27
MinorThe dataset description calls SAFHS a breast-cancer-related natural language processing dataset developed by Harvard Medical School. Its cited reference instead describes the San Antonio Family Heart Study and a genome-wide scan for microalbuminuria.
“The SAFHS [27] is a large-scale population-based natural language processing dataset developed by Harvard Medical School.”
“27. Arar N. Nath S. Thameem F. Bauer R. Voruganti S. Comuzzie A. Cole S. Blangero J. MacCluer J. Abboud H. Genome-wide scans for microalbuminuria in Mexican Americans: The San Antonio family heart study”
Reference 8 is ANDI, not ADNI
MinorThe paper cites reference 8 for the ADNI dataset, but reference 8 identifies a different database, ANDI.
“the ADNI dataset, cited in references [7,8]”
“8. de Vent N.R. Agelink van Rentergem J.A. Schmand B.A. Murre J.M. Consortium A. Huizenga H.M. Advanced Neuropsychological Diagnostics Infrastructure (ANDI): A normative database created from control datasets”
Logistic regression cites a random-forests reference
MinorThe logistic-regression description cites reference 45, whose listed work is titled "Random decision forests."
“Logistic regression (LR). LR [45] maps the results of linear regression to the range (0, 1) using a logistic function”
“45. Ho T.K. Random decision forests Proceedings of the 3rd International Conference on Document Analysis and Recognition”
An epilepsy EEG study is presented as evidence for Alzheimer's diagnosis
MinorThe Alzheimer's section cites Djemili et al. [90] as demonstrating EEG-based Alzheimer's diagnosis. The cited paper's title specifies classification of normal and epileptic EEG signals instead.
“Notable research by Djemili et al. [90], Pandya et al. [91], Kim et al. [92], along with studies cited as [93], have demonstrated the utility of EEG in capturing the neurophysiological hallmarks of Alzheimer’s disease”
“90. Djemili R. Bourouba H. Korba M.A. Application of empirical mode decomposition and artificial neural network for the classification of normal and epileptic EEG signals”
Correct and incorrect predictions are assigned to the wrong totals
Minor"True Classification" includes false positives, which the paper defines as incorrect predictions. "False Classification" includes true negatives, which it defines as correct predictions.
“True Classification is the sum of TP and FP, and False Classification is the sum of FN and TN.”
“TN represents the correctly identified negative instances, i.e., the negative class correctly classified as negative. FP represents the falsely identified positive instances”
Cites
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · see citing works
Jakubův, J., Chvalovský, K., Goertzel, Z., et al. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.itp.2023.19 · Cited by 76,648 · see citing works
Hallmarks of Cancer: The Next Generation
Hanahan, D., & Weinberg, R. A. (2011) · Cell · doi:10.1016/j.cell.2011.02.013 · Cited by 66,590 · see citing works
You Only Look Once: Unified, Real-Time Object Detection
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.91 · Cited by 53,803 · see citing works
Goldberger, A. L., Amaral, L. A. N., Glass, L., Hausdorff, J. M., Ivanov, P. C., Mark, R. G., Mietus, J. E., Moody, G. B., Peng, C.-K., & Stanley, H. E. (2000) · Circulation · doi:10.1161/01.cir.101.23.e215 · Cited by 14,809 · see citing works
Kaugeranna, R., Kaugeranna, E., & Aion, (. S. 4. 6. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.giscience.2023.43 · Cited by 13,695 · see citing works
Gene Expression Omnibus: NCBI gene expression and hybridization array data repository
Edgar, R. (2002) · Nucleic Acids Research · doi:10.1093/nar/30.1.207 · Cited by 13,288 · see citing works
Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data
Lafferty, J., McCallum, A., & Pereira, F. C. N. (2001) · ScholarlyCommons (University of Pennsylvania) · Cited by 12,953 · see citing works
Sudlow, C., Gallacher, J., Allen, N., Beral, V., Burton, P., Danesh, J., Downey, P., Elliott, P., Green, J., Landray, M., et al. (2015) · PLOS Medicine · doi:10.1371/journal.pmed.1001779 · Cited by 12,728 · see citing works
Advances and Open Problems in Federated Learning
Kairouz, P., & McMahan, H. B. (2021) · Foundations and Trends in Machine Learning · doi:10.1561/2200000083 · Cited by 7,772 · see citing works
An overview of statistical learning theory
Vapnik, V. N. (1999) · IEEE Transactions on Neural Networks · doi:10.1109/72.788640 · Cited by 6,305 · see citing works
The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups
METABRIC Group, Curtis, C., Shah, S. P., Chin, S.-F., Turashvili, G., Rueda, O. M., Dunning, M. J., Speed, D., Lynch, A. G., Samarajiwa, S., et al. (2012) · Nature · doi:10.1038/nature10983 · Cited by 6,097 · see citing works
Training language models to follow instructions with human feedback
Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C. L., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al. (2022) · arXiv · doi:10.48550/arxiv.2203.02155 · Cited by 4,208 · see citing works
Large language models encode clinical knowledge
Singhal, K., Azizi, S., Tu, T., Mahdavi, S. S., Wei, J., Chung, H. W., Scales, N., Tanwani, A., Cole-Lewis, H., Pfohl, S., et al. (2023) · Nature · doi:10.1038/s41586-023-06291-2 · Cited by 4,083 · see citing works
SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
Park, D. S., Chan, W., Zhang, Y., Chiu, C.-C., Zoph, B., Cubuk, E. D., & Le, Q. V. (2019) · Interspeech 2019 · doi:10.21437/interspeech.2019-2680 · Cited by 3,558 · see citing works
Review The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge
Tomczak, K. J., Czerwińska, P., & Wiznerowicz, M. (2015) · Współczesna Onkologia · doi:10.5114/wo.2014.47136 · Cited by 3,503 · see citing works
Andrzejak, R. G., Lehnertz, K., Mormann, F., Rieke, C., David, P., & Elger, C. E. (2001) · Physical Review E · doi:10.1103/physreve.64.061907 · Cited by 3,032 · see citing works
PaLM: Scaling Language Modeling with Pathways
Chowdhery, A., Narang, S., Devlin, J., Bosma, M., Mishra, G., Roberts, A., Barham, P., Chung, H. W., Sutton, C., Gehrmann, S., et al. (2022) · arXiv · doi:10.48550/arxiv.2204.02311 · Cited by 2,590 · see citing works
