1
|
Zanfardino M, Punzo B, Maffei E, Saba L, Bossone E, Nistri S, La Grutta L, Franzese M, Cavaliere C, Cademartiri F. Unsupervised machine learning for risk stratification and identification of relevant subgroups of ascending aorta dimensions using cardiac CT and clinical data. Comput Struct Biotechnol J 2024; 23:287-294. [PMID: 38173875 PMCID: PMC10762320 DOI: 10.1016/j.csbj.2023.11.021] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/14/2023] [Revised: 11/10/2023] [Accepted: 11/10/2023] [Indexed: 01/05/2024] Open
Abstract
The potential of precision population health lies in its capacity to utilize robust patient data for customized prevention and care targeted at specific groups. Machine learning has the potential to automatically identify clinically relevant subgroups of individuals, considering heterogeneous data sources. This study aimed to assess whether unsupervised machine learning (UML) techniques could interpret different clinical data to uncover clinically significant subgroups of patients suspected of coronary artery disease and identify different ranges of aorta dimensions in the different identified subgroups. We employed a random forest-based cluster analysis, utilizing 14 variables from 1170 (717 men/453 women) participants. The unsupervised clustering approach successfully identified four distinct subgroups of individuals with specific clinical characteristics, and this allows us to interpret and assess different ranges of aorta dimensions for each cluster. By employing flexible UML algorithms, we can effectively process heterogeneous patient data and gain deeper insights into clinical interpretation and risk assessment.
Collapse
Affiliation(s)
| | | | - Erica Maffei
- Department of Imaging, Fondazione Monasterio/CNR, Pisa, 56124, Italy
| | - Luca Saba
- Department of Radiology, University Hospital of Cagliari, Cagliari, 09042, Italy
| | - Eduardo Bossone
- Department of Public Health, University of Naples Federico II, Naples, 80131, Italy
| | | | | | | | | | | |
Collapse
|
2
|
Linmans J, Raya G, van der Laak J, Litjens G. Diffusion models for out-of-distribution detection in digital pathology. Med Image Anal 2024; 93:103088. [PMID: 38228075 DOI: 10.1016/j.media.2024.103088] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/04/2023] [Revised: 12/29/2023] [Accepted: 01/10/2024] [Indexed: 01/18/2024]
Abstract
The ability to detect anomalies, i.e. anything not seen during training or out-of-distribution (OOD), in medical imaging applications is essential for successfully deploying machine learning systems. Filtering out OOD data using unsupervised learning is especially promising because it does not require costly annotations. A new class of models called AnoDDPMs, based on denoising diffusion probabilistic models (DDPMs), has recently achieved significant progress in unsupervised OOD detection. This work provides a benchmark for unsupervised OOD detection methods in digital pathology. By leveraging fast sampling techniques, we apply AnoDDPM on a large enough scale for whole-slide image analysis on the complete test set of the Camelyon16 challenge. Based on ROC analysis, we show that AnoDDPMs can detect OOD data with an AUC of up to 94.13 and 86.93 on two patch-level OOD detection tasks, outperforming the other unsupervised methods. We observe that AnoDDPMs alter the semantic properties of inputs, replacing anomalous data with more benign-looking tissue. Furthermore, we highlight the flexibility of AnoDDPM towards different information bottlenecks by evaluating reconstruction errors for inputs with different signal-to-noise ratios. While there is still a significant performance gap with fully supervised learning, AnoDDPMs show considerable promise in the field of OOD detection in digital pathology.
Collapse
Affiliation(s)
- Jasper Linmans
- Department of Pathology, RadboudUMC Graduate School, Radboud University Medical Center, Nijmegen, The Netherlands.
| | - Gabriel Raya
- Jheronimus Academy of Data Science, 's-Hertogenbosch, The Netherlands
| | - Jeroen van der Laak
- Department of Pathology, RadboudUMC Graduate School, Radboud University Medical Center, Nijmegen, The Netherlands; Center for Medical Image Science and Visualization, Linköping University, Linköping, Sweden
| | - Geert Litjens
- Department of Pathology, RadboudUMC Graduate School, Radboud University Medical Center, Nijmegen, The Netherlands
| |
Collapse
|
3
|
Bai J, Jin A, Adams M, Yang C, Nabavi S. Unsupervised feature correlation model to predict breast abnormal variation maps in longitudinal mammograms. Comput Med Imaging Graph 2024; 113:102341. [PMID: 38277769 DOI: 10.1016/j.compmedimag.2024.102341] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/29/2023] [Revised: 01/18/2024] [Accepted: 01/18/2024] [Indexed: 01/28/2024]
Abstract
Breast cancer continues to be a significant cause of mortality among women globally. Timely identification and precise diagnosis of breast abnormalities are critical for enhancing patient prognosis. In this study, we focus on improving the early detection and accurate diagnosis of breast abnormalities, which is crucial for improving patient outcomes and reducing the mortality rate of breast cancer. To address the limitations of traditional screening methods, a novel unsupervised feature correlation network was developed to predict maps indicating breast abnormal variations using longitudinal 2D mammograms. The proposed model utilizes the reconstruction process of current year and prior year mammograms to extract tissue from different areas and analyze the differences between them to identify abnormal variations that may indicate the presence of cancer. The model incorporates a feature correlation module, an attention suppression gate, and a breast abnormality detection module, all working together to improve prediction accuracy. The proposed model not only provides breast abnormal variation maps but also distinguishes between normal and cancer mammograms, making it more advanced compared to the state-of-the-art baseline models. The results of the study show that the proposed model outperforms the baseline models in terms of Accuracy, Sensitivity, Specificity, Dice score, and cancer detection rate.
Collapse
Affiliation(s)
- Jun Bai
- Department of Computer Science and Engineering, University of Connecticut, 371 Fairfield Way, Storrs, CT 06269, USA
| | - Annie Jin
- University of Connecticut School of Medicine, 263 Farmington Ave. Farmington, CT 06030, USA
| | - Madison Adams
- University of Connecticut School of Medicine, 263 Farmington Ave. Farmington, CT 06030, USA
| | - Clifford Yang
- University of Connecticut School of Medicine, 263 Farmington Ave. Farmington, CT 06030, USA; Department of Radiology, UConn Health, 263 Farmington Ave. Farmington, CT 06030, USA
| | - Sheida Nabavi
- Department of Computer Science and Engineering, University of Connecticut, 371 Fairfield Way, Storrs, CT 06269, USA.
| |
Collapse
|
4
|
Hyun CM, Kim TG, Lee K. Unsupervised sequence-to-sequence learning for automatic signal quality assessment in multi-channel electrical impedance-based hemodynamic monitoring. Comput Methods Programs Biomed 2024; 247:108079. [PMID: 38394789 DOI: 10.1016/j.cmpb.2024.108079] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 05/23/2023] [Revised: 11/08/2023] [Accepted: 02/11/2024] [Indexed: 02/25/2024]
Abstract
BACKGROUND AND OBJECTIVE This study proposes an unsupervised sequence-to-sequence learning approach that automatically assesses the motion-induced reliability degradation of the cardiac volume signal (CVS) in multi-channel electrical impedance-based hemodynamic monitoring. The proposed method attempts to tackle shortcomings in existing learning-based assessment approaches, such as the requirement of manual annotation for motion influence and the lack of explicit mechanisms for realizing motion-induced abnormalities under contextual variations in CVS over time. METHOD By utilizing long-short term memory and variational auto-encoder structures, an encoder-decoder model is trained not only to self-reproduce an input sequence of the CVS but also to extrapolate the future in a parallel fashion. By doing so, the model can capture contextual knowledge lying in a temporal CVS sequence while being regularized to explore a general relationship over the entire time-series. A motion-influenced CVS of low-quality is detected, based on the residual between the input sequence and its neural representation with a cut-off value determined from the two-sigma rule of thumb over the training set. RESULT Our experimental observations validated two claims: (i) in the learning environment of label-absence, assessment performance is achievable at a competitive level to the supervised setting, and (ii) the contextual information across a time series of CVS is advantageous for effectively realizing motion-induced unrealistic distortions in signal amplitude and morphology. We also investigated the capability as a pseudo-labeling tool to minimize human-craft annotation by preemptively providing strong candidates for motion-induced anomalies. Empirical evidence has shown that machine-guided annotation can reduce inevitable human-errors during manual assessment while minimizing cumbersome and time-consuming processes. CONCLUSION The proposed method has a particular significance in the industrial field, where it is unavoidable to gather and utilize a large amount of CVS data to achieve high accuracy and robustness in real-world applications.
Collapse
Affiliation(s)
- Chang Min Hyun
- Department of Radiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, USA; School of Mathematics and Computing (Computational Science and Engineering), Yonsei University, Seoul, Republic of Korea.
| | - Tae-Geun Kim
- Department of Physics, Yonsei University, Seoul, Republic of Korea
| | - Kyounghun Lee
- Medical Science Research Institute, Kyung Hee University Medical Center, Seoul 02447, Republic of Korea.
| |
Collapse
|
5
|
Hack JB, Watkins JC, Hammer MF. Machine learning models reveal distinct disease subgroups and improve diagnostic and prognostic accuracy for individuals with pathogenic SCN8A gain-of-function variants. Biol Open 2024:bio.060286. [PMID: 38466077 DOI: 10.1242/bio.060286] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/28/2023] [Accepted: 03/01/2024] [Indexed: 03/12/2024] Open
Abstract
Distinguishing clinical subgroups for patients suffering with diseases characterized by a wide phenotypic spectrum is essential for developing precision therapies. Patients with gain-of-function (GOF) variants in the SCN8A gene exhibit substantial clinical heterogeneity, viewed historically as a linear spectrum ranging from mild to severe. To test for hidden clinical subgroups, we applied two machine learning algorithms to analyze a dataset of patient features collected by the International SCN8A Patient Registry. We utilized two research methodologies: a supervised approach that incorporated feature severity cutoffs based on clinical conventions, and an unsupervised approach employing an entirely data-driven strategy. Both approaches found statistical support for three distinct subgroups and were validated by correlation analyses utilizing external variables. However, distinguishing features of the three subgroups within each approach were not concordant, suggesting a more complex phenotypic landscape. The unsupervised approach yielded strong support for a model involving three partially-ordered subgroups rather than a linear spectrum. Application of these machine-learning approaches may lead to improved prognosis and clinical management of individuals with SCN8A GOF variants and provide insights into the underlying mechanisms of the disease.
Collapse
Affiliation(s)
- Joshua B Hack
- BIO5 Institute, University of Arizona, Tucson AZ, USA
| | - Joseph C Watkins
- Department of Mathematics, University of Arizona, Tucson AZ, USA
| | - Michael F Hammer
- BIO5 Institute, University of Arizona, Tucson AZ, USA
- Neurology Department, University of Arizona, Tucson AZ, USA
| |
Collapse
|
6
|
Fang Y, Yap PT, Lin W, Zhu H, Liu M. Source-free unsupervised domain adaptation: A survey. Neural Netw 2024; 174:106230. [PMID: 38490115 DOI: 10.1016/j.neunet.2024.106230] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/31/2023] [Revised: 01/14/2024] [Accepted: 03/07/2024] [Indexed: 03/17/2024]
Abstract
Unsupervised domain adaptation (UDA) via deep learning has attracted appealing attention for tackling domain-shift problems caused by distribution discrepancy across different domains. Existing UDA approaches highly depend on the accessibility of source domain data, which is usually limited in practical scenarios due to privacy protection, data storage and transmission cost, and computation burden. To tackle this issue, many source-free unsupervised domain adaptation (SFUDA) methods have been proposed recently, which perform knowledge transfer from a pre-trained source model to the unlabeled target domain with source data inaccessible. A comprehensive review of these works on SFUDA is of great significance. In this paper, we provide a timely and systematic literature review of existing SFUDA approaches from a technical perspective. Specifically, we categorize current SFUDA studies into two groups, i.e., white-box SFUDA and black-box SFUDA, and further divide them into finer subcategories based on different learning strategies they use. We also investigate the challenges of methods in each subcategory, discuss the advantages/disadvantages of white-box and black-box SFUDA methods, conclude the commonly used benchmark datasets, and summarize the popular techniques for improved generalizability of models learned without using source data. We finally discuss several promising future directions in this field.
Collapse
Affiliation(s)
- Yuqi Fang
- Department of Radiology and Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, United States
| | - Pew-Thian Yap
- Department of Radiology and Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, United States
| | - Weili Lin
- Department of Radiology and Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, United States
| | - Hongtu Zhu
- Department of Biostatistics and Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, United States
| | - Mingxia Liu
- Department of Radiology and Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, United States.
| |
Collapse
|
7
|
Cui Y, Huang H, Liu J, Zhao M, Li C, Han X, Luo N, Gao J, Yan DM, Zhang C, Jiang T, Yu S. FFCM-MRF: An accurate and generalizable cerebrovascular segmentation pipeline for humans and rhesus monkeys based on TOF-MRA. Comput Biol Med 2024; 170:107996. [PMID: 38266465 DOI: 10.1016/j.compbiomed.2024.107996] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/04/2023] [Revised: 12/14/2023] [Accepted: 01/13/2024] [Indexed: 01/26/2024]
Abstract
PURPOSE Cerebrovascular segmentation and quantification of vascular morphological features in humans and rhesus monkeys are essential for prevention, diagnosis, and treatment of brain diseases. However, current automated whole-brain vessel segmentation methods are often not generalizable to independent datasets, limiting their usefulness in real-world environments with their heterogeneity in participants, scanners, and species. MATERIALS AND METHODS In this study, we proposed an automated, accurate and generalizable segmentation method for magnetic resonance angiography images called FFCM-MRF. This method integrated fast fuzzy c-means clustering and Markov random field optimization by vessel shape priors and spatial constraints. We used a total of 123 human and 44 macaque MRA images scanned at 1.5 T, 3 T, and 7 T MRI from 9 datasets to develop and validate the method. RESULTS FFCM-MRF achieved average Dice similarity coefficients ranging from 69.16 % to 89.63 % across multiple independent datasets, with improvements ranging from 3.24 % to 7.3 % compared to state-of-the-art methods. Quantitative analysis showed that FFCM-MRF can accurately segment major arteries in the Circle of Willis at the base of the brain and small distal pial arteries while effectively reducing noise. Test-retest analysis showed that the model yielded high vascular volume and diameter reliability. CONCLUSIONS Our results have demonstrated that FFCM-MRF is highly accurate and reliable and largely independent of variations in field strength, scanner platforms, acquisition parameters, and species. The macaque MRA data and user-friendly open-source toolbox are freely available at OpenNeuro and GitHub to facilitate studies of imaging biomarkers for cerebrovascular and neurodegenerative diseases.
Collapse
Affiliation(s)
- Yue Cui
- Laboratory of Brain Atlas and Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China; School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China.
| | - Haibin Huang
- Laboratory of Brain Atlas and Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China; School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China
| | - Jialu Liu
- Laboratory of Brain Atlas and Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China; School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China
| | - Mingyang Zhao
- Hong Kong Institute of Science & Innovation, Chinese Academy of Sciences, Hong Kong, China
| | - Chengyi Li
- Laboratory of Brain Atlas and Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China; School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China
| | - Xinyong Han
- Laboratory of Brain Atlas and Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China; School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China
| | - Na Luo
- Brainnetome Center, Laboratory of Brain Atlas and Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China
| | - Jinquan Gao
- Model R&D Center, Beijing Life Biosciences Company Limited, Beijing, China; Technology Management Center, SAFE Pharmaceutical Technology Company Limited, Beijing, China
| | - Dong-Ming Yan
- School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China; State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS), Institute of Automation, Chinese Academy of Sciences, Beijing, China
| | - Chen Zhang
- Department of Neurobiology, School of Basic Medical Sciences, Beijing Key Laboratory of Neural Regeneration and Repair, Advanced Innovation Center for Human Brain Protection, Capital Medical University, Beijing, China
| | - Tianzi Jiang
- School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China; Brainnetome Center, Laboratory of Brain Atlas and Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China; Research Center for Augmented Intelligence, Zhejiang Lab, Hangzhou, China
| | - Shan Yu
- Laboratory of Brain Atlas and Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China; School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China.
| |
Collapse
|
8
|
Wang Y, Hu S, Liu J, Zhong G, Yang C. A multi-module algorithm for heartbeat classification based on unsupervised learning and adaptive feature transfer. Comput Biol Med 2024; 170:108072. [PMID: 38301518 DOI: 10.1016/j.compbiomed.2024.108072] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/16/2023] [Revised: 12/29/2023] [Accepted: 01/27/2024] [Indexed: 02/03/2024]
Abstract
The scarcity of annotated data is a common issue in the realm of heartbeat classification based on deep learning. Transfer learning (TL) has emerged as an effective strategy for addressing this issue. However, current TL techniques in this realm overlook the probability distribution differences between the source domain (SD) and target domain (TD) databases. The motivation of this paper is to address the challenge of labeled data scarcity at the model level while exploring an effective method to eliminate domain discrepancy between SD and TD databases, especially when SD and TD are derived from inconsistent tasks. This study proposes a multi-module heartbeat classification algorithm. Initially, unsupervised feature extractors are designed to extract rich features from unlabeled SD and TD data. Subsequently, a novel adaptive transfer method is proposed to effectively eliminate domain discrepancy between features of SD for pre-training (PTF-SD) and features of TD for fine-tuning (FTF-TD). Finally, the adapted PTF-SD is employed to pre-train a designed classifier, and FTF-TD is used for classifier fine-tuning, with the objective of evaluating the algorithm's performance on the TD task. In our experiments, MNIST-DB serves as the SD database for handwritten digit image classification task, MIT-DB as the TD database for heartbeat classification task. The overall accuracy of classifying heartbeats into normal heartbeats, supraventricular ectopic beats (SVEBs), and ventricular ectopic beats (VEBs) reaches 96.7 %. Specifically, the sensitivity (Sen), positive predictive value (PPV), and F1 score for SVEBs are 0.802, 0.701, and 0.748, respectively. For VEBs, Sen, PPV, and F1 score are 0.976, 0.840, and 0.903, respectively. The results indicate that the proposed multi-module algorithm effectively addresses the challenge labeled data scarcity in heartbeat classification through unsupervised learning and adaptive feature transfer methods.
Collapse
Affiliation(s)
- Yanan Wang
- Center for Biomedical Engineering, School of Information Science and Technology, Fudan University, Shanghai, 200433, China.
| | - Shuaicong Hu
- Center for Biomedical Engineering, School of Information Science and Technology, Fudan University, Shanghai, 200433, China.
| | - Jian Liu
- Center for Biomedical Engineering, School of Information Science and Technology, Fudan University, Shanghai, 200433, China.
| | - Gaoyan Zhong
- Center for Biomedical Engineering, School of Information Science and Technology, Fudan University, Shanghai, 200433, China.
| | - Cuiwei Yang
- Center for Biomedical Engineering, School of Information Science and Technology, Fudan University, Shanghai, 200433, China; Key Laboratory of Medical Imaging Computing and Computer Assisted Intervention of Shanghai, 200093, China.
| |
Collapse
|
9
|
Liu X, Zhang T, Liu M. Joint estimation of pose, depth, and optical flow with a competition-cooperation transformer network. Neural Netw 2024; 171:263-275. [PMID: 38103436 DOI: 10.1016/j.neunet.2023.12.020] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/15/2023] [Revised: 10/31/2023] [Accepted: 12/12/2023] [Indexed: 12/19/2023]
Abstract
Estimating depth, ego-motion, and optical flow from consecutive frames is a critical task in robot navigation and has received significant attention in recent years. In this study, we propose PDF-Former, an unsupervised joint estimation network comprising a full transformer-based framework, as well as a competition and cooperation mechanism. The transformer framework captures global feature dependencies and is customized for different task types, thereby improving the performance of sequential tasks. The competition and cooperation mechanisms enable the network to obtain additional supervisory information at different training stages. Specifically, the competition mechanism is implemented early in training to achieve iterative optimization of 6 DOF poses (rotation and translation information from the target image to the two reference images), the depth of target image, and optical flow (from the target image to the two reference images) estimation in a competitive manner. In contrast, the cooperation mechanism is implemented later in training to facilitate the transmission of results among the three networks and mutually optimize the estimation results. We conducted experiments on the KITTI dataset, and the results indicate that PDF-Former has significant potential to enhance the accuracy and robustness of sequential tasks in robot navigation.
Collapse
Affiliation(s)
- Xiaochen Liu
- School of Instrument Science & Engineering, Southeast University, Nanjing, 210096, China; Key Laboratory of Micro-Inertial Instrument and Advanced Navigation Technology, Ministry of Education, Southeast University, Nanjing, 210096, Jiangsu, China
| | - Tao Zhang
- School of Instrument Science & Engineering, Southeast University, Nanjing, 210096, China; Key Laboratory of Micro-Inertial Instrument and Advanced Navigation Technology, Ministry of Education, Southeast University, Nanjing, 210096, Jiangsu, China.
| | - Mingming Liu
- Department of Orthopedic Surgery, The Second People's Hospital of Lianyungang, Lianyungang, 222003, Jiangsu, China; Department of Orthopedic Surgery, The First People's Hospital of Xining, Xining, 810000, Qinghai, China.
| |
Collapse
|
10
|
Thenier-Villa JL, Martínez-Ricarte FR, Figueroa-Vezirian M, Arikan-Abelló F. Glioblastoma Pseudoprogression Discrimination Using Multiparametric Magnetic Resonance Imaging, Principal Component Analysis, and Supervised and Unsupervised Machine Learning. World Neurosurg 2024; 183:e953-e962. [PMID: 38253179 DOI: 10.1016/j.wneu.2024.01.074] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/01/2023] [Revised: 01/12/2024] [Accepted: 01/13/2024] [Indexed: 01/24/2024]
Abstract
BACKGROUND One of the most frequent phenomena in the follow-up of glioblastoma is pseudoprogression, present in up to half of cases. The clinical usefulness of discriminating this phenomenon through magnetic resonance imaging and nuclear medicine has not yet been standardized; in this study, we used machine learning on multiparametric magnetic resonance imaging to explore discriminators of this phenomenon. METHODS For the study, 30 patients diagnosed with IDH wild-type glioblastoma operated on at both study centers in 2011-2020 were selected; 15 patients corresponded to early tumor progression and 15 patients to pseudoprogression. Using unsupervised learning, the number of clusters and tumor segmentation was recorded using gap-stat and k-means method, adjusting to voxel adjacency. In a second phase, a class prediction was carried out with a multinomial logistic regression supervised learning method; the outcome variables were the percentage of assignment, class overrepresentation, and degree of voxel adjacency. RESULTS Unsupervised learning of the tumor in its diagnosis shows up to 14 well-differentiated tumor areas. In the supervised learning phase, there is a higher percentage of assigned classes (P < 0.01), less overrepresentation of classes (P < 0.01), and greater adjacency (55% vs. 33%) in cases of true tumor progression compared with pseudoprogression. CONCLUSIONS True tumor progression preserves the multidimensional characteristics of the basal tumor at the voxel and region of interest level, resulting in a characteristic differential pattern when supervised learning is used.
Collapse
Affiliation(s)
- José Luis Thenier-Villa
- Department of Neurosurgery, University Hospital Arnau de Vilanova, Lleida, Spain; Department of Neurosurgery, Vall d'Hebron University Hospital, Barcelona, Spain; Neurotrauma and Neurosurgery Research Unit (UNINN), Vall d'Hebron Research Institute (VHIR), Barcelona, Spain.
| | - Francisco Ramón Martínez-Ricarte
- Department of Neurosurgery, Vall d'Hebron University Hospital, Barcelona, Spain; Neurotrauma and Neurosurgery Research Unit (UNINN), Vall d'Hebron Research Institute (VHIR), Barcelona, Spain
| | | | - Fuat Arikan-Abelló
- Department of Neurosurgery, University Hospital Arnau de Vilanova, Lleida, Spain; Department of Neurosurgery, Vall d'Hebron University Hospital, Barcelona, Spain; Neurotrauma and Neurosurgery Research Unit (UNINN), Vall d'Hebron Research Institute (VHIR), Barcelona, Spain
| |
Collapse
|
11
|
Bahrami MK, Nazari S. Digital design of a spatial-pow-STDP learning block with high accuracy utilizing pow CORDIC for large-scale image classifier spatiotemporal SNN. Sci Rep 2024; 14:3388. [PMID: 38337032 PMCID: PMC10858263 DOI: 10.1038/s41598-024-54043-7] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/24/2023] [Accepted: 02/07/2024] [Indexed: 02/12/2024] Open
Abstract
The paramount concern of highly accurate energy-efficient computing in machines with significant cognitive capabilities aims to enhance the accuracy and efficiency of bio-inspired Spiking Neural Networks (SNNs). This paper addresses this main objective by introducing a novel spatial power spike-timing-dependent plasticity (Spatial-Pow-STDP) learning rule as a digital block with high accuracy in a bio-inspired SNN model. Motivated by the demand for precise and accelerated computation that reduces high-cost resources in neural network applications, this paper presents a methodology based on COordinate Rotation DIgital Computer (CORDIC) definitions. The proposed designs of CORDIC algorithms for exponential (Exp CORDIC), natural logarithm (Ln CORDIC), and arbitrary power function (Pow CORDIC) are meticulously detailed and evaluated to ensure optimal acceleration and accuracy, which respectively show average errors near 10-9, 10-6, and 10-5 with 4, 4, and 6 iterations. The engineered architectures for the Exp, Ln, and Pow CORDIC implementations are illustrated and assessed, showcasing the efficiency achieved through high frequency, leading to the introduction of a Spatial-Pow-STDP learning block design based on Pow CORDIC that facilitates efficient and accurate hardware computation with 6.93 × 10-3 average error with 9 iterations. The proposed learning mechanism integrates this structure into a large-scale spatiotemporal SNN consisting of three layers with reduced hyper-parameters, enabling unsupervised training in an event-based paradigm using excitatory and inhibitory synapses. As a result, the application of the developed methodology and equations in the computational SNN model for image classification reveals superior accuracy and convergence speed compared to existing spiking networks by achieving up to 97.5%, 97.6%, 93.4%, and 93% accuracy, respectively, when trained on the MNIST, EMNIST digits, EMNIST letters, and CIFAR10 datasets with 6, 2, 2, and 6 training epochs.
Collapse
Affiliation(s)
| | - Soheila Nazari
- Faculty of Electrical Engineering, Shahid Beheshti University, Tehran, 1983969411, Iran.
| |
Collapse
|
12
|
Nguyen Q, Lejeune E. Segmenting mechanically heterogeneous domains via unsupervised learning. Biomech Model Mechanobiol 2024; 23:349-372. [PMID: 38217746 DOI: 10.1007/s10237-023-01779-2] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/28/2023] [Accepted: 09/30/2023] [Indexed: 01/15/2024]
Abstract
From biological organs to soft robotics, highly deformable materials are essential components of natural and engineered systems. These highly deformable materials can have heterogeneous material properties, and can experience heterogeneous deformations with or without underlying material heterogeneity. Many recent works have established that computational modeling approaches are well suited for understanding and predicting the consequences of material heterogeneity and for interpreting observed heterogeneous strain fields. In particular, there has been significant work toward developing inverse analysis approaches that can convert observed kinematic quantities (e.g., displacement, strain) to material properties and mechanical state. Despite the success of these approaches, they are not necessarily generalizable and often rely on tight control and knowledge of boundary conditions. Here, we will build on the recent advances (and ubiquity) of machine learning approaches to explore alternative approaches to detect patterns in heterogeneous material properties and mechanical behavior. Specifically, we will explore unsupervised learning approaches to clustering and ensemble clustering to identify heterogeneous regions. Overall, we find that these approaches are effective, yet limited in their abilities. Through this initial exploration (where all data and code are published alongside this manuscript), we set the stage for future studies that more specifically adapt these methods to mechanical data.