Bernard, O., Lalande, A., Zotti, C., Cervenansky, F., Yang, X., Heng, P.-A., Cetin, I., Lekadir, K., Camara, O., Gonzalez Ballester, M. A., et al. (2018) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2018.2837502 · Cited by 2,431 · see citing works
Classification assessment methods
Tharwat, A. (2021) · Applied Computing and Informatics · doi:10.1016/j.aci.2018.08.003 · Cited by 2,417 · see citing works
Dataset of breast ultrasound images
Al-Dhabyani, W., Gomaa, M., Khaled, H., & Fahmy, A. (2020) · Data in Brief · doi:10.1016/j.dib.2019.104863 · Cited by 2,405 · see citing works
Marcus, D. S., Wang, T. H., Parker, J., Csernansky, J. G., Morris, J. C., & Buckner, R. L. (2007) · Journal of Cognitive Neuroscience · doi:10.1162/jocn.2007.19.9.1498 · Cited by 2,163 · see citing works
Health disparities and health equity: Concepts and measurement
Braveman, P. (2006) · Annual Review of Public Health · doi:10.1146/annurev.publhealth.27.021405.102103 · Cited by 1,570 · see citing works
The Alzheimer's Disease Neuroimaging Initiative
Mueller, S. G., Weiner, M. W., Thal, L. J., Petersen, R. C., Jack, C., Jagust, W., Trojanowski, J. Q., Toga, A. W., & Beckett, L. (2005) · Neuroimaging Clinics of North America · doi:10.1016/j.nic.2005.09.008 · Cited by 1,525 · see citing works
Secure, privacy-preserving and federated machine learning in medical imaging
Kaissis, G. A., Makowski, M. R., Rückert, D., & Braren, R. F. (2020) · Nature Machine Intelligence · doi:10.1038/s42256-020-0186-1 · Cited by 1,501 · see citing works
Explainable AI: From black box to glass box
Rai, A. (2020) · Journal of the Academy of Marketing Science · doi:10.1007/s11747-019-00710-5 · Cited by 1,317 · see citing works
BioGPT: generative pre-trained transformer for biomedical text generation and mining
Luo, R., Sun, L., Xia, Y., Qin, T., Zhang, S., Poon, H., & Liu, T.-Y. (2022) · Briefings in Bioinformatics · doi:10.1093/bib/bbac409 · Cited by 1,150 · see citing works
The first step for neuroimaging data analysis: DICOM to NIfTI conversion
Li, X., Morgan, P. S., Ashburner, J., Smith, J., & Rorden, C. (2016) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2016.03.001 · Cited by 1,073 · see citing works
Zhang, J., Gajjala, S., Agrawal, P., Tison, G. H., Hallock, L. A., Beussink-Nelson, L., Lassen, M. H., Fan, E., Aras, M. A., Jordan, C., et al. (2018) · Circulation · doi:10.1161/circulationaha.118.034338 · Cited by 1,006 · see citing works
Machine learning in medicine: Addressing ethical challenges
Vayena, E., Blasimme, A., & Cohen, I. G. (2018) · PLOS Medicine · doi:10.1371/journal.pmed.1002689 · Cited by 912 · see citing works
Peng, Y., Yan, S., & Lu, Z. (2019) · Proceedings of the 18th BioNLP Workshop and Shared Task · doi:10.18653/v1/w19-5006 · Cited by 899 · see citing works
Swarm Learning for decentralized and confidential clinical machine learning
Warnat-Herresthal, S., Schultze, H., Shastry, K. L., Manamohan, S., Mukherjee, S., Garg, V., Sarveswara, R., Händler, K., Pickkers, P., Aziz, N. A., et al. (2021) · Nature · doi:10.1038/s41586-021-03583-3 · Cited by 884 · see citing works
Shen, R., Olshen, A. B., & Ladanyi, M. (2009) · Bioinformatics · doi:10.1093/bioinformatics/btp543 · Cited by 849 · see citing works
Artificial Intelligence and Human Trust in Healthcare: Focus on Clinicians
Asan, O., Bayrak, A. E., & Choudhury, A. (2020) · Journal of Medical Internet Research · doi:10.2196/15154 · Cited by 825 · see citing works
A framework for feature selection in clustering
Witten, D. M., & Tibshirani, R. (2010) · Journal of the American Statistical Association · doi:10.1198/jasa.2010.tm09415 · Cited by 696 · see citing works
Li, Y., Li, Z., Zhang, K., Dan, R., Jiang, S. B., & Zhang, Y. (2023) · Cureus · doi:10.7759/cureus.40895 · Cited by 616 · see citing works
The distress analysis interview corpus of human and computer interviews
Gratch, J., Artstein, R., Lucas, G., Stratou, G., Scherer, S., Nazarian, A., Wood, R., Boberg, J., DeVault, D., Marsella, S., et al. (2014) · Proceedings of the Language Resources and Evaluation Conference · doi:10.63317/3o7bccg9xequ · Cited by 599 · see citing works
Shen, R., Olshen, A. B., & Ladanyi, M. (2010) · Bioinformatics · doi:10.1093/bioinformatics/btp659 · Cited by 566 · see citing works
Lian, C., Liu, M., Zhang, J., & Shen, D. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2018.2889096 · Cited by 544 · see citing works
The temple university hospital EEG data corpus
Obeid, I., & Picone, J. (2016) · Frontiers in Neuroscience · doi:10.3389/fnins.2016.00196 · Cited by 528 · see citing works
UK Biobank: Current status and what it means for epidemiology
Allen, N., Sudlow, C., Downey, P., Peakman, T., Danesh, J., Elliott, P., Gallacher, J., Green, J., Matthews, P., Pell, J., et al. (2012) · Health Policy and Technology · doi:10.1016/j.hlpt.2012.07.003 · Cited by 521 · see citing works
Deep learning interpretation of echocardiograms
Ghorbani, A., Ouyang, D., Abid, A., He, B., Chen, J. H., Harrington, R. A., Liang, D. H., Ashley, E. A., & Zou, J. Y. (2020) · npj Digital Medicine · doi:10.1038/s41746-019-0216-8 · Cited by 518 · see citing works
Cardiovascular risk factors in an Iranian urban population: Tehran Lipid and Glucose Study (Phase 1)
Azizi, F., Rahmani, M., Emami, H., Mirmiran, P., Hajipour, R., Madjid, M., Ghanbili, J., Ghanbarian, A., Mehrabi, J., Saadat, N., et al. (2002) · Sozial- und Präventivmedizin · doi:10.1007/s000380200008 · Cited by 499 · see citing works
Biobank, O. B. O. D. P. U. A. U., Wilkinson, T., Schnier, C., Bush, K., Rannikmäe, K., Henshall, D. E., Lerpiniere, C., Allen, N. E., Flaig, R., Russ, T. C., et al. (2019) · European Journal of Epidemiology · doi:10.1007/s10654-019-00499-1 · Cited by 452 · see citing works
Shi, J., Zheng, X., Li, Y., Zhang, Q., & Ying, S. (2018) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2017.2655720 · Cited by 444 · see citing works
Ko, W.-Y., Siontis, K. C., Attia, Z. I., Carter, R. E., Kapa, S., Ommen, S. R., Demuth, S. J., Ackerman, M. J., Gersh, B. J., Arruda-Olson, A. M., et al. (2020) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2019.12.030 · Cited by 404 · see citing works
Evaluation Framework for Algorithms Segmenting Short Axis Cardiac MRI.