Collapse
Affiliation(s)
- Quan Nguyen
- Department of Mechanical Engineering, Boston University, Boston, MA, 02215, USA
| | - Emma Lejeune
- Department of Mechanical Engineering, Boston University, Boston, MA, 02215, USA.
| |
Collapse
|
13
|
Deperrois N, Petrovici MA, Senn W, Jordan J. Learning beyond sensations: How dreams organize neuronal representations. Neurosci Biobehav Rev 2024; 157:105508. [PMID: 38097096 DOI: 10.1016/j.neubiorev.2023.105508] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/19/2023] [Revised: 12/05/2023] [Accepted: 12/09/2023] [Indexed: 12/25/2023]
Abstract
Semantic representations in higher sensory cortices form the basis for robust, yet flexible behavior. These representations are acquired over the course of development in an unsupervised fashion and continuously maintained over an organism's lifespan. Predictive processing theories propose that these representations emerge from predicting or reconstructing sensory inputs. However, brains are known to generate virtual experiences, such as during imagination and dreaming, that go beyond previously experienced inputs. Here, we suggest that virtual experiences may be just as relevant as actual sensory inputs in shaping cortical representations. In particular, we discuss two complementary learning principles that organize representations through the generation of virtual experiences. First, "adversarial dreaming" proposes that creative dreams support a cortical implementation of adversarial learning in which feedback and feedforward pathways engage in a productive game of trying to fool each other. Second, "contrastive dreaming" proposes that the invariance of neuronal representations to irrelevant factors of variation is acquired by trying to map similar virtual experiences together via a contrastive learning process. These principles are compatible with known cortical structure and dynamics and the phenomenology of sleep thus providing promising directions to explain cortical learning beyond the classical predictive processing paradigm.
Collapse
Affiliation(s)
| | | | - Walter Senn
- Department of Physiology, University of Bern, Bern, Switzerland
| | - Jakob Jordan
- Department of Physiology, University of Bern, Bern, Switzerland; Electrical Engineering, Yale University, New Haven, CT, United States
| |
Collapse
|
14
|
Choi J, Lee B. Quantitative Topic Analysis of Materials Science Literature Using Natural Language Processing. ACS Appl Mater Interfaces 2024; 16:1957-1968. [PMID: 38059688 DOI: 10.1021/acsami.3c12301] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 12/08/2023]
Abstract
Materials science research has garnered extensive attention from industry, society, policy, and academia. However, understanding the research landscape and extracting strategic insights are challenging due to the increasing diversity and volume of publications. This study proposes a natural language processing-based protocol for extracting text-encoded topics from a large volume of scientific literature, uncovering research interests of scientific communities, as well as convergence trends. We report a topic map, representing the materials science research landscape with text-mined 257 topics regarding biocompatible materials, structural materials, electrochemistry, or photonics. We analyze the topic map in terms of national research interests in materials science, revealing competitive positions and strategies of active nations. For example, it is found that the increasing trend of research interest in machine learning topic was captured in the United States earlier than other nations. Similarly, our journal-level analyses serve as reference information for journal recommendations and trend guidance, showing that the main topics and research interests of materials science journals slightly changed over time. Moreover, we build the topic association network which can highlight the status and future potential of interdisciplinary research, revealing research fields with high centrality in the network such as machine learning-enabled composite modeling, energy policy, or wearable electronics. This study offers insightful results on current and near-future materials science research landscapes, facilitating the understanding of stakeholders, amidst the fast-evolving and diverse knowledge of materials science.
Collapse
Affiliation(s)
- Jaewoong Choi
- Computational Science Research Center, Korea Institute of Science and Technology, Seoul 02792, Republic of Korea
| | - Byungju Lee
- Computational Science Research Center, Korea Institute of Science and Technology, Seoul 02792, Republic of Korea
| |
Collapse
|
15
|
Kong S, Ai J, Lu M, Gong Y. GRAND: GAN-based software runtime anomaly detection method using trace information. Neural Netw 2024; 169:365-377. [PMID: 37924606 DOI: 10.1016/j.neunet.2023.10.036] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/06/2023] [Revised: 08/13/2023] [Accepted: 10/22/2023] [Indexed: 11/06/2023]
Abstract
Software runtime anomaly detection can detect manifestations (known as anomalies) caused by faults in complex systems before they lead to failure. Whereas most existing methods use external performance indicators, this study uses internal execution traces to reveal failures not only related to software performance issues but also functional errors. A neural network model called GRAND, which combines a variational autoencoder and a generative adversarial network, is proposed to mine anomalies in the execution trace. Cassandra, a widely used database system, was used as a representation to conduct the empirical study. The dataset was collected under a well-designed operational profile that contained 5180 time series, each containing more than ten million data points. GRAND achieved a higher detection performance than the other two SOTA baseline models, with a 99% F1-score compared with 93% and 87%. Ablation studies show that the workload information used in GRAND can determine whether the current internal status is consistent with the task, thus achieving a 16% improvement in the F1-score. The attention mechanism used for data fusion can achieve a 32% improvement in the F1-score.
Collapse
Affiliation(s)
- Shiyi Kong
- School of Reliability and Systems Engineering, Beihang University, No. 37, Xueyuan Road, Haidian District, Beijing 100191, China; Beijing Institute of Astronautical Systems Engineering, No. 1, Nandahongmen Road, Fengtai District, Beijing 100076, China.
| | - Jun Ai
- School of Reliability and Systems Engineering, Beihang University, No. 37, Xueyuan Road, Haidian District, Beijing 100191, China.
| | - Minyan Lu
- School of Reliability and Systems Engineering, Beihang University, No. 37, Xueyuan Road, Haidian District, Beijing 100191, China.
| | - Yiang Gong
- School of Reliability and Systems Engineering, Beihang University, No. 37, Xueyuan Road, Haidian District, Beijing 100191, China
| |
Collapse
|
16
|
Aragones DG, Palomino-Segura M, Sicilia J, Crainiciuc G, Ballesteros I, Sánchez-Cabo F, Hidalgo A, Calvo GF. Variable selection for nonlinear dimensionality reduction of biological datasets through bootstrapping of correlation networks. Comput Biol Med 2024; 168:107827. [PMID: 38086138 DOI: 10.1016/j.compbiomed.2023.107827] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/30/2023] [Revised: 11/15/2023] [Accepted: 12/04/2023] [Indexed: 01/10/2024]
Abstract
Identifying the most relevant variables or features in massive datasets for dimensionality reduction can lead to improved and more informative display, faster computation times, and more explainable models of complex systems. Despite significant advances and available algorithms, this task generally remains challenging, especially in unsupervised settings. In this work, we propose a method that constructs correlation networks using all intervening variables and then selects the most informative ones based on network bootstrapping. The method can be applied in both supervised and unsupervised scenarios. We demonstrate its functionality by applying Uniform Manifold Approximation and Projection for dimensionality reduction to several high-dimensional biological datasets, derived from 4D live imaging recordings of hundreds of morpho-kinetic variables, describing the dynamics of thousands of individual leukocytes at sites of prominent inflammation. We compare our method with other standard ones in the field, such as Principal Component Analysis and Elastic Net, showing that it outperforms them. The proposed method can be employed in a wide range of applications, encompassing data analysis and machine learning.
Collapse
Affiliation(s)
- David G Aragones
- Department of Mathematics & MOLAB-Mathematical Oncology Laboratory, Universidad de Castilla-La Mancha, Ciudad Real, Spain
| | - Miguel Palomino-Segura
- Area of Cell and Developmental Biology, Centro Nacional de Investigaciones Cardiovasculares Carlos III, Madrid, Spain; Immunophysiology Research Group, Instituto Universitario de Investigación Biosanitaria de Extremadura (INUBE), Badajoz, Spain; Department of Physiology, Faculty of Sciences, University of Extremadura, Badajoz, Spain
| | - Jon Sicilia
- Area of Cell and Developmental Biology, Centro Nacional de Investigaciones Cardiovasculares Carlos III, Madrid, Spain
| | - Georgiana Crainiciuc
- Area of Cell and Developmental Biology, Centro Nacional de Investigaciones Cardiovasculares Carlos III, Madrid, Spain
| | - Iván Ballesteros
- Area of Cell and Developmental Biology, Centro Nacional de Investigaciones Cardiovasculares Carlos III, Madrid, Spain
| | - Fátima Sánchez-Cabo
- Bioinformatics Unit, Centro Nacional de Investigaciones Cardiovasculares Carlos III, Madrid, Spain
| | - Andrés Hidalgo
- Vascular Biology and Therapeutics Program and Department of Immunobiology, Yale University School of Medicine, New Haven, CT, USA
| | - Gabriel F Calvo
- Department of Mathematics & MOLAB-Mathematical Oncology Laboratory, Universidad de Castilla-La Mancha, Ciudad Real, Spain.
| |
Collapse
|
17
|
Cartucho J, Weld A, Tukra S, Xu H, Matsuzaki H, Ishikawa T, Kwon M, Jang YE, Kim KJ, Lee G, Bai B, Kahrs LA, Boecking L, Allmendinger S, Müller L, Zhang Y, Jin Y, Bano S, Vasconcelos F, Reiter W, Hajek J, Silva B, Lima E, Vilaça JL, Queirós S, Giannarou S. SurgT challenge: Benchmark of soft-tissue trackers for robotic surgery. Med Image Anal 2024; 91:102985. [PMID: 37844472 DOI: 10.1016/j.media.2023.102985] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/28/2023] [Revised: 08/30/2023] [Accepted: 09/28/2023] [Indexed: 10/18/2023]
Abstract
This paper introduces the "SurgT: Surgical Tracking" challenge which was organized in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI 2022). There were two purposes for the creation of this challenge: (1) the establishment of the first standardized benchmark for the research community to assess soft-tissue trackers; and (2) to encourage the development of unsupervised deep learning methods, given the lack of annotated data in surgery. A dataset of 157 stereo endoscopic videos from 20 clinical cases, along with stereo camera calibration parameters, have been provided. Participants were assigned the task of developing algorithms to track the movement of soft tissues, represented by bounding boxes, in stereo endoscopic videos. At the end of the challenge, the developed methods were assessed on a previously hidden test subset. This assessment uses benchmarking metrics that were purposely developed for this challenge, to verify the efficacy of unsupervised deep learning algorithms in tracking soft-tissue. The metric used for ranking the methods was the Expected Average Overlap (EAO) score, which measures the average overlap between a tracker's and the ground truth bounding boxes. Coming first in the challenge was the deep learning submission by ICVS-2Ai with a superior EAO score of 0.617. This method employs ARFlow to estimate unsupervised dense optical flow from cropped images, using photometric and regularization losses. Second, Jmees with an EAO of 0.583, uses deep learning for surgical tool segmentation on top of a non-deep learning baseline method: CSRT. CSRT by itself scores a similar EAO of 0.563. The results from this challenge show that currently, non-deep learning methods are still competitive. The dataset and benchmarking tool created for this challenge have been made publicly available at https://surgt.grand-challenge.org/. This challenge is expected to contribute to the development of autonomous robotic surgery and other digital surgical technologies.
Collapse
Affiliation(s)
- João Cartucho
- The Hamlyn Centre for Robotic Surgery, Imperial College London, United Kingdom.
| | - Alistair Weld
- The Hamlyn Centre for Robotic Surgery, Imperial College London, United Kingdom
| | - Samyakh Tukra
- The Hamlyn Centre for Robotic Surgery, Imperial College London, United Kingdom
| | - Haozheng Xu
- The Hamlyn Centre for Robotic Surgery, Imperial College London, United Kingdom
| | | | | | - Minjun Kwon
- Electronics and Telecommunications Research Institute (ETRI), Daejeon, South Korea
| | - Yong Eun Jang
- Electronics and Telecommunications Research Institute (ETRI), Daejeon, South Korea
| | - Kwang-Ju Kim
- Electronics and Telecommunications Research Institute (ETRI), Daejeon, South Korea
| | - Gwang Lee
- Ajou University, Gyeonggi-do, South Korea
| | - Bizhe Bai
- Medical Computer Vision and Robotics Lab, University of Toronto, Canada
| | - Lueder A Kahrs
- Medical Computer Vision and Robotics Lab, University of Toronto, Canada
| | | | | | | | - Yitong Zhang
- Surgical Robot Vision, University College London, United Kingdom
| | - Yueming Jin
- Surgical Robot Vision, University College London, United Kingdom
| | - Sophia Bano
- Surgical Robot Vision, University College London, United Kingdom
| | | | | | | | - Bruno Silva
- Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal; ICVS/3B's - PT Government Associate Laboratory, Braga/Guimarães, Portugal; 2Ai - School of Technology, IPCA, Barcelos, Portugal
| | - Estevão Lima
- Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal; ICVS/3B's - PT Government Associate Laboratory, Braga/Guimarães, Portugal
| | - João L Vilaça
- 2Ai - School of Technology, IPCA, Barcelos, Portugal
| | - Sandro Queirós
- Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal; ICVS/3B's - PT Government Associate Laboratory, Braga/Guimarães, Portugal
| | - Stamatia Giannarou
- The Hamlyn Centre for Robotic Surgery, Imperial College London, United Kingdom
| |
Collapse
|
18
|
Martini L, Amprimo G, Di Carlo S, Olmo G, Ferraris C, Savino A, Bardini R. Neuronal Spike Shapes (NSS): A straightforward approach to investigate heterogeneity in neuronal excitability states. Comput Biol Med 2024; 168:107783. [PMID: 38056213 DOI: 10.1016/j.compbiomed.2023.107783] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/26/2023] [Revised: 10/23/2023] [Accepted: 11/28/2023] [Indexed: 12/08/2023]
Abstract
The mammalian brain exhibits a remarkable diversity of neurons, contributing to its intricate architecture and functional complexity. The analysis of multimodal single-cell datasets enables the investigation of cell types and states heterogeneity. In this study, we introduce the Neuronal Spike Shapes (NSS), a straightforward approach for the exploration of excitability states of neurons based on their Action Potential (AP) waveforms. The NSS method describes the AP waveform based on a triangular representation complemented by a set of derived electrophysiological (EP) features. To support this hypothesis, we validate the proposed approach on two datasets of murine cortical neurons, focusing it on GABAergic neurons. The validation process involves a combination of NSS-based clustering analysis, features exploration, Differential Expression (DE), and Gene Ontology (GO) enrichment analysis. Results show that the NSS-based analysis captures neuronal excitability states that possess biological relevance independently of cell subtype. In particular, Neuronal Spike Shapes (NSS) captures, among others, a well-characterized fast-spiking excitability state, supported by both electrophysiological and transcriptomic validation. Gene Ontology Enrichment Analysis reveals voltage-gated potassium (K+) channels as specific markers of the identified NSS partitions. This finding strongly corroborates the biological relevance of NSS partitions as excitability states, as the expression of voltage-gated K+ channels regulates the hyperpolarization phase of the AP, being directly implicated in the regulation of neuronal excitability.
Collapse
Affiliation(s)
- Lorenzo Martini
- Politecnico di Torino - Control and Computer Engineering Department, Corso Duca degli Abruzzi, 24, Turin, 10129, Italy.
| | - Gianluca Amprimo
- Politecnico di Torino - Control and Computer Engineering Department, Corso Duca degli Abruzzi, 24, Turin, 10129, Italy; Institute of Electronics, Information Engineering and Telecommunications, National Research Council, Corso Duca degli Abruzzi, 24, Turin, 10029, Italy.
| | - Stefano Di Carlo
- Politecnico di Torino - Control and Computer Engineering Department, Corso Duca degli Abruzzi, 24, Turin, 10129, Italy. https://www.smilies.polito.it
| | - Gabriella Olmo
- Politecnico di Torino - Control and Computer Engineering Department, Corso Duca degli Abruzzi, 24, Turin, 10129, Italy. https://www.sysbio.polito.it/analytics-technologies-health/
| | - Claudia Ferraris
- Institute of Electronics, Information Engineering and Telecommunications, National Research Council, Corso Duca degli Abruzzi, 24, Turin, 10029, Italy. https://www.ieiit.cnr.it/people/Ferraris-Claudia
| | - Alessandro Savino
- Politecnico di Torino - Control and Computer Engineering Department, Corso Duca degli Abruzzi, 24, Turin, 10129, Italy. https://www.smilies.polito.it
| | - Roberta Bardini
- Politecnico di Torino - Control and Computer Engineering Department, Corso Duca degli Abruzzi, 24, Turin, 10129, Italy.
| |
Collapse
|
19
|
Mukashyaka P, Sheridan TB, Foroughi Pour A, Chuang JH. SAMPLER: unsupervised representations for rapid analysis of whole slide tissue images. EBioMedicine 2024; 99:104908. [PMID: 38101298 PMCID: PMC10733087 DOI: 10.1016/j.ebiom.2023.104908] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/23/2023] [Revised: 11/27/2023] [Accepted: 11/27/2023] [Indexed: 12/17/2023] Open
Abstract
BACKGROUND Deep learning has revolutionized digital pathology, allowing automatic analysis of hematoxylin and eosin (H&E) stained whole slide images (WSIs) for diverse tasks. WSIs are broken into smaller images called tiles, and a neural network encodes each tile. Many recent works use supervised attention-based models to aggregate tile-level features into a slide-level representation, which is then used for downstream analysis. Training supervised attention-based models is computationally intensive, architecture optimization of the attention module is non-trivial, and labeled data are not always available. Therefore, we developed an unsupervised and fast approach called SAMPLER to generate slide-level representations. METHODS Slide-level representations of SAMPLER are generated by encoding the cumulative distribution functions of multiscale tile-level features. To assess effectiveness of SAMPLER, slide-level representations of breast carcinoma (BRCA), non-small cell lung carcinoma (NSCLC), and renal cell carcinoma (RCC) WSIs of The Cancer Genome Atlas (TCGA) were used to train separate classifiers distinguishing tumor subtypes in FFPE and frozen WSIs. In addition, BRCA and NSCLC classifiers were externally validated on frozen WSIs. Moreover, SAMPLER's attention maps identify regions of interest, which were evaluated by a pathologist. To determine time efficiency of SAMPLER, we compared runtime of SAMPLER with two attention-based models. SAMPLER concepts were used to improve the design of a context-aware multi-head attention model (context-MHA). FINDINGS SAMPLER-based classifiers were comparable to state-of-the-art attention deep learning models to distinguish subtypes of BRCA (AUC = 0.911 ± 0.029), NSCLC (AUC = 0.940 ± 0.018), and RCC (AUC = 0.987 ± 0.006) on FFPE WSIs (internal test sets). However, training SAMLER-based classifiers was >100 times faster. SAMPLER models successfully distinguished tumor subtypes on both internal and external test sets of frozen WSIs. Histopathological review confirmed that SAMPLER-identified high attention tiles contained subtype-specific morphological features. The improved context-MHA distinguished subtypes of BRCA and RCC (BRCA-AUC = 0.921 ± 0.027, RCC-AUC = 0.988 ± 0.010) with increased accuracy on internal test FFPE WSIs. INTERPRETATION Our unsupervised statistical approach is fast and effective for analyzing WSIs, with greatly improved scalability over attention-based deep learning methods. The high accuracy of SAMPLER-based classifiers and interpretable attention maps suggest that SAMPLER successfully encodes the distinct morphologies within WSIs and will be applicable to general histology image analysis problems. FUNDING This study was supported by the National Cancer Institute (Grant No. R01CA230031 and P30CA034196).
Collapse
Affiliation(s)
- Patience Mukashyaka
- The Jackson Laboratory for Genomic Medicine, Farmington, CT, USA; Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, CT, USA
| | - Todd B Sheridan
- The Jackson Laboratory for Genomic Medicine, Farmington, CT, USA; Department of Pathology, Hartford Hospital, Hartford, CT, USA
| | | | - Jeffrey H Chuang
- The Jackson Laboratory for Genomic Medicine, Farmington, CT, USA; Department of Genetics and Genome Sciences, University of Connecticut Health Center, Farmington, CT, USA.
| |
Collapse
|
20
|
Dou Y, Xia J, Fu M, Cai Y, Meng X, Zhan Y. Identification of epileptic networks with graph convolutional network incorporating oscillatory activities and evoked synaptic responses. Neuroimage 2023; 284:120439. [PMID: 37939889 DOI: 10.1016/j.neuroimage.2023.120439] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/11/2023] [Revised: 10/01/2023] [Accepted: 10/31/2023] [Indexed: 11/10/2023] Open
Abstract
Stereoelectroencephalography (SEEG) offers unique neural data from in-depth brain structures with fine temporal resolutions to better investigate the origin of epileptic brain activities. Although oscillatory patterns from different frequency bands and functional connectivity computed from the SEEG datasets are employed to study the epileptic zones, direct electrical stimulation-evoked electrophysiological recordings of synaptic responses, namely cortical-cortical evoked potentials (CCEPs), from the same SEEG electrodes are not explored for the localization of epileptic zones. Here we proposed a two-stream model with unsupervised learning and graph convolutional network tailored to the SEEG and CCEP datasets in individual patients to perform localization of epileptic zones. We compared our localization results with the clinically marked electrode sites determined for surgical resections. Our model had good classification capability when compared to other state-of-the-art methods. Furthermore, based on our prediction results we performed group-level brain-area mapping analysis for temporal, frontal and parietal epilepsy patients and found that epileptic and non-epileptic brain networks were distinct in patients with different types of focal epilepsy. Our unsupervised data-driven model provides personalized localization analysis for the epileptic zones. The epileptic and non-epileptic brain areas disclosed by the prediction model provide novel insights into the network-level pathological characteristics of epilepsy.
Collapse
Affiliation(s)
- Yonglin Dou
- The Brain Cognition and Brain Disease Institute, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| | - Jing Xia
- The Brain Cognition and Brain Disease Institute, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China; CAS Key Laboratory of Brain Connectome and Manipulation, The Brain Cognition and Brain Disease Institute, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| | - Mengmeng Fu
- Department of Neurosurgery, Shenzhen University General Hospital, Shenzhen University, Shenzhen, China
| | - Yunpeng Cai
- Institute of Advanced Computing and Digital Engineering, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
| | - Xianghong Meng
- Department of Neurosurgery, Shenzhen University General Hospital, Shenzhen University, Shenzhen, China.
| | - Yang Zhan
- The Brain Cognition and Brain Disease Institute, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China; CAS Key Laboratory of Brain Connectome and Manipulation, The Brain Cognition and Brain Disease Institute, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China; Shenzhen-Hong Kong Institute of Brain Science-Shenzhen Fundamental Research Institutions, Shenzhen 518055, China.
| |
Collapse
|
21
|
Abstract
BACKGROUND Biclustering is increasingly used in biomedical data analysis, recommendation tasks, and text mining domains, with hundreds of biclustering algorithms proposed. When assessing the performance of these algorithms, more than real datasets are required as they do not offer a solid ground truth. Synthetic data surpass this limitation by producing reference solutions to be compared with the found patterns. However, generating synthetic datasets is challenging since the generated data must ensure reproducibility, pattern representativity, and real data resemblance. RESULTS We propose G-Bic, a dataset generator conceived to produce synthetic benchmarks for the normative assessment of biclustering algorithms. Beyond expanding on aspects of pattern coherence, data quality, and positioning properties, it further handles specificities related to mixed-type datasets and time-series data.G-Bic has the flexibility to replicate real data regularities from diverse domains. We provide the default configurations to generate reproducible benchmarks to evaluate and compare diverse aspects of biclustering algorithms. Additionally, we discuss empirical strategies to simulate the properties of real data. CONCLUSION G-Bic is a parametrizable generator for biclustering analysis, offering a solid means to assess biclustering solutions according to internal and external metrics robustly.
Collapse
Affiliation(s)
- Eduardo N Castanho
- LASIGE, Faculdade de Ciências, Universidade de Lisboa, Campo Grande 016, 1749-016, Lisbon, Portugal.
| | - João P Lobo
- LASIGE, Faculdade de Ciências, Universidade de Lisboa, Campo Grande 016, 1749-016, Lisbon, Portugal
| | - Rui Henriques
- INESC-ID, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais 1, 1900-001, Lisbon, Portugal
| | - Sara C Madeira
- LASIGE, Faculdade de Ciências, Universidade de Lisboa, Campo Grande 016, 1749-016, Lisbon, Portugal
| |
Collapse
|
22
|
Abstract
Learners flexibly update category boundaries to adjust to the range of experiences they encounter. However, little is known about whether the degree of flexibility is consistent across domains. We examined whether categorization of social input, specifically emotions, is afforded more flexibility as compared to other biological input. To address this question, children (6-12 years; 32 female, 37 male; 7 Hispanic or Latino, 62 not Hispanic or Latino; 8 Black or African American, 14 multiracial, 46 White, 1 selected "other") categorized faces morphed from calm to upset and animals morphed from a horse to a cow across task phases that differed in the distribution of stimuli presented. Learners flexibly adjusted both emotion and animal category boundaries according to distributional information, yet children showed more flexibility when updating their category boundaries for emotions. These results provide support for the idea that children-who must adjust to the vast and varied emotional signals of their social partners-respond to social signals dynamically in order to make predictions about the internal states and future behaviors of others.
Collapse
Affiliation(s)
- Rista C. Plate
- Department of Psychology, University of Pennsylvania, 3720 Walnut St, Philadelphia, PA 19104 USA
| | - Kristina Woodard
- Department of Psychology, University of Wisconsin-Madison, 1202 West Johnson Street, Madison, WI 53706 USA
| | - Seth D. Pollak
- Department of Psychology, University of Wisconsin-Madison, 1202 West Johnson Street, Madison, WI 53706 USA
| |
Collapse
|
23
|
Kascenas A, Sanchez P, Schrempf P, Wang C, Clackett W, Mikhael SS, Voisey JP, Goatman K, Weir A, Pugeault N, Tsaftaris SA, O'Neil AQ. The role of noise in denoising models for anomaly detection in medical images. Med Image Anal 2023; 90:102963. [PMID: 37769551 DOI: 10.1016/j.media.2023.102963] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/15/2022] [Revised: 08/22/2023] [Accepted: 09/07/2023] [Indexed: 10/03/2023]
Abstract
Pathological brain lesions exhibit diverse appearance in brain images, in terms of intensity, texture, shape, size, and location. Comprehensive sets of data and annotations are difficult to acquire. Therefore, unsupervised anomaly detection approaches have been proposed using only normal data for training, with the aim of detecting outlier anomalous voxels at test time. Denoising methods, for instance classical denoising autoencoders (DAEs) and more recently emerging diffusion models, are a promising approach, however naive application of pixelwise noise leads to poor anomaly detection performance. We show that optimization of the spatial resolution and magnitude of the noise improves the performance of different model training regimes, with similar noise parameter adjustments giving good performance for both DAEs and diffusion models. Visual inspection of the reconstructions suggests that the training noise influences the trade-off between the extent of the detail that is reconstructed and the extent of erasure of anomalies, both of which contribute to better anomaly detection performance. We validate our findings on two real-world datasets (tumor detection in brain MRI and hemorrhage/ischemia/tumor detection in brain CT), showing good detection on diverse anomaly appearances. Overall, we find that a DAE trained with coarse noise is a fast and simple method that gives state-of-the-art accuracy. Diffusion models applied to anomaly detection are as yet in their infancy and provide a promising avenue for further research. Code for our DAE model and coarse noise is provided at: https://github.com/AntanasKascenas/DenoisingAE.