Radau, P. E., Lu, Y., Connelly, K. A., Paul, G. A., Dick, A. J., & Wright, G. A. (2009) · The MIDAS Journal · doi:10.54294/g80ruo · Cited by 373 · see citing works
Zeng, Q. T., Goryachev, S., Weiss, S., Sordo, M., Murphy, S. N., & Lazarus, R. (2006) · BMC Medical Informatics and Decision Making · doi:10.1186/1472-6947-6-30 · Cited by 360 · see citing works
Jabeen, K., Khan, M. A., Alhaisoni, M., Tariq, U., Zhang, Y.-D., Hamza, A., Mickus, A., & Damaševičius, R. (2022) · Sensors · doi:10.3390/s22030807 · Cited by 314 · see citing works
A multi-modal open dataset for mental-disorder analysis
Cai, H., Yuan, Z., Gao, Y., Sun, S., Li, N., Tian, F., Xiao, H., Li, J., Yang, Z., Li, X., et al. (2022) · Scientific Data · doi:10.1038/s41597-022-01211-x · Cited by 293 · see citing works
Self-citationBhattacharyya, A., Pachori, R. B., Upadhyay, A. B., & Acharya, U. R. (2017) · Applied Sciences · doi:10.3390/app7040385 · Cited by 285 · see citing works
Dual attention multi-instance deep learning for Alzheimer’s disease diagnosis with structural MRI
Zhu, W., Sun, L., Huang, J., Han, L., & Zhang, D. (2021) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2021.3077079 · Cited by 278 · see citing works
Mental-LLM: Leveraging Large Language Models for Mental Health Prediction via Online Text Data
Xu, X., Yao, B., Dong, Y., Gabriel, S., Yu, H., Hendler, J., Ghassemi, M., Dey, A. K., & Wang, D. (2024) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3643540 · Cited by 278 · see citing works
Self-citationUK Biobank: Opportunities for cardiovascular research
Littlejohns, T. J., Sudlow, C., Allen, N. E., & Collins, R. (2019) · European Heart Journal · doi:10.1093/eurheartj/ehx254 · Cited by 270 · see citing works
Al-antari, M. A., Han, S.-M., & Kim, T.-S. (2020) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2020.105584 · Cited by 264 · see citing works
Automated depression analysis using convolutional neural networks from speech
He, L., & Cui, C. (2018) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2018.05.007 · Cited by 250 · see citing works
Madani, A., Ong, J. R., Tibrewal, A., & Mofrad, M. R. K. (2018) · npj Digital Medicine · doi:10.1038/s41746-018-0065-x · Cited by 250 · see citing works
The influence of racial factors on psychiatric diagnosis: A review and suggestions for research
Neighbors, H. W., Jackson, J. S., Campbell, L., & Williams, D. (1989) · Community Mental Health Journal · doi:10.1007/bf00755677 · Cited by 244 · see citing works
Zhou, T., Liu, M., Thung, K., & Shen, D. G. (2019) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2019.2913158 · Cited by 226 · see citing works
Jan, A., Meng, H., Gaus, Y. F. B. A., & Zhang, F. (2018) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2017.2721552 · Cited by 219 · see citing works
Automated and interpretable patient ECG profiles for disease detection, tracking, and discovery
Tison, G. H., Zhang, J., Delling, F. N., & Deo, R. C. (2019) · Circulation: Cardiovascular Quality and Outcomes · doi:10.1161/circoutcomes.118.005289 · Cited by 208 · see citing works
Levine, D. M., Tuwani, R., Kompa, B., Varma, A., Finlayson, S. G., Mehrotra, A., & Beam, A. (2024) · The Lancet Digital Health · doi:10.1016/s2589-7500(24)00097-9 · Cited by 199 · see citing works
Park, C., Ha, J., & Park, S. (2020) · Expert Systems with Applications · doi:10.1016/j.eswa.2019.112873 · Cited by 184 · see citing works
Multimodal Measurement of Depression Using Deep Learning Models
Yang, L., Jiang, D., Xia, X., Pei, E., Oveneke, M. C., & Hichem Sahli, B. (2017) · doi:10.1145/3133944.3133948 · Cited by 175 · see citing works
Convolution Neural Network for Breast Cancer Detection and Classification Using Deep Learning
Abunasser, B., AL-Hiealy, M. R., Zaqout, I., & Abu-Naser, S. (2023) · Asian Pacific Journal of Cancer Prevention · doi:10.31557/apjcp.2023.24.2.531 · Cited by 174 · see citing works
Ragab, M., Albukhari, A., Alyami, J., & Mansour, R. F. (2022) · Biology · doi:10.3390/biology11030439 · Cited by 172 · see citing works
Multimodal Attention-Based Deep Learning for Alzheimer’s Disease Diagnosis
Golovanevsky, M., Eickhoff, C., & Singh, R. (2022) · Journal of the American Medical Informatics Association · doi:10.1093/jamia/ocac168 · Cited by 165 · see citing works
Mahato, S., & Paul, S. (2020) · Journal of Medical Systems · doi:10.1007/s10916-019-1486-z · Cited by 148 · see citing works
Yang, X., Liu, G., Feng, G., Bu, D., Wang, P., Jiang, J., Chen, S., Yang, Q., Zhang, Y., Man, Z., et al. (2023) · doi:10.1101/2023.09.26.559542 · Cited by 148 · see citing works
Self-citationYang, X., Liu, G., Feng, G., Bu, D., Wang, P., Jiang, J., Chen, S., Yang, Q., Miao, H., Zhang, Y., et al. (2024) · Cell Research · doi:10.1038/s41422-024-01034-y · Cited by 148 · see citing works
Self-citationČukić, M., Stokić, M., Simić, S., & Pokrajac, D. (2020) · Cognitive Neurodynamics · doi:10.1007/s11571-020-09581-x · Cited by 130 · see citing works
Serrano, B. G., Spiers, A. D. V., Ruotong, L., Gangadia, S., Toledano, M. B., & Simplicio, M. C. D. (2022) · European Child & Adolescent Psychiatry · doi:10.1007/s00787-022-02012-8 · Cited by 125 · see citing works
Zhou, S., Wang, N., Wang, L., Liu, H., & Zhang, R. (2022) · Journal of the American Medical Informatics Association · doi:10.1093/jamia/ocac040 · Cited by 119 · see citing works
Artificial intelligence-enabled decision support in nephrology
Loftus, T. J., Shickel, B., Ozrazgat‐Baslanti, T., Ren, Y., Glicksberg, B. S., Cao, J., Singh, K., Chan, L., Nadkarni, G. N., & Bihorac, A. (2022) · Nature Reviews Nephrology · doi:10.1038/s41581-022-00562-3 · Cited by 109 · see citing works
An Efficient Classifier for Alzheimer’s Disease Genes Identification
Xu, L., Liang, G., Liao, C., Chen, G.-D., & Chang, C.-C. (2018) · Molecules · doi:10.3390/molecules23123140 · Cited by 107 · see citing works
Detecting Depression Using an Ensemble Logistic Regression Model Based on Multiple Speech Features
Jiang, H., Hu, B., Liu, Z., Wang, G., Zhang, L., Li, X., & Kang, H. (2018) · Computational and Mathematical Methods in Medicine · doi:10.1155/2018/6508319 · Cited by 107 · see citing works
Fürbass, F., Kural, M. A., Gritsch, G., Hartmann, M., Kluge, T., & Beniczky, S. (2020) · Clinical Neurophysiology · doi:10.1016/j.clinph.2020.02.032 · Cited by 106 · see citing works
Relation-induced multi-modal shared representation learning for Alzheimer’s disease diagnosis
Ning, Z., Xiao, Q., Feng, Q., Chen, W., & Zhang, Y. (2021) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2021.3063150 · Cited by 106 · see citing works
Mao, K., Zhang, W., Wang, D. B., Li, A., Jiao, R., Zhu, Y., Wu, B., Zheng, T., Lei, Q., Lyu, W., et al. (2022) · IEEE Transactions on Affective Computing · doi:10.1109/taffc.2022.3154332 · Cited by 106 · see citing works
The Diagnostic and Triage Accuracy of the GPT-3 Artificial Intelligence Model
Levine, D. M., Tuwani, R., Kompa, B., Varma, A., Finlayson, S. G., Mehrotra, A., & Beam, A. (2023) · doi:10.1101/2023.01.30.23285067 · Cited by 106 · see citing works
A multitask framework to detect depression, sentiment and multi-label emotion from suicide notes
Ghosh, S., Ekbal, A., & Bhattacharyya, P. (2022) · Cognitive Computation · doi:10.1007/s12559-021-09828-7 · Cited by 102 · see citing works
A hierarchical depression detection model based on vocal and emotional cues
Dong, Y., & Yang, X. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.02.019 · Cited by 99 · see citing works
Deep Multi-Modal Network Based Automated Depression Severity Estimation
Uddin, M. A., Joolee, J. B., & Sohn, K.-A. (2022) · IEEE Transactions on Affective Computing · doi:10.1109/taffc.2022.3179478 · Cited by 99 · see citing works
Principles of Health Interoperability
Benson, T., & Grieve, G. (2020) · Health information technology standards · doi:10.1007/978-3-030-56883-2 · Cited by 96 · see citing works
Epileptic seizure detection and prediction using stacked bidirectional long short term memory
D.K., T., B.G., P., & Xiong, F. (2019) · Pattern Recognition Letters · doi:10.1016/j.patrec.2019.10.034 · Cited by 93 · see citing works
Djemili, R., Bourouba, H., & Korba, M. C. A. (2015) · Journal of Applied Biomedicine · doi:10.1016/j.bbe.2015.10.006 · Cited by 93 · see citing works
An integrative deep learning framework for classifying molecular subtypes of breast cancer
Mohaiminul Islam, M., Huang, S., Ajwad, R., Chi, C., Wang, Y., & Hu, P. (2020) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2020.08.005 · Cited by 90 · see citing works
Huang, Y., Wei, L., Hu, Y., Shao, N., Lin, Y., He, S., Shi, H., Zhang, X., & Lin, Y. (2021) · Frontiers in Oncology · doi:10.3389/fonc.2021.706733 · Cited by 82 · see citing works
Dai, H., Hwang, H.-G., & Tseng, V. S. (2021) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2021.106035 · Cited by 81 · see citing works
Jo, T., Nho, K., Bice, P., Saykin, A. J., & Initiative, F. T. A. D. N. (2022) · Briefings in Bioinformatics · doi:10.1093/bib/bbac022 · Cited by 80 · see citing works
Detection of Early Stage Alzheimer’s Disease using EEG Relative Power with Deep Neural Network
Kim, D., & Kim, K. (2018) · 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2018.8512231 · Cited by 76 · see citing works
Depression Detection in Speech Using Transformer and Parallel Convolutional Neural Networks
Yin, F., Du, J. K., Xu, X., & Li, Z. (2023) · Electronics · doi:10.3390/electronics12020328 · Cited by 75 · see citing works
Self-citationAl-Tam, R. M., Al-Hejri, A. M., Narangale, S. M., Samee, N. A., Mahmoud, N. F., Al-masni, M. A., & Al-antari, M. A. (2022) · Biomedicines · doi:10.3390/biomedicines10112971 · Cited by 74 · see citing works
Benchmarking machine learning models for late-onset alzheimer’s disease prediction from genomic data
De Velasco Oriol, J., Vallejo, E. E., Estrada, K., Taméz Peña, J. G., & Disease Neuroimaging Initiative, T. A. (2019) · BMC Bioinformatics · doi:10.1186/s12859-019-3158-x · Cited by 71 · see citing works
Deep-Asymmetry: Asymmetry Matrix Image for Deep Learning Method in Pre-Screening Depression
Kang, M., Kwon, H., Park, J.-H., Kang, S., & Lee, Y. (2020) · Sensors · doi:10.3390/s20226526 · Cited by 71 · see citing works
de Vent, N. R., van Rentergem, J. A. A., Schmand, B. A., Murre, J. M. J., & Huizenga, H. M. (2016) · Frontiers in Psychology · doi:10.3389/fpsyg.2016.01601 · Cited by 69 · see citing works
Study on feature selection methods for depression detection using three-electrode EEG data
Cai, H., Chen, Y., Han, J., Zhang, X., & Hu, B. (2018) · Interdisciplinary Sciences: Computational Life Sciences · doi:10.1007/s12539-018-0292-5 · Cited by 68 · see citing works
Self-citationAn automatic Alzheimer’s disease classifier based on spontaneous spoken English