Collapse
Affiliation(s)
- Antanas Kascenas
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom; University of Glasgow, Glasgow G12 8QQ, United Kingdom.
| | - Pedro Sanchez
- University of Edinburgh, Kings Buildings, Edinburgh EH9 3FG, United Kingdom
| | - Patrick Schrempf
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom
| | - Chaoyang Wang
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom
| | - William Clackett
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom
| | - Shadia S Mikhael
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom
| | - Jeremy P Voisey
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom
| | - Keith Goatman
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom
| | - Alexander Weir
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom
| | | | - Sotirios A Tsaftaris
- University of Edinburgh, Kings Buildings, Edinburgh EH9 3FG, United Kingdom; The Alan Turing Institute, London, United Kingdom
| | - Alison Q O'Neil
- Canon Medical Research Europe, Bonnington Bond, 2 Anderson Pl, Edinburgh EH6 5NP, United Kingdom; University of Edinburgh, Kings Buildings, Edinburgh EH9 3FG, United Kingdom
| |
Collapse
|
24
|
Wang SC, Ting CK, Chen CY, Liu C, Lin NC, Loong CC, Wu HT, Lin YT. Arterial blood pressure waveform in liver transplant surgery possesses variability of morphology reflecting recipients' acuity and predicting short term outcomes. J Clin Monit Comput 2023; 37:1521-1531. [PMID: 37436598 DOI: 10.1007/s10877-023-01047-9] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/05/2023] [Accepted: 06/13/2023] [Indexed: 07/13/2023]
Abstract
We investigated clinical information underneath the beat-to-beat fluctuation of the arterial blood pressure (ABP) waveform morphology. We proposed the Dynamical Diffusion Map algorithm (DDMap) to quantify the variability of morphology. The underlying physiology could be the compensatory mechanisms involving complex interactions between various physiological mechanisms to regulate the cardiovascular system. As a liver transplant surgery contains distinct periods, we investigated its clinical behavior in different surgical steps. Our study used DDmap algorithm, based on unsupervised manifold learning, to obtain a quantitative index for the beat-to-beat variability of morphology. We examined the correlation between the variability of ABP morphology and disease acuity as indicated by Model for End-Stage Liver Disease (MELD) scores, the postoperative laboratory data, and 4 early allograft failure (EAF) scores. Among the 85 enrolled patients, the variability of morphology obtained during the presurgical phase was best correlated with MELD-Na scores. The neohepatic phase variability of morphology was associated with EAF scores as well as postoperative bilirubin levels, international normalized ratio, aspartate aminotransferase levels, and platelet count. Furthermore, variability of morphology presents more associations with the above clinical conditions than the common BP measures and their BP variability indices. The variability of morphology obtained during the presurgical phase is indicative of patient acuity, whereas those during the neohepatic phase are indicative of short-term surgical outcomes.
Collapse
Affiliation(s)
- Shen-Chih Wang
- Department of Anesthesiology, Taipei Veterans General Hospital, Taipei, Taiwan
- School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan
| | - Chien-Kun Ting
- Department of Anesthesiology, Taipei Veterans General Hospital, Taipei, Taiwan
- School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan
| | - Cheng-Yen Chen
- School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan
- Division of Transplantation Surgery, Taipei Veterans General Hospital, Taipei, Taiwan
| | - Chinsu Liu
- School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan
- Division of Transplantation Surgery, Taipei Veterans General Hospital, Taipei, Taiwan
| | - Niang-Cheng Lin
- School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan
- Division of Transplantation Surgery, Taipei Veterans General Hospital, Taipei, Taiwan
| | - Che-Chuan Loong
- School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan
- Division of Transplantation Surgery, Taipei Veterans General Hospital, Taipei, Taiwan
| | - Hau-Tieng Wu
- Department of Mathematics, Duke University, Durham, NC, USA.
- Department of Statistical Science, Duke University, Durham, NC, USA.
| | - Yu-Ting Lin
- Department of Anesthesiology, Taipei Veterans General Hospital, Taipei, Taiwan.
- School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan.
| |
Collapse
|
25
|
Slator PJ, Cromb D, Jackson LH, Ho A, Counsell SJ, Story L, Chappell LC, Rutherford M, Hajnal JV, Hutter J, Alexander DC. Non-invasive mapping of human placenta microenvironments throughout pregnancy with diffusion-relaxation MRI. Placenta 2023; 144:29-37. [PMID: 37952367 DOI: 10.1016/j.placenta.2023.11.002] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 05/10/2023] [Revised: 10/13/2023] [Accepted: 11/01/2023] [Indexed: 11/14/2023]
Abstract
INTRODUCTION In-vivo measurements of placental structure and function have the potential to improve prediction, diagnosis, and treatment planning for a wide range of pregnancy complications, such as fetal growth restriction and pre-eclampsia, and hence inform clinical decision making, ultimately improving patient outcomes. MRI is emerging as a technique with increased sensitivity to placental structure and function compared to the current clinical standard, ultrasound. METHODS We demonstrate and evaluate a combined diffusion-relaxation MRI acquisition and analysis pipeline on a sizable cohort of 78 normal pregnancies with gestational ages ranging from 15 + 5 to 38 + 4 weeks. Our acquisition comprises a combined T2*-diffusion MRI acquisition sequence - which is simultaneously sensitive to oxygenation, microstructure and microcirculation. We analyse our scans with a data-driven unsupervised machine learning technique, InSpect, that parsimoniously identifies distinct components in the data. RESULTS We identify and map seven potential placental microenvironments and reveal detailed insights into multiple microstructural and microcirculatory features of the placenta, and assess their trends across gestation. DISCUSSION By demonstrating direct observation of micro-scale placental structure and function, and revealing clear trends across pregnancy, our work contributes towards the development of robust imaging biomarkers for pregnancy complications and the ultimate goal of a normative model of placental development.
Collapse
Affiliation(s)
- Paddy J Slator
- Cardiff University Brain Research Imaging Centre, School of Psychology, Maindy Road, Cardiff, CF24 4HQ, UK; School of Computer Science and Informatics, Cardiff University, Cardiff, UK; Centre for Medical Image Computing and Department of Computer Science, University College London, London, UK.
| | - Daniel Cromb
- Centre for the Developing Brain, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK
| | - Laurence H Jackson
- Centre for the Developing Brain, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK; Biomedical Engineering Department, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK
| | - Alison Ho
- Department of Women and Children's Health, School of Life Course Sciences, King's College London, London, UK
| | - Serena J Counsell
- Centre for the Developing Brain, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK
| | - Lisa Story
- Department of Women and Children's Health, School of Life Course Sciences, King's College London, London, UK
| | - Lucy C Chappell
- Department of Women and Children's Health, School of Life Course Sciences, King's College London, London, UK
| | - Mary Rutherford
- Centre for the Developing Brain, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK
| | - Joseph V Hajnal
- Centre for the Developing Brain, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK; Biomedical Engineering Department, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK
| | - Jana Hutter
- Centre for the Developing Brain, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK; Biomedical Engineering Department, School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK
| | - Daniel C Alexander
- Centre for Medical Image Computing and Department of Computer Science, University College London, London, UK
| |
Collapse
|
26
|
Chen YT, Gao LL. Testing for a difference in means of a single feature after clustering. ArXiv 2023:arXiv:2311.16375v1. [PMID: 38076519 PMCID: PMC10705581] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Figures] [Subscribe] [Scholar Register] [Indexed: 12/23/2023]
Abstract
For many applications, it is critical to interpret and validate groups of observations obtained via clustering. A common validation approach involves testing differences in feature means between observations in two estimated clusters. In this setting, classical hypothesis tests lead to an inflated Type I error rate. To overcome this problem, we propose a new test for the difference in means in a single feature between a pair of clusters obtained using hierarchical or k-means clustering. The test based on the proposed p-value controls the selective Type I error rate in finite samples and can be efficiently computed. We further illustrate the validity and power of our proposal in simulation and demonstrate its use on single-cell RNA-sequencing data.
Collapse
Affiliation(s)
- Yiqun T Chen
- Department of Biomedical Data Science, Stanford University
| | - Lucy L Gao
- Department of Statistics, University of British Columbia, November 29, 2023
| |
Collapse
|
27
|
Ilagan MJ, Falk CF. Model-agnostic unsupervised detection of bots in a Likert-type questionnaire. Behav Res Methods 2023:10.3758/s13428-023-02246-7. [PMID: 37985637 DOI: 10.3758/s13428-023-02246-7] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 09/07/2023] [Indexed: 11/22/2023]
Abstract
To detect bots in online survey data, there is a wealth of literature on statistical detection using only responses to Likert-type items. There are two traditions in the literature. One tradition requires labeled data, forgoing strong model assumptions. The other tradition requires a measurement model, forgoing collection of labeled data. In the present article, we consider the problem where neither requirement is available, for an inventory that has the same number of Likert-type categories for all items. We propose a bot detection algorithm that is both model-agnostic and unsupervised. Our proposed algorithm involves a permutation test with leave-one-out calculations of outlier statistics. For each respondent, it outputs a p value for the null hypothesis that the respondent is a bot. Such an algorithm offers nominal sensitivity calibration that is robust to the bot response distribution. In a simulation study, we found our proposed algorithm to improve upon naive alternatives in terms of 95% sensitivity calibration and, in many scenarios, in terms of classification accuracy.
Collapse
Affiliation(s)
- Michael John Ilagan
- Department of Psychology, McGill University, 2001 McGill College, 7th Floor, H3A 1G1, Montreal, QC, Canada
| | - Carl F Falk
- Department of Psychology, McGill University, 2001 McGill College, 7th Floor, H3A 1G1, Montreal, QC, Canada.
| |
Collapse
|
28
|
Biswas R, Thoma M, Kong X. Functional data analysis to characterize disease patterns in frequent longitudinal data: application to bacterial vaginal microbiota patterns using weekly Nugent scores and identification of pattern-specific risk factors. BMC Med Res Methodol 2023; 23:251. [PMID: 37884907 PMCID: PMC10604810 DOI: 10.1186/s12874-023-02063-8] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/30/2023] [Accepted: 10/10/2023] [Indexed: 10/28/2023] Open
Abstract
BACKGROUND Technology advancement has allowed more frequent monitoring of biomarkers. The resulting data structure entails more frequent follow-ups compared to traditional longitudinal studies where the number of follow-up is often small. Such data allow explorations of the role of intra-person variability in understanding disease etiology and characterizing disease processes. A specific example was to characterize pathogenesis of bacterial vaginosis (BV) using weekly vaginal microbiota Nugent assay scores collected over 2 years in post-menarcheeal women from Rakai, Uganda, and to identify risk factors for each vaginal microbiota pattern to inform epidemiological and etiological understanding of the pathogenesis of BV. METHODS We use a fully data-driven approach to characterize the longitudinal patters of vaginal microbiota by considering the densely sampled Nugent scores to be random functions over time and performing dimension reduction by functional principal components. Extending a current functional data clustering method, we use a hierarchical functional clustering framework considering multiple data features to help identify clinically meaningful patterns of vaginal microbiota fluctuations. Additionally, multinomial logistic regression was used to identify risk factors for each vaginal microbiota pattern to inform epidemiological and etiological understanding of the pathogenesis of BV. RESULTS Using weekly Nugent scores over 2 years of 211 sexually active and post-menarcheal women in Rakai, four patterns of vaginal microbiota variation were identified: persistent with a BV state (high Nugent scores), persistent with normal ranged Nugent scores, large fluctuation of Nugent scores which however are predominantly in the BV state; large fluctuation of Nugent scores but predominantly the scores are in the normal state. Higher Nugent score at the start of an interval, younger age group of less than 20 years, unprotected source for bathing water, a woman's partner's being not circumcised, use of injectable/Norplant hormonal contraceptives for family planning were associated with higher odds of persistent BV in women. CONCLUSION The hierarchical functional data clustering method can be used for fully data driven unsupervised clustering of densely sampled longitudinal data to identify clinically informative clusters and risk-factors associated with each cluster.
Collapse
Affiliation(s)
| | - Marie Thoma
- University of Maryland, College Park, MD, USA
| | | |
Collapse
|
29
|
Li D, Peng Y, Sun J, Guo Y. Unsupervised deep consistency learning adaptation network for cardiac cross-modality structural segmentation. Med Biol Eng Comput 2023; 61:2713-2732. [PMID: 37450212 DOI: 10.1007/s11517-023-02833-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/24/2022] [Accepted: 04/05/2023] [Indexed: 07/18/2023]
Abstract
Deep neural networks have recently been succeessful in the field of medical image segmentation; however, they are typically subject to performance degradation problems when well-trained models are tested in another new domain with different data distributions. Given that annotated cross-domain images may inaccessible, unsupervised domain adaptation methods that transfer learnable information from annotated source domains to unannotated target domains with different distributions have attracted substantial attention. Many methods leverage image-level or pixel-level translation networks to align domain-invariant information and mitigate domain shift issues. However, These methods rarely perform well when there is a large domain gap. A new unsupervised deep consistency learning adaptation network, which adopts input space consistency learning and output space consistency learning to realize unsupervised domain adaptation and cardiac structural segmentation, is introduced in this paper The framework mainly includes a domain translation path and a cross-modality segmentation path. In domain translation path, a symmetric alignment generator network with attention to cross-modality features and anatomy is introduced to align bidirectional domain features. In the segmentation path, entropy map minimization, output probability map minimization and segmentation prediction minimization are leveraged to align the output space features. The model conducts supervised learning to extract source domain features and conducts unsupervised deep consistency learning to extract target domain features. Through experimental testing on two challenging cross-modality segmentation tasks, our method has robust performance compared to that of previous methods. Furthermore, ablation experiments are conducted to confirm the effectiveness of our framework.