Bertini, F., Allevi, D., Lutero, G., Calzà, L., & Montesi, D. (2022) · Computer Speech & Language · doi:10.1016/j.csl.2021.101298 · Cited by 66 · see citing works
Lopez-de-Ipina, K., Martinez-de-Lizarduy, U., Calvo, P. M., Mekyska, J., Beitia, B., Barroso, N., Estanga, A., Tainta, M., & Ecay-Torres, M. (2018) · Current Alzheimer Research · doi:10.2174/1567205014666171120143800 · Cited by 65 · see citing works
Multimodal mild depression recognition based on EEG-EM synchronization acquisition network
Zhu, J., Wang, Y., La, R., Zhan, J., Niu, J., Zeng, S., & Hu, X. (2019) · IEEE Access · doi:10.1109/access.2019.2901950 · Cited by 56 · see citing works
Computer-assisted EEG diagnostic review for idiopathic generalized epilepsy
Clarke, S., Karoly, P. J., Nurse, E., Seneviratne, U., Taylor, J., Knight-Sadler, R., Kerr, R., Moore, B., Hennessy, P., Mendis, D., et al. (2021) · Epilepsy & Behavior · doi:10.1016/j.yebeh.2019.106556 · Cited by 55 · see citing works
Martinc, M., Haider, F., Pollak, S., & Luz, S. (2021) · Frontiers in Aging Neuroscience · doi:10.3389/fnagi.2021.642647 · Cited by 55 · see citing works
Sherafatian, M. (2018) · Gene · doi:10.1016/j.gene.2018.07.057 · Cited by 51 · see citing works
Multi-Head Attention-Based Long Short-Term Memory for Depression Detection From Speech
Zhao, Y. D., Liang, Z., Du, J. K., Zhang, L., Liu, C., & Li, Z. (2021) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2021.684037 · Cited by 50 · see citing works
Cheng, D., & Liu, M. (2017) · 2017 IEEE International Conference on Imaging Systems and Techniques (IST) · doi:10.1109/ist.2017.8261461 · Cited by 49 · see citing works
Pandya, R., Nadiadwala, S., Shah, R., & Shah, M. (2019) · Augmented Human Research · doi:10.1007/s41133-019-0021-6 · Cited by 49 · see citing works
Agbavor, F., & Liang, H. (2022) · Brain Sciences · doi:10.3390/brainsci13010028 · Cited by 47 · see citing works
Sharma, R., Goel, T., Tanveer, M., Suganthan, P. N., Razzak, I., & Murugan, R. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2022.3215533 · Cited by 44 · see citing works
Hamid, H., Fodeh, S. J., Lizama, A. G., Czlapinski, R., Pugh, M. J., LaFrance, W. C., & Brandt, C. A. (2013) · Epilepsy & Behavior · doi:10.1016/j.yebeh.2013.09.025 · Cited by 42 · see citing works
A hierarchical multimodal system for motion analysis in patients with epilepsy
Ahmedt‐Aristizabal, D., Fookes, C., Denman, S., et al. (2018) · Epilepsy & Behavior · doi:10.1016/j.yebeh.2018.07.028 · Cited by 42 · see citing works
Tison, G. H., Siontis, K. C., Abreau, S., Attia, Z., Agarwal, P., Balasubramanyam, A., Li, Y., Sehnert, A. J., Edelberg, J. M., Friedman, P. A., et al. (2022) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2022.01.005 · Cited by 42 · see citing works
Pominova, M., Artemov, A., Sharaev, M., Kondrateva, E., Bernstein, A., & Burnaev, E. V. (2018) · IEEE ... International Conference on Data Mining workshops · doi:10.1109/icdmw.2018.00050 · Cited by 41 · see citing works
Learning Voice Source Related Information for Depression Detection
Dubagunta, S. P., Vlasenko, B., & Magimai.-Doss, M. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683498 · Cited by 41 · see citing works
Fairness in Group Recommendations in the Health Domain
Stratigi, M., Kondylakis, H., & Stefanidis, K. (2017) · Proceedings - International Conference on Data Engineering · doi:10.1109/icde.2017.217 · Cited by 39 · see citing works
Detecting depression in speech: Comparison and combination between different speech types
Long, H., Guo, Z., Wu, X., Hu, B. X., Liu, Z., & Cai, H. (2017) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm.2017.8217802 · Cited by 38 · see citing works
Contrastive Learning for Prediction of Alzheimer’s Disease Using Brain 18F-FDG PET
Chen, Y., Wang, H., Zhang, G., Liu, X., Huang, W., Han, X., Li, X., Martin, M., & Tao, L. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2022.3231905 · Cited by 36 · see citing works
Self-citationLiu, Z., Guo, Z., Ling, Z., & Li, Y. (2021) · ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp39728.2021.9413566 · Cited by 34 · see citing works
Baydargil, H. B., Park, J.-S., & Kang, D.-Y. (2021) · Applied Sciences · doi:10.3390/app11052187 · Cited by 30 · see citing works
An unsupervised domain adaptation brain CT segmentation method across image modalities and diseases
Dong, D., Fu, G., Li, J., Pei, Y., & Chen, Y. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.118016 · Cited by 30 · see citing works
Self-citationFairGRecs: Fair Group Recommendations by Exploiting Personal Health Information
Stratigi, M., Kondylakis, H., & Stefanidis, K. (2018) · Lecture notes in computer science · doi:10.1007/978-3-319-98812-2_11 · Cited by 30 · see citing works
Cognitive impairment prediction in the elderly based on vocal biomarkers
Yu, B., Quatieri, T. F., Williamson, J. R., & Mundt, J. C. (2015) · Interspeech 2015 · doi:10.21437/interspeech.2015-741 · Cited by 28 · see citing works
Improving Breast Cancer Detection Using Symmetry Information with Deep Learning
Brhane Hagos, Y., Gubern Mérida, A., & Teuwen, J. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-00946-5_10 · Cited by 27 · see citing works
Connolly, B., Matykiewicz, P., Bretonnel Cohen, K., Standridge, S. M., Glauser, T. A., Dlugos, D. J., Koh, S., Tham, E., & Pestian, J. (2014) · Journal of the American Medical Informatics Association · doi:10.1136/amiajnl-2013-002601 · Cited by 26 · see citing works