Collapse
Affiliation(s)
- Dapeng Li
- College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, 266590, Shandong, China
| | - Yanjun Peng
- College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, 266590, Shandong, China.
- Shandong Province Key Laboratory of Wisdom Mining Information Technology, Shandong University of Science and Technology, Qingdao, 266590, Shandong, China.
| | - Jindong Sun
- College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, 266590, Shandong, China
| | - Yanfei Guo
- College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, 266590, Shandong, China
| |
Collapse
|
30
|
Yang A, Liu Y, Wang J, Li X, Cao J, Ji Z, Pang Y. Visual-quality-driven unsupervised image dehazing. Neural Netw 2023; 167:1-9. [PMID: 37598543 DOI: 10.1016/j.neunet.2023.08.010] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/19/2022] [Revised: 05/11/2023] [Accepted: 08/06/2023] [Indexed: 08/22/2023]
Abstract
Most of the existing learning-based dehazing methods require a diverse and large collection of paired hazy/clean images, which is intractable to obtain. Therefore, existing dehazing methods resort to training on synthetic images. This may result in a possible domain shift when treating real scenes. In this paper, we propose a novel unsupervised dehazing (lightweight) network without any reference images to directly predict clear images from the original hazy images, which consists of an interactive fusion module (IFM) and an iterative optimization module (IOM). Specifically, IFM interactively fuses multi-level features to make up for the missing information among deep and shallow features while IOM iteratively optimizes dehazed results to obtain pleasing visual effects. Particularly, based on the observation that hazy images usually suffer from quality degradation, four non-reference visual-quality-driven loss functions are designed to enable the network trained in an unsupervised way, including dark channel loss, contrast loss, saturation loss, and edge sharpness loss. Extensive experiments on two synthetic datasets and one real-world dataset demonstrate that our method performs favorably against the state-of-the-art unsupervised dehazing methods and even matches some supervised methods in terms of metrics such as PSNR, SSIM, and UQI.
Collapse
Affiliation(s)
- Aiping Yang
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China; Shanghai Artificial Intelligence Laboratory, China
| | - Yumeng Liu
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| | - Jinbin Wang
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China.
| | - Xiaoxiao Li
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| | - Jiale Cao
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| | - Zhong Ji
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| | - Yanwei Pang
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| |
Collapse
|
31
|
Riza LS, Zain MI, Izzuddin A, Prasetyo Y, Hidayat T, Abu Samah KAF. Implementation of machine learning in DNA barcoding for determining the plant family taxonomy. Heliyon 2023; 9:e20161. [PMID: 37767518 PMCID: PMC10520734 DOI: 10.1016/j.heliyon.2023.e20161] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/09/2022] [Revised: 09/05/2023] [Accepted: 09/13/2023] [Indexed: 09/29/2023] Open
Abstract
The DNA barcoding approach has been used extensively in taxonomy and phylogenetics. The differences in certain DNA sequences are able to differentiate and help classify organisms into taxa. It has been used in cases of taxonomic disputes where morphology by itself is insufficient. This research aimed to utilize hierarchical clustering, an unsupervised machine learning method, to determine and resolve disputes in plant family taxonomy. We take a case study of Leguminosae that historically some classify into three families (Fabaceae, Caesalpiniaceae, and Mimosaceae) but others classify into one family (Leguminosae). This study is divided into several phases, which are: (i) data collection, (ii) data preprocessing, (iii) finding the best distance method, and (iv) determining disputed family. The data used are collected from several sources, including National Center for Biotechnology Information (NCBI), journals, and websites. The data for validation of the methods were collected from NCBI. This was used to determine the best distance method for differentiating families or genera. The data for the case study in the Leguminosae group was collected from journals and a website. From the experiment that we have conducted, we found that the Pearson method is the best distance method to do clustering ITS sequence of plants, both in accuracy and computational cost. We use the Pearson method to determine the disputed family between Leguminosae. We found that the case study of Leguminosae should be grouped into one family based on our research.
Collapse
Affiliation(s)
- Lala Septem Riza
- Department of Computer Science Education, Universitas Pendidikan Indonesia, Bandung, Indonesia
| | - Muhammad Iqbal Zain
- Department of Computer Science Education, Universitas Pendidikan Indonesia, Bandung, Indonesia
| | - Ahmad Izzuddin
- Department of Computer Science Education, Universitas Pendidikan Indonesia, Bandung, Indonesia
| | - Yudi Prasetyo
- Department of Computer Science Education, Universitas Pendidikan Indonesia, Bandung, Indonesia
| | - Topik Hidayat
- Department of Biology Education, Universitas Pendidikan Indonesia, Bandung, Indonesia
| | | |
Collapse
|
32
|
Liu Z, Wang Q, Ma A, Feng S, Chung D, Zhao J, Ma Q, Liu B. Inference of disease-associated microbial gene modules based on metagenomic and metatranscriptomic data. Comput Biol Med 2023; 165:107458. [PMID: 37703713 DOI: 10.1016/j.compbiomed.2023.107458] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/10/2023] [Revised: 08/22/2023] [Accepted: 09/04/2023] [Indexed: 09/15/2023]
Abstract
The identification of microbial characteristics associated with diseases is crucial for disease diagnosis and therapy. However, the presence of heterogeneity, high dimensionality, and large amounts of microbial data presents tremendous challenges in discovering key microbial features. In this paper, we present IDAM, a novel computational method for inferring disease-associated gene modules from metagenomic and metatranscriptomic data. This method integrates gene context conservation (uber-operons) and regulatory mechanisms (gene co-expression patterns) within a mathematical graph model to explore gene modules associated with specific diseases. It alleviates reliance on prior meta-data. We applied IDAM to publicly available datasets from inflammatory bowel disease, melanoma, type 1 diabetes mellitus, and irritable bowel syndrome. The results demonstrated the superior performance of IDAM in inferring disease-associated characteristics compared to existing popular tools. Furthermore, we showcased the high reproducibility of the gene modules inferred by IDAM using independent cohorts with inflammatory bowel disease. We believe that IDAM can be a highly advantageous method for exploring disease-associated microbial characteristics. The source code of IDAM is freely available at https://github.com/OSU-BMBL/IDAM, and the web server can be accessed at https://bmblx.bmi.osumc.edu/idam/.
Collapse
Affiliation(s)
- Zhaoqian Liu
- School of Mathematics, Shandong University, Jinan, Shandong, 250100, China
| | - Qi Wang
- School of Mathematics, Shandong University, Jinan, Shandong, 250100, China
| | - Anjun Ma
- Department of Biomedical Informatics, The Ohio State University, Columbus, OH, 43210, USA
| | - Shaohong Feng
- Department of Biomedical Informatics, The Ohio State University, Columbus, OH, 43210, USA
| | - Dongjun Chung
- Department of Biomedical Informatics, The Ohio State University, Columbus, OH, 43210, USA; Pelotonia Institute for Immuno-Oncology, The Ohio State University, Columbus, OH, 43210, USA
| | - Jing Zhao
- Department of Biomedical Informatics, The Ohio State University, Columbus, OH, 43210, USA
| | - Qin Ma
- Department of Biomedical Informatics, The Ohio State University, Columbus, OH, 43210, USA; Pelotonia Institute for Immuno-Oncology, The Ohio State University, Columbus, OH, 43210, USA.
| | - Bingqiang Liu
- School of Mathematics, Shandong University, Jinan, Shandong, 250100, China; Shandong National Center for Applied Mathematics, Jinan, Shandong, 250100, China.
| |
Collapse
|
33
|
Suesserman M, Gorny S, Lasaga D, Helms J, Olson D, Bowen E, Bhattacharya S. Procedure code overutilization detection from healthcare claims using unsupervised deep learning methods. BMC Med Inform Decis Mak 2023; 23:196. [PMID: 37770866 PMCID: PMC10536726 DOI: 10.1186/s12911-023-02268-3] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/17/2022] [Accepted: 08/17/2023] [Indexed: 09/30/2023] Open
Abstract
BACKGROUND Fraud, Waste, and Abuse (FWA) in medical claims have a negative impact on the quality and cost of healthcare. A major component of FWA in claims is procedure code overutilization, where one or more prescribed procedures may not be relevant to a given diagnosis and patient profile, resulting in unnecessary and unwarranted treatments and medical payments. This study aims to identify such unwarranted procedures from millions of healthcare claims. In the absence of labeled examples of unwarranted procedures, the study focused on the application of unsupervised machine learning techniques. METHODS Experiments were conducted with deep autoencoders to find claims containing anomalous procedure codes indicative of FWA, and were compared against a baseline density-based clustering model. Diagnoses, procedures, and demographic data associated with healthcare claims were used as features for the models. A dataset of one hundred thousand claims sampled from a larger claims database is used to initially train and tune the models, followed by experimentations on a dataset with thirty-three million claims. Experimental results show that the autoencoder model, when trained with a novel feature-weighted loss function, outperforms the density-based clustering approach in finding potential outlier procedure codes. RESULTS Given the unsupervised nature of our experiments, model performance was evaluated using a synthetic outlier test dataset, and a manually annotated outlier test dataset. Precision, recall and F1-scores on the synthetic outlier test dataset for the autoencoder model trained on one hundred thousand claims were 0.87, 1.0 and 0.93, respectively, while the results for these metrics on the manually annotated outlier test dataset were 0.36, 0.86 and 0.51, respectively. The model performance on the manually annotated outlier test dataset improved further when trained on the larger thirty-three million claims dataset with precision, recall and F1-scores of 0.48, 0.90 and 0.63, respectively. CONCLUSIONS This study demonstrates the feasibility of leveraging unsupervised, deep-learning methods to identify potential procedure overutilization from healthcare claims.
Collapse
Affiliation(s)
| | - Samantha Gorny
- Program Integrity, Deloitte & Touche LLP, New York, NY, USA
| | - Daniel Lasaga
- Program Integrity, Deloitte & Touche LLP, New York, NY, USA
| | - John Helms
- AI Center of Excellence, Deloitte & Touche LLP, New York, NY, USA
| | - Dan Olson
- Program Integrity, Deloitte & Touche LLP, New York, NY, USA
| | - Edward Bowen
- AI Center of Excellence, Deloitte & Touche LLP, New York, NY, USA
| | | |
Collapse
|
34
|
Sutherland S, Egger B, Tenenbaum J. Building 3D Generative Models from Minimal Data. Int J Comput Vis 2023; 132:555-580. [PMID: 38303742 PMCID: PMC10827923 DOI: 10.1007/s11263-023-01870-2] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/03/2022] [Accepted: 07/28/2023] [Indexed: 02/03/2024]
Abstract
We propose a method for constructing generative models of 3D objects from a single 3D mesh and improving them through unsupervised low-shot learning from 2D images. Our method produces a 3D morphable model that represents shape and albedo in terms of Gaussian processes. Whereas previous approaches have typically built 3D morphable models from multiple high-quality 3D scans through principal component analysis, we build 3D morphable models from a single scan or template. As we demonstrate in the face domain, these models can be used to infer 3D reconstructions from 2D data (inverse graphics) or 3D data (registration). Specifically, we show that our approach can be used to perform face recognition using only a single 3D template (one scan total, not one per person). We extend our model to a preliminary unsupervised learning framework that enables the learning of the distribution of 3D faces using one 3D template and a small number of 2D images. Our approach is motivated as a potential model for the origins of face perception in human infants, who appear to start with an innate face template and subsequently develop a flexible system for perceiving the 3D structure of any novel face from experience with only 2D images of a relatively small number of familiar faces.
Collapse
Affiliation(s)
- Skylar Sutherland
- Department of Psychology, Yale University, 1 Hillhouse Ave, New Haven, CT 06511 USA
- Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139 USA
| | - Bernhard Egger
- Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139 USA
- Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, Cauerstraße 11, 91058 Erlangen, Bavaria Germany
| | - Joshua Tenenbaum
- Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139 USA
| |
Collapse
|
35
|
Dong J, Cheng G, Zhang Y, Peng C, Song Y, Tong R, Lin L, Chen YW. Tailored multi-organ segmentation with model adaptation and ensemble. Comput Biol Med 2023; 166:107467. [PMID: 37725849 DOI: 10.1016/j.compbiomed.2023.107467] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/13/2023] [Revised: 08/10/2023] [Accepted: 09/04/2023] [Indexed: 09/21/2023]
Abstract
Multi-organ segmentation, which identifies and separates different organs in medical images, is a fundamental task in medical image analysis. Recently, the immense success of deep learning motivated its wide adoption in multi-organ segmentation tasks. However, due to expensive labor costs and expertise, the availability of multi-organ annotations is usually limited and hence poses a challenge in obtaining sufficient training data for deep learning-based methods. In this paper, we aim to address this issue by combining off-the-shelf single-organ segmentation models to develop a multi-organ segmentation model on the target dataset, which helps get rid of the dependence on annotated data for multi-organ segmentation. To this end, we propose a novel dual-stage method that consists of a Model Adaptation stage and a Model Ensemble stage. The first stage enhances the generalization of each off-the-shelf segmentation model on the target domain, while the second stage distills and integrates knowledge from multiple adapted single-organ segmentation models. Extensive experiments on four abdomen datasets demonstrate that our proposed method can effectively leverage off-the-shelf single-organ segmentation models to obtain a tailored model for multi-organ segmentation with high accuracy.
Collapse
Affiliation(s)
- Jiahua Dong
- College of Computer Science and Technology, Zhejiang University, Hangzhou, 310027, China
| | - Guohua Cheng
- College of Computer Science and Technology, Zhejiang University, Hangzhou, 310027, China
| | - Yue Zhang
- Center for Medical Imaging, Robotics, Analytic Computing & Learning (MIRACLE), Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, 215163, China; School of Biomedical Engineering, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, Anhui, 230026, China.
| | - Chengtao Peng
- Department of Electronic Engineering and Information Science, University of Science and Technology of China, Hefei, 230026, China
| | - Yu Song
- Graduate School of Information Science and Engineering, Ritsumeikan University, Shiga, 525-8577, Japan
| | - Ruofeng Tong
- College of Computer Science and Technology, Zhejiang University, Hangzhou, 310027, China
| | - Lanfen Lin
- College of Computer Science and Technology, Zhejiang University, Hangzhou, 310027, China
| | - Yen-Wei Chen
- Graduate School of Information Science and Engineering, Ritsumeikan University, Shiga, 525-8577, Japan
| |
Collapse
|
36
|
Kurucu MC, Rekik I. Graph neural network based unsupervised influential sample selection for brain multigraph population fusion. Comput Med Imaging Graph 2023; 108:102274. [PMID: 37531812 DOI: 10.1016/j.compmedimag.2023.102274] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/27/2023] [Revised: 07/17/2023] [Accepted: 07/17/2023] [Indexed: 08/04/2023]
Abstract
Graph neural networks (GNNs) have witnessed remarkable proliferation due to the increasing number of applications where data is represented as graphs. GNN-based multigraph population fusion methods for estimating population representative connectional brain templates (CBT) have recently led to improvements, especially in network neuroscience. However, prior studies do not consider how an individual training brain multigraph influences the quality of GNN training for brain multigraph population fusion. To address this issue, we propose two major sample selection methods to quantify the influence of a training brain multigraph on the brain multigraph population fusion task using GNNs, in a fully unsupervised manner: (1) GraphGradIn, in which we use gradients w.r.t GNN weights to trace changes in the centeredness loss of connectional brain template during the training phase; (2) GraphTestIn, in which we exclude a training brain multigraph of interest during the refinement process in the test phase to infer its influence on the CBT centeredness loss. Next, we select the most influential multigraphs to build the training set for brain multigraph population fusion into a CBT. We conducted extensive experiments on brain multigraph datasets to show that using a dataset of influential training samples improves the learned connectional brain template in terms of centeredness, discriminativeness, and topological soundness. Finally, we demonstrate the use of our methods by discovering the connectional fingerprints of healthy and neurologically disordered brain multigraph populations including Alzheimer's disease and Autism spectrum disorder patients. Our source code is available at https://github.com/basiralab/GraphGradIn.
Collapse
Affiliation(s)
- Mert Can Kurucu
- BASIRA Lab, Imperial-X and Computing Department, Imperial College London, London, UK; Istanbul Technical University, Istanbul, Turkey
| | - Islem Rekik
- BASIRA Lab, Imperial-X and Computing Department, Imperial College London, London, UK.
| |
Collapse
|
37
|
Gao CX, Dwyer D, Zhu Y, Smith CL, Du L, Filia KM, Bayer J, Menssink JM, Wang T, Bergmeir C, Wood S, Cotton SM. An overview of clustering methods with guidelines for application in mental health research. Psychiatry Res 2023; 327:115265. [PMID: 37348404 DOI: 10.1016/j.psychres.2023.115265] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 12/15/2022] [Revised: 05/20/2023] [Accepted: 05/21/2023] [Indexed: 06/24/2023]
Abstract
Cluster analyzes have been widely used in mental health research to decompose inter-individual heterogeneity by identifying more homogeneous subgroups of individuals. However, despite advances in new algorithms and increasing popularity, there is little guidance on model choice, analytical framework and reporting requirements. In this paper, we aimed to address this gap by introducing the philosophy, design, advantages/disadvantages and implementation of major algorithms that are particularly relevant in mental health research. Extensions of basic models, such as kernel methods, deep learning, semi-supervised clustering, and clustering ensembles are subsequently introduced. How to choose algorithms to address common issues as well as methods for pre-clustering data processing, clustering evaluation and validation are then discussed. Importantly, we also provide general guidance on clustering workflow and reporting requirements. To facilitate the implementation of different algorithms, we provide information on R functions and libraries.
Collapse
Affiliation(s)
- Caroline X Gao
- Centre for Youth Mental Health, The University of Melbourne, Parkville, VIC, Australia; Orygen, Parkville, VIC, Australia; Department of Epidemiology and Preventative Medicine, School of Public Health and Preventive Medicine, Monash University, Melbourne, VIC, Australia.
| | - Dominic Dwyer
- Centre for Youth Mental Health, The University of Melbourne, Parkville, VIC, Australia; Orygen, Parkville, VIC, Australia
| | - Ye Zhu
- School of Information Technology, Deakin University, Geelong, VIC, Australia
| | - Catherine L Smith
- Department of Epidemiology and Preventative Medicine, School of Public Health and Preventive Medicine, Monash University, Melbourne, VIC, Australia
| | - Lan Du
- Faculty of Information Technology, Monash University, Clayton, VIC, Australia
| | - Kate M Filia
- Centre for Youth Mental Health, The University of Melbourne, Parkville, VIC, Australia; Orygen, Parkville, VIC, Australia
| | - Johanna Bayer
- Centre for Youth Mental Health, The University of Melbourne, Parkville, VIC, Australia; Orygen, Parkville, VIC, Australia
| | - Jana M Menssink
- Centre for Youth Mental Health, The University of Melbourne, Parkville, VIC, Australia; Orygen, Parkville, VIC, Australia
| | - Teresa Wang
- Faculty of Information Technology, Monash University, Clayton, VIC, Australia
| | - Christoph Bergmeir
- Faculty of Information Technology, Monash University, Clayton, VIC, Australia; Department of Computer Science and Artificial Intelligence, University of Granada, Granada, Spain
| | - Stephen Wood
- Centre for Youth Mental Health, The University of Melbourne, Parkville, VIC, Australia; Orygen, Parkville, VIC, Australia
| | - Sue M Cotton
- Centre for Youth Mental Health, The University of Melbourne, Parkville, VIC, Australia; Orygen, Parkville, VIC, Australia
| |
Collapse
|
38
|
Veillon R, Shabushnig J, Aabye-Hansen L, Duvinage M, Eckstein C, Li Z, Sardella A, Soto M, Torres JD, Turnquist B. Applying Machine Learning to the Visual Inspection of Filled Injectable Drug Products. PDA J Pharm Sci Technol 2023; 77:376-401. [PMID: 37321861 DOI: 10.5731/pdajpst.2022.012796] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/10/2022] [Accepted: 06/07/2023] [Indexed: 06/17/2023]
Abstract
With machine learning (ML), we see the potential to better harness the intelligence and decision-making abilities of human inspectors performing manual visual inspection (MVI) and apply this to automated visual inspection (AVI) with the inherent improvements in throughput and consistency. This article is intended to capture current experience with this new technology and provides points to consider for successful application to AVI of injectable drug products. The technology is available today for such AVI applications. Machine vision companies have integrated ML as an additional visual inspection tool with minimal upgrades to existing hardware. Studies have demonstrated superior results in defect detection and reduction in false rejects, when compared with conventional inspection tools. ML implementation does not require modifications to current AVI qualification strategies. The utilization of this technology for AVI will accelerate recipe development by use of faster computers rather than by direct human configuration and coding of vision tools. By freezing the model developed with artificial intelligence tools and subjecting it to current validation strategies, assurance of reliable performance in the production environment can be achieved.
Collapse
Affiliation(s)
| | | | | | | | | | - Zheng Li
- Genentech, South San Francisco, CA
| | | | | | | | | |
Collapse
|
39
|
Moon J, Posada-Quintero HF, Chon KH. Genetic data visualization using literature text-based neural networks: Examples associated with myocardial infarction. Neural Netw 2023; 165:562-595. [PMID: 37364469 DOI: 10.1016/j.neunet.2023.05.015] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/17/2022] [Revised: 04/11/2023] [Accepted: 05/09/2023] [Indexed: 06/28/2023]
Abstract
Data visualization is critical to unraveling hidden information from complex and high-dimensional data. Interpretable visualization methods are critical, especially in the biology and medical fields, however, there are limited effective visualization methods for large genetic data. Current visualization methods are limited to lower-dimensional data and their performance suffers if there is missing data. In this study, we propose a literature-based visualization method to reduce high-dimensional data without compromising the dynamics of the single nucleotide polymorphisms (SNP) and textual interpretability. Our method is innovative because it is shown to (1) preserves both global and local structures of SNP while reducing the dimension of the data using literature text representations, and (2) enables interpretable visualizations using textual information. For performance evaluations, we examined the proposed approach to classify various classification categories including race, myocardial infarction event age groups, and sex using several machine learning models on the literature-derived SNP data. We used visualization approaches to examine clustering of data as well as quantitative performance metrics for the classification of the risk factors examined above. Our method outperformed all popular dimensionality reduction and visualization methods for both classification and visualization, and it is robust against missing and higher-dimensional data. Moreover, we found it feasible to incorporate both genetic and other risk information obtained from literature with our method.
Collapse
Affiliation(s)
- Jihye Moon
- Department of Biomedical Engineering, University of Connecticut, Storrs, CT 06269, USA.
| | | | - Ki H Chon
- Department of Biomedical Engineering, University of Connecticut, Storrs, CT 06269, USA.
| |
Collapse
|
40
|
Tang K, Zhang S, Wang Y, Zhang X, Liu Z, Liang Z, Wang H, Chen L, Chen W, Qi L. Learning spatially variant degradation for unsupervised blind photoacoustic tomography image restoration. Photoacoustics 2023; 32:100536. [PMID: 37575971 PMCID: PMC10413197 DOI: 10.1016/j.pacs.2023.100536] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 05/05/2023] [Revised: 07/18/2023] [Accepted: 07/19/2023] [Indexed: 08/15/2023]
Abstract
Photoacoustic tomography (PAT) images contain inherent distortions due to the imaging system and heterogeneous tissue properties. Improving image quality requires the removal of these system distortions. While model-based approaches and data-driven techniques have been proposed for PAT image restoration, achieving accurate and robust image recovery remains challenging. Recently, deep-learning-based image deconvolution approaches have shown promise for image recovery. However, PAT imaging presents unique challenges, including spatially varying resolution and the absence of ground truth data. Consequently, there is a pressing need for a novel learning strategy specifically tailored for PAT imaging. Herein, we propose a configurable network model named Deep hybrid Image-PSF Prior (DIPP) that builds upon the physical image degradation model of PAT. DIPP is an unsupervised and deeply learned network model that aims to extract the ideal PAT image from complex system degradation. Our DIPP framework captures the degraded information solely from the acquired PAT image, without relying on ground truth or labeled data for network training. Additionally, we can incorporate the experimentally measured Point Spread Functions (PSFs) of the specific PAT system as a reference to further enhance performance. To evaluate the algorithm's effectiveness in addressing multiple degradations in PAT, we conduct extensive experiments using simulation images, publicly available datasets, phantom images, and in vivo small animal imaging data. Comparative analyses with classical analytical methods and state-of-the-art deep learning models demonstrate that our DIPP approach achieves significantly improved restoration results in terms of image details and contrast.
Collapse
Affiliation(s)
- Kaiyi Tang
- School of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, Guangdong, China
| | - Shuangyang Zhang
- School of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, Guangdong, China
| | - Yang Wang
- School of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, Guangdong, China
| | - Xiaoming Zhang
- School of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, Guangdong, China
| | - Zhenyang Liu
- School of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, Guangdong, China
| | - Zhichao Liang
- School of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, Guangdong, China
| | - Huafeng Wang
- Research Center of Narrative Medicine, Shunde Hospital, Southern Medical University, Foshan, Guangdong, China
| | - Lingjian Chen
- Research Center of Narrative Medicine, Shunde Hospital, Southern Medical University, Foshan, Guangdong, China
| | - Wufan Chen
- School of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, Guangdong, China
| | - Li Qi
- School of Biomedical Engineering, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China
- Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, Guangdong, China
| |
Collapse
|
41
|
Briscik M, Dillies MA, Déjean S. Improvement of variables interpretability in kernel PCA. BMC Bioinformatics 2023; 24:282. [PMID: 37438763 DOI: 10.1186/s12859-023-05404-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/27/2023] [Accepted: 06/27/2023] [Indexed: 07/14/2023] Open
Abstract
BACKGROUND Kernel methods have been proven to be a powerful tool for the integration and analysis of high-throughput technologies generated data. Kernels offer a nonlinear version of any linear algorithm solely based on dot products. The kernelized version of principal component analysis is a valid nonlinear alternative to tackle the nonlinearity of biological sample spaces. This paper proposes a novel methodology to obtain a data-driven feature importance based on the kernel PCA representation of the data. RESULTS The proposed method, kernel PCA Interpretable Gradient (KPCA-IG), provides a data-driven feature importance that is computationally fast and based solely on linear algebra calculations. It has been compared with existing methods on three benchmark datasets. The accuracy obtained using KPCA-IG selected features is equal to or greater than the other methods' average. Also, the computational complexity required demonstrates the high efficiency of the method. An exhaustive literature search has been conducted on the selected genes from a publicly available Hepatocellular carcinoma dataset to validate the retained features from a biological point of view. The results once again remark on the appropriateness of the computed ranking. CONCLUSIONS The black-box nature of kernel PCA needs new methods to interpret the original features. Our proposed methodology KPCA-IG proved to be a valid alternative to select influential variables in high-dimensional high-throughput datasets, potentially unravelling new biological and medical biomarkers.