Samadishadlou, M., Rahbarghazi, R., Piryaei, Z., Esmaeili, M., Avcı, Ç. B., Bani, F., & Kavousi, K. (2023) · Cardiovascular Diabetology · doi:10.1186/s12933-023-01957-7 · Cited by 26 · see citing works
Single-channel EEG-based machine learning method for prescreening major depressive disorder
Wan, Z., Zhang, H., Huang, J., Zhou, H., Yang, J., & Zhong, N. (2019) · International Journal of Information Technology & Decision Making · doi:10.1142/s0219622019500342 · Cited by 24 · see citing works
Self-citationZhu, Z., Li, J., Zhao, Q., & Akhtar, F. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.120709 · Cited by 24 · see citing works
Self-citationSi, X., Zhang, X., Zhou, Y., et al. (2020) · Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society · doi:10.1109/embc44109.2020.9175467 · Cited by 23 · see citing works
Pevy, N., Christensen, H., Walker, T., & Reuber, M. (2021) · Seizure · doi:10.1016/j.seizure.2021.06.009 · Cited by 23 · see citing works
A Machine Learning Model to Predict Seizure Susceptibility from Resting-State fMRI Connectivity
Garner, R., Rocca, M. L., Barisano, G., Toga, A. W., Duncan, D., & Vespa, P. (2019) · 2019 Spring Simulation Conference (SpringSim) · doi:10.23919/springsim.2019.8732859 · Cited by 22 · see citing works
Guan, Y., Wen, P., Li, J., Zhang, J., & Xie, X. (2024) · Tsinghua Science and Technology · doi:10.26599/tst.2022.9010016 · Cited by 22 · see citing works
Self-citationMultimodal Approaches for Alzheimer’s Detection Using Patients’ Speech and Transcript
Cai, H., Huang, X., Liu, Z., Liao, W., Dai, H., Wu, Z., Zhu, D., Ren, H., Li, Q., Liu, T., et al. (2023) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-43075-6_34 · Cited by 22 · see citing works
The future of brain–machine interfaces is optical
Ersaro, N. T., Yalcin, C., & Muller, R. (2023) · Nature Electronics · doi:10.1038/s41928-023-00926-y · Cited by 22 · see citing works
A computer-aided system for ocular myasthenia gravis diagnosis
Liu, G., Wei, Y., Xie, Y., Li, J., Qiao, L., & Yang, J.-J. (2021) · Tsinghua Science and Technology · doi:10.26599/tst.2021.9010025 · Cited by 21 · see citing works
Self-citationPeng, W., Sun, Y., & Zhang, L. (2022) · BMC Cardiovascular Disorders · doi:10.1186/s12872-022-02481-4 · Cited by 21 · see citing works
Hybrid optimization enabled deep learning-based ensemble classification for heart disease detection
Jayasudha, R., Suragali, C., Thirukrishna, J. T., & Santhosh Kumar, B. (2023) · Signal, Image and Video Processing · doi:10.1007/s11760-023-02656-2 · Cited by 21 · see citing works
Computational approach for deriving cancer progression roadmaps from static sample data
Sun, Y., Yao, J. B., Yang, L., Chen, R., Nowak, N. J., & Steve Goodison, B. (2017) · Nucleic Acids Research · doi:10.1093/nar/gkx003 · Cited by 20 · see citing works
Shah, Z., Qi, S.-A., Wang, F., Farrokh, M., Tasnim, M., Stroulia, E., Greiner, R., Plitsis, M., & Katsamanis, A. (2023) · ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp49357.2023.10095593 · Cited by 19 · see citing works
The Ustc System for Adress-m Challenge
Mei, K., Ding, X., Liu, Y., Guo, Z., Xu, F., Li, X., Naren, T., Yuan, J., & Ling, Z. (2023) · ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp49357.2023.10094714 · Cited by 19 · see citing works
Zhang, L., Liu, Y., Wang, K., Ou, X., Zhou, J., Zhang, H., Huang, M., Du, Z., & Qiang, S. (2023) · Journal of Translational Medicine · doi:10.1186/s12967-023-04573-x · Cited by 18 · see citing works
Vision-Based Mouth Motion Analysis in Epilepsy: A 3D Perspective
Ahmedt-Aristizabal, D., Nguyen, K., Denman, S., Sarfraz, M. S., Sridharan, S., Dionisio, S., & Fookes, C. (2019) · 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2019.8857656 · Cited by 17 · see citing works
Nasimova, N., Muminov, B., Nasimov, R., Abdurashidova, K., & Abdullaev, M. (2021) · 2021 International Conference on Information Science and Communications Technologies (ICISCT) · doi:10.1109/icisct52966.2021.9670134 · Cited by 17 · see citing works
Supervised Machine Learning Approaches for Medical Data Set Classification - A Review
Anto, S., & Krishna, S. H. (2011) · Cited by 17 · see citing works
KHyperLogLog: Estimating Reidentifiability and Joinability of Large Data at Scale
Chia, P. H., Desfontaines, D., Perera, I. M., Simmons-Marengo, D., Li, C., Day, W.-Y., Wang, Q., & Guevara, M. (2019) · 2019 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2019.00046 · Cited by 15 · see citing works
Zhang, J., Liu, B., Wu, J., Wang, Z., & Li, J. (2023) · BMC Bioinformatics · doi:10.1186/s12859-023-05469-9 · Cited by 15 · see citing works
Self-citationEpileptic seizure classification using the NeuroMov database
Maia, P., Hartl, E., Vollmar, C., Noachtar, S., & Silva Cunha, J. P. (2019) · 2019 IEEE 6th Portuguese Meeting on Bioengineering (ENBENG) · doi:10.1109/enbeng.2019.8692465 · Cited by 14 · see citing works
(2020) · SSRN Electronic Journal · Cited by 14 · see citing works
Karacsony, T., Loesch-Biffar, A. M., Vollmar, C., Noachtar, S., & Cunha, J. P. S. (2021) · 2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI) · doi:10.1109/bhi50953.2021.9508555 · Cited by 14 · see citing works
Genome-wide scans for microalbuminuria in Mexican Americans: The San Antonio Family Heart Study
Arar, N. H., Nath, S. D., Thameem, F., et al. (2007) · Genetics in Medicine · doi:10.1097/gim.0b013e31803068ec · Cited by 14 · see citing works
Ensemble learning with speaker embeddings in multiple speech task stimuli for depression detection
Liu, Z., Yu, H., Li, G., Chen, Q., Ding, Z., Feng, L., Yao, Z., & Hu, B. X. (2023) · Frontiers in Neuroscience · doi:10.3389/fnins.2023.1141621 · Cited by 14 · see citing works
Torres-Velazquez, M., Hwang, G., Cook, C. J., Hermann, B., Prabhakaran, V., Meyerand, M. E., & McMillan, A. B. (2020) · 2020 IEEE 17th International Symposium on Biomedical Imaging Workshops (ISBI Workshops) · doi:10.1109/isbiworkshops50223.2020.9153359 · Cited by 13 · see citing works
Qi, H., Zhao, Q., Li, J., Song, C., Zhai, W., Dan, L., Liu, S., Yu, Y. J., Wang, F., Zou, H., et al. (2023) · Research Square · doi:10.21203/rs.3.rs-3523508/v1 · Cited by 13 · see citing works
Self-citationSantoso, I. B., Adrianto, Y., Sensusiati, A. D., Wulandari, D. P., & Purnama, I. K. E. (2022) · IEEE Access · doi:10.1109/access.2022.3159923 · Cited by 13 · see citing works
Deep Learning with Multi Modal Ensemble Fusion for Epilepsy Diagnosis
Kunekar, P. R., Gupta, M. K., & Agarwal, B. (2020) · doi:10.1109/icetce48199.2020.9091742 · Cited by 11 · see citing works
Depression Detection Among Social Media Users Using Machine Learning
Verma, P., Sharma, K., & Walia, G. S. (2021) · Advances in Intelligent Systems and Computing · doi:10.1007/978-981-15-5113-0_72 · Cited by 10 · see citing works
Sundaram, D. S. B., Arunachalam, S. P., Damani, D. N., Farahani, N. Z., Enayati, M., Pasupathy, K. S., & Arruda-Olson, A. M. (2021) · Design of Medical Devices Conference (DMD) · doi:10.1115/dmd2021-1076 · Cited by 9 · see citing works
Cost-effective Models for Detecting Depression from Speech
Tasnim, M., & Novikova, J. (2022) · International Conference on Machine Learning and Applications · doi:10.1109/icmla55696.2022.00259 · Cited by 9 · see citing works
Achilles, F., Belagiannis, V., Tombari, F., Loesch, A. M., Cunha, J. P. S., Navab, N., & Noachtar, S. (2015) · Journal of the Neurological Sciences · doi:10.1016/j.jns.2015.09.065 · Cited by 8 · see citing works
Machine Learning Based Analysis of Structural MRI for Epilepsy Diagnosis
Sahebzamani, G., Saffar, M., & Soltanian-Zadeh, H. (2019) · 2019 4th International Conference on Pattern Recognition and Image Analysis (IPRIA) · doi:10.1109/pria.2019.8785985 · Cited by 8 · see citing works
Liu, J., Wang, X., Lin, J., Li, S., Deng, G., & Wei, J. (2021) · International Journal of General Medicine · doi:10.2147/ijgm.s329005 · Cited by 8 · see citing works
Comparison of Different Classification Methods for Breast Cancer Subtypes Prediction
Xu, J., Wu, P., Chen, Y., & Zhang, L. (2018) · 2018 International Conference on Security, Pattern Analysis, and Cybernetics (SPAC) · doi:10.1109/spac46244.2018.8965553 · Cited by 5 · see citing works
Self-citationSpeech evaluation of patients with alzheimer’s disease using an automatic interviewer
Alonso Hernández, J. B., Barragán Pulido, M. L., Gil Bordón, J. M., Ferrer Ballester, M. Á., & Travieso González, C. M. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2021.116386 · Cited by 5 · see citing works
Role of serum cytokines in the prediction of heart failure in patients with coronary artery disease
Hou, Q., Sun, Z., Zhao, L., Liu, Y., Zhang, J., Huang, J., Luo, Y., Xiao, Y., Hu, Z., & Shen, A. (2023) · ESC Heart Failure · doi:10.1002/ehf2.14491 · Cited by 5 · see citing works
Self-citationHCM-Dynamic-Echo: A Framework for Detecting Hypertrophic Cardiomyopathy (HCM) in Echocardiograms
Almadani, A., Agu, E., Sarwar, A., Ahluwalia, M., & Kpodonu, J. (2023) · 2023 IEEE International Conference on Digital Health (ICDH) · doi:10.1109/icdh60066.2023.00039 · Cited by 5 · see citing works
GPT-4 in breast cancer combat: A dazzling leap forward or merely a whim?
Deng, L., Zhang, Y., Luo, S., & Xu, J. (2023) · International Journal of Surgery · doi:10.1097/js9.0000000000000668 · Cited by 5 · see citing works
Mishra, J., Tiwari, M., Singh, S. T., & Goswami, S. (2022) · 2021 4th International Conference on Recent Trends in Computer Science and Technology (ICRTCST) · doi:10.1109/icrtcst54752.2022.9782009 · Cited by 4 · see citing works
The importance of SNOMED CT concept specificity in healthcare analytics
Roberts, L., Lanes, S., Peatman, O., et al. (2023) · Health Information Management Journal · doi:10.1177/18333583221144662 · Cited by 4 · see citing works
BDCN: Semantic Embedding Self-explanatory Breast Diagnostic Capsules Network
Chen, D., Zhong, K., & He, J. (2021) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-84186-7_28 · Cited by 3 · see citing works
Yoo, H. J., & Shin, S. (2023) · CIN Computers Informatics Nursing · doi:10.1097/cin.0000000000001055 · Cited by 1 · see citing works
Flagged first, then by citations
Research works printed but not linked: 26
- O’Shea, K., and Nash, R. (2015). An introduction to convolutional neural networks. arXiv.
- Vaswani, 2017, Attention is all you need, Adv. Neural Inf. Process. Syst., 30, 1
- Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., and Anadkat, S. (2023). Gpt-4 technical report. arXiv.
- Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.A., Lacroix, T., Rozière, B., Goyal, N., Hambro, E., and Azhar, F. (2023). Llama: Open and efficient foundation language models. arXiv.
- Touvron, H., Martin, L., Stone, K., Albert, P., Almahairi, A., Babaei, Y., Bashlykov, N., Batra, S., Bhargava, P., and Bhosale, S. (2023). Llama 2: Open foundation and fine-tuned chat models. arXiv.
- Singhal, K., Tu, T., Gottweis, J., Sayres, R., Wulczyn, E., Hou, L., Clark, K., Pfohl, S., Cole-Lewis, H., and Neal, D. (2023). Towards expert-level medical question answering with large language models. arXiv.
- Wang, H., Liu, C., Xi, N., Qiang, Z., Zhao, S., Qin, B., and Liu, T. (2023). Huatuo: Tuning llama model with chinese medical knowledge. arXiv.
- Yunxiang, L., Zihan, L., Kai, Z., Ruilong, D., and You, Z. (2023). Chatdoctor: A medical chat model fine-tuned on llama model using medical domain knowledge. arXiv.
- Xiong, H., Wang, S., Zhu, Y., Zhao, Z., Liu, Y., Wang, Q., and Shen, D. (2023). Doctorglm: Fine-tuning your chinese doctor is not a herculean task. arXiv.
- Chen, Y., Wang, Z., Xing, X., Xu, Z., Fang, K., Wang, J., Li, S., Wu, J., Liu, Q., and Xu, X. (2023). BianQue: Balancing the Questioning and Suggestion Ability of Health LLMs with Multi-turn Health Conversations Polished by ChatGPT. arXiv.
- Ho, T.K. (1995, January 14–16). Random decision forests. Proceedings of the 3rd International Conference on Document Analysis and Recognition, Montreal, QC, Canada.
- Anbarasi, 2010, Enhanced prediction of heart disease with feature subset selection using genetic algorithm, Int. J. Eng. Sci. Technol., 2, 5370
- Li, H., Habes, M., and Fan, Y. (2017). Deep ordinal ranking for multi-category diagnosis of Alzheimer’s disease using hippocampal MRI data. arXiv.
- Bommasani, R., Hudson, D.A., Adeli, E., Altman, R., Arora, S., von Arx, S., Bernstein, M.S., Bohg, J., Bosselut, A., and Brunskill, E. (2021). On the opportunities and risks of foundation models. arXiv.
- Zhou, C., Li, Q., Li, C., Yu, J., Liu, Y., Wang, G., Zhang, K., Ji, C., Yan, Q., and He, L. (2023). A comprehensive survey on pretrained foundation models: A history from bert to chatgpt. arXiv.
- Brown, 2020, Language models are few-shot learners, Adv. Neural Inf. Process. Syst., 33, 1877
- Chung, H.W., Hou, L., Longpre, S., Zoph, B., Tay, Y., Fedus, W., Li, Y., Wang, X., Dehghani, M., and Brahma, S. (2022). Scaling instruction-finetuned language models. arXiv.
- Zhang, C., Liu, L., Cui, Y., Huang, G., Lin, W., Yang, Y., and Hu, Y. (2023). A Comprehensive Survey on Segment Anything Model for Vision and Beyond. arXiv.
- Garza, A., and Mergenthaler-Canseco, M. (2023). TimeGPT-1. arXiv.
- Thawkar, O., Shaker, A., Mullappilly, S.S., Cholakkal, H., Anwer, R.M., Khan, S., Laaksonen, J., and Khan, F.S. (2023). Xraygpt: Chest radiographs summarization using medical vision-language models. arXiv.
- Freitag, 2018, audeep: Unsupervised learning of representations from audio with deep recurrent neural networks, J. Mach. Learn. Res., 18, 1
- Kumar, 2021, Phoenix@ SMM4H Task-8: Adversities Make Ordinary Models Do Extraordinary Things, NAACL-HLT, 2, 112
- Liu, Z., Wang, D., Zhang, L., and Hu, B. (2020). A novel decision tree for depression recognition in speech. arXiv.
- Ehghaghi, M., Rudzicz, F., and Novikova, J. (2022). Data-driven Approach to Differentiating between Depression and Dementia from Noisy Speech and Language Data. arXiv.
- Diep, B., Stanojevic, M., and Novikova, J. (2022). Multi-modal deep learning system for depression and anxiety detection. arXiv.
- Yao, X., Cheng, Q., and Zhang, G.Q. (2019). Automated classification of seizures against nonseizures: A deep learning approach. arXiv.
Cited by
Sorted by citations
Abugabah, A., Shukla, P. K., Shukla, P. K., & Pandey, A. (2025) · BioData Mining · doi:10.1186/s13040-025-00487-0 · Cited by 11 · see citing works
Anbukkarasi, S., Hemalatha, S., Balakrishnan, A., Varadhaganapathy, S., & Easwaramoorthy, S. V. (2026) · Scientific reports · doi:10.1038/s41598-026-47146-w · Cited by 1 · see citing works
“Extensive reviews have highlighted that the multimodal synergy, which covers the text, imaging, labs, and ontology-driven knowledge, is progressively becoming the core of the future medical AI systems 40 .”
Nadeem, A., Khan, M. Z., Sirshar, M., & Thirunavukkarasu, U. (2026) · Diagnostics · doi:10.3390/diagnostics16152377
- Authors
- Xu, X., Li, J., Zhu, Z., Zhao, L., Wang, H., Song, C., Chen, Y., Zhao, Q., Yang, J., & Pei, Y.
- Affiliations
- Beijing University of Technology (CN)
- Tsinghua National Laboratory for Information Science and Technology, Tsinghua University, Beijing 100084, China
- University of Aizu (JP)
- Corresponding author
- Zhao, Qing
- ORCID
- Venue
- Bioengineering
- Publisher
- MDPI AG
- Published
- 2024-02-25
- Volume
- 11
- Issue
- 3
- Pages
- 219
- ISSN
- 2306-5354 (ISSN-L)
- PMC
- PMC10967767
- OpenAlex
- W4392168796
- Semantic Scholar
- adc4bc346850a8047869666cc2112934f3f89d82
- Cited by
- 242
- References
- 168
- Citations by year
- 2026: 96 · 2025: 129 · 2024: 17
- FWCI
- 5.46
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Bioengineering (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Funders
- National Natural Science Foundation of China (62266041)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Keywords
- multimodal fusion, diagnosis, artificial intelligence in healthcare, Alzheimer's disease, breast cancer, depression, heart disease, epilepsy, multimodal medical data, clinical decision support, feature engineering, classification models, medical datasets
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.
Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-06.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.
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