Collapse
Affiliation(s)
- Mitja Briscik
- Institut de Mathématiques de Toulouse, UMR5219, CNRS, UPS, Université de Toulouse, Cedex 9, 31062, Toulouse, France.
| | - Marie-Agnès Dillies
- Institut Pasteur, Université Paris Cité, Bioinformatics and Biostatistics Hub, F-75015, Paris, France
| | - Sébastien Déjean
- Institut de Mathématiques de Toulouse, UMR5219, CNRS, UPS, Université de Toulouse, Cedex 9, 31062, Toulouse, France
| |
Collapse
|
42
|
Xu Z, Dai Y, Liu F, Chen W, Liu Y, Shi L, Liu S, Zhou Y. Swin MAE: Masked autoencoders for small datasets. Comput Biol Med 2023; 161:107037. [PMID: 37230020 DOI: 10.1016/j.compbiomed.2023.107037] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/17/2023] [Revised: 04/23/2023] [Accepted: 05/11/2023] [Indexed: 05/27/2023]
Abstract
The development of deep learning models in medical image analysis is majorly limited by the lack of large-sized and well-annotated datasets. Unsupervised learning does not require labels and is more suitable for solving medical image analysis problems. However, most unsupervised learning methods must be applied to large datasets. To make unsupervised learning applicable to small datasets, we proposed Swin MAE, a masked autoencoder with Swin Transformer as its backbone. Even on a dataset of only a few thousand medical images, Swin MAE can still learn useful semantic features purely from images without using any pre-trained models. It can equal or even slightly outperform the supervised model obtained by Swin Transformer trained on ImageNet in the transfer learning results of downstream tasks. Compared to MAE, Swin MAE brought a performance improvement of twice and five times for downstream tasks on BTCV and our parotid dataset, respectively. The code is publicly available at https://github.com/Zian-Xu/Swin-MAE.
Collapse
Affiliation(s)
- Zi'an Xu
- Northeastern University, Shenyang, China
| | - Yin Dai
- Northeastern University, Shenyang, China.
| | - Fayu Liu
- China Medical University, Shenyang, China
| | | | - Yue Liu
- Northeastern University, Shenyang, China
| | - Lifu Shi
- Liaoning Jiayin Medical Technology Co., China
| | - Sheng Liu
- China Medical University, Shenyang, China
| | | |
Collapse
|
43
|
Froese T, Weber N, Shpurov I, Ikegami T. From autopoiesis to self-optimization: Toward an enactive model of biological regulation. Biosystems 2023:104959. [PMID: 37380066 DOI: 10.1016/j.biosystems.2023.104959] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/03/2023] [Revised: 06/08/2023] [Accepted: 06/15/2023] [Indexed: 06/30/2023]
Abstract
The theory of autopoiesis has been influential in many areas of theoretical biology, especially in the fields of artificial life and origins of life. However, it has not managed to productively connect with mainstream biology, partly for theoretical reasons, but arguably mainly because deriving specific working hypotheses has been challenging. The theory has recently undergone significant conceptual development in the enactive approach to life and mind. Hidden complexity in the original conception of autopoiesis has been explicated in the service of other operationalizable concepts related to self-individuation: precariousness, adaptivity, and agency. Here we advance these developments by highlighting the interplay of these concepts with considerations from thermodynamics: reversibility, irreversibility, and path-dependence. We interpret this interplay in terms of the self-optimization model, and present modeling results that illustrate how these minimal conditions enable a system to re-organize itself such that it tends toward coordinated constraint satisfaction at the system level. Although the model is still very abstract, these results point in a direction where the enactive approach could productively connect with cell biology.
Collapse
Affiliation(s)
- Tom Froese
- Embodied Cognitive Science Unit, Okinawa Institute of Science and Technology Graduate University, Tancha, Okinawa, Japan.
| | - Natalya Weber
- Embodied Cognitive Science Unit, Okinawa Institute of Science and Technology Graduate University, Tancha, Okinawa, Japan
| | - Ivan Shpurov
- Embodied Cognitive Science Unit, Okinawa Institute of Science and Technology Graduate University, Tancha, Okinawa, Japan
| | - Takashi Ikegami
- Theoretical Sciences Visiting Program, Okinawa Institute of Science and Technology Graduate University, Tancha, Okinawa, Japan; Ikegami Lab, Department of General Systems Studies, University of Tokyo, Komaba, Tokyo, Japan
| |
Collapse
|
44
|
Mattioni L, Ferri F, Nikčević AV, Spada MM, Sestieri C. Twisted memories: Addiction-related engrams are strengthened by desire thinking. Addict Behav 2023; 145:107782. [PMID: 37348176 DOI: 10.1016/j.addbeh.2023.107782] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/18/2022] [Revised: 06/08/2023] [Accepted: 06/15/2023] [Indexed: 06/24/2023]
Abstract
Associative learning plays a central role in addiction by reinforcing associations between environmental cues and addiction-related information. Unsupervised learning models posit that memories are adjusted based on how strongly these representations are coactivated during the retrieval process. From a different perspective, clinical models of addiction posit that the escalation and persistence of craving may depend on desire thinking, a thinking style orienting to prefigure information about positive addiction-related experiences. In the present work, we tested the main hypothesis that desire thinking is a key factor in the strengthening of addiction-related associations. A group of adult smoking volunteers (N = 26) engaged in a period of desire thinking before performing an associative learning task in which neutral words (cues) were shown along with images (smoking-related vs. neutral context) at different frequencies. Two retrieval tests were administered, one immediately after encoding and the other after 24 h, to test how the recall of associations changed as a function of retention interval. Two control groups, smokers (N = 21) and non-smokers (N = 22), performed a similar procedure, with a neutral imagination task replacing desire thinking. Participants who engaged in desire thinking increased their performance from the first to the second retrieval test only for the most frequent smoking-related associations. Crucially, this selective effect was not observed in the two control groups. These results provide behavioral evidence in support of the idea that desire thinking plays a role in strengthening addiction-related associations. Thus, this thinking process may be considered a target for reconsolidation-based conceptualizations of, and treatments for, addiction.
Collapse
Affiliation(s)
- Lorenzo Mattioni
- Department of Neuroscience, Imaging and Clinical Sciences - and ITAB, Institute for Advanced Biomedical Technologies, G. d'Annunzio University, Chieti, Italy.
| | - Francesca Ferri
- Department of Neuroscience, Imaging and Clinical Sciences - and ITAB, Institute for Advanced Biomedical Technologies, G. d'Annunzio University, Chieti, Italy
| | - Ana V Nikčević
- Department of Psychology, Kingston University, Kingston upon Thames, UK
| | | | - Carlo Sestieri
- Department of Neuroscience, Imaging and Clinical Sciences - and ITAB, Institute for Advanced Biomedical Technologies, G. d'Annunzio University, Chieti, Italy
| |
Collapse
|
45
|
Shokoohi F, Khaniki SH. Uncovering Alterations in Cancer Epigenetics via Trans-Dimensional Markov Chain Monte Carlo and Hidden Markov Models. bioRxiv 2023:2023.06.15.545168. [PMID: 37398181 PMCID: PMC10312753 DOI: 10.1101/2023.06.15.545168] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 07/04/2023]
Abstract
Epigenetic alterations are key drivers in the development and progression of cancer. Identifying differentially methylated cytosines (DMCs) in cancer samples is a crucial step toward understanding these changes. In this paper, we propose a trans-dimensional Markov chain Monte Carlo (TMCMC) approach that uses hidden Markov models (HMMs) with binomial emission, and bisulfite sequencing (BS-Seq) data, called DMCTHM, to identify DMCs in cancer epigenetic studies. We introduce the Expander-Collider penalty to tackle under and over-estimation in TMCMC-HMMs. We address all known challenges inherent in BS-Seq data by introducing novel approaches for capturing functional patterns and autocorrelation structure of the data, as well as for handling missing values, multiple covariates, multiple comparisons, and family-wise errors. We demonstrate the effectiveness of DMCTHM through comprehensive simulation studies. The results show that our proposed method outperforms other competing methods in identifying DMCs. Notably, with DMCTHM, we uncovered new DMCs and genes in Colorectal cancer that were significantly enriched in the Tp53 pathway.
Collapse
Affiliation(s)
- Farhad Shokoohi
- Department of Mathematical Sciences, University of Nevada-Las Vegas, Las Vega, NV 89154, USA
| | - Saeedeh Hajebi Khaniki
- Department of Biostatistics, School of Health, Mashhad University of Medical Sciences, Mashhad, Iran
| |
Collapse
|
46
|
Song L, Ma M, Liu G. TS-Net: Two-stage deformable medical image registration network based on new smooth constraints. Magn Reson Imaging 2023; 99:26-33. [PMID: 36709011 DOI: 10.1016/j.mri.2023.01.013] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/22/2022] [Revised: 05/27/2022] [Accepted: 01/14/2023] [Indexed: 01/27/2023]
Abstract
Medical image registration can establish the spatial consistency of the corresponding anatomical structures between different medical images, which is important in medical image analysis. In recent years, with the rapid development of deep learning, the image registration methods based on deep learning greatly improve the speed, accuracy, and robustness of registration. Regrettably, these methods typically do not work well for large deformations and complex deformations in the image, and neglect to preserve the topological properties of the image during deformation. Aiming at these problems, we propose a new network TS-Net that learns deformation from coarse to fine and transmits information of different scales in the two stages. Two-stage network learning deformation from coarse to fine can gradually learn the large and complex deformations in images. In the second stage, the feature maps downsampled in the first stage for skip connection can expand the local receptive field and obtain more local information. The smooth constraints function used in the past is to impose the same restriction on the global, which is not targeted. In this paper, we propose a new smooth constraints function for each voxel deformation, which can better ensure the smoothness of the transformation and maintain the topological properties of the image. The experiments on brain datasets with complex deformations and heart datasets with large deformations show that our proposed method achieves better results while maintaining the topological properties of deformations compared to existing deep learning-based registration methods.
Collapse
Affiliation(s)
- Lei Song
- College of Computer Science and Technology, Jilin University, Changchun 130012, Jilin, PR China; Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, Jilin, PR China.
| | - Mingrui Ma
- College of Computer Science and Technology, Jilin University, Changchun 130012, Jilin, PR China; Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, Jilin, PR China.
| | - Guixia Liu
- College of Computer Science and Technology, Jilin University, Changchun 130012, Jilin, PR China; Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, Jilin, PR China.
| |
Collapse
|
47
|
Zhu W, Qiu P, Dumitrascu OM, Sobczak JM, Farazi M, Yang Z, Nandakumar K, Wang Y. OTRE: Where Optimal Transport Guided Unpaired Image-to-Image Translation Meets Regularization by Enhancing. Inf Process Med Imaging 2023; 13939:415-427. [PMID: 37426457 PMCID: PMC10329768 DOI: 10.1007/978-3-031-34048-2_32] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 07/11/2023]
Abstract
Non-mydriatic retinal color fundus photography (CFP) is widely available due to the advantage of not requiring pupillary dilation, however, is prone to poor quality due to operators, systemic imperfections, or patient-related causes. Optimal retinal image quality is mandated for accurate medical diagnoses and automated analyses. Herein, we leveraged the Optimal Transport (OT) theory to propose an unpaired image-to-image translation scheme for mapping low-quality retinal CFPs to high-quality counterparts. Furthermore, to improve the flexibility, robustness, and applicability of our image enhancement pipeline in the clinical practice, we generalized a state-of-the-art model-based image reconstruction method, regularization by denoising, by plugging in priors learned by our OT-guided image-to-image translation network. We named it as regularization by enhancing (RE). We validated the integrated framework, OTRE, on three publicly available retinal image datasets by assessing the quality after enhancement and their performance on various downstream tasks, including diabetic retinopathy grading, vessel segmentation, and diabetic lesion segmentation. The experimental results demonstrated the superiority of our proposed framework over some state-of-the-art unsupervised competitors and a state-of-the-art supervised method.
Collapse
Affiliation(s)
- Wenhui Zhu
- School of Computing and Augmented Intelligence, Arizona State Univ., AZ, USA
| | - Peijie Qiu
- McKeley School of Engineering, Washington Univ. in St. Louis, St. Louis, MO, USA
| | | | | | - Mohammad Farazi
- School of Computing and Augmented Intelligence, Arizona State Univ., AZ, USA
| | - Zhangsihao Yang
- School of Computing and Augmented Intelligence, Arizona State Univ., AZ, USA
| | - Keshav Nandakumar
- School of Computing and Augmented Intelligence, Arizona State Univ., AZ, USA
| | - Yalin Wang
- School of Computing and Augmented Intelligence, Arizona State Univ., AZ, USA
| |
Collapse
|
48
|
Zhu S, Zheng W, Pang H. CPAE: Contrastive predictive autoencoder for unsupervised pre-training in health status prediction. Comput Methods Programs Biomed 2023; 234:107484. [PMID: 37030137 DOI: 10.1016/j.cmpb.2023.107484] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 12/12/2022] [Revised: 02/20/2023] [Accepted: 03/12/2023] [Indexed: 06/19/2023]
Abstract
BACKGROUND AND OBJECTIVE Fully-supervised learning approaches have shown promising results in some health status prediction tasks using Electronic Health Records (EHRs). These traditional approaches rely on sufficient labeled data to learn from. However, in practice, acquiring large-scaled labeled medical data for various prediction tasks is often not feasible. Thus, it is of great interest to utilize contrastive pre-training to leverage the unlabeled information. METHODS In this work, we propose a novel data-efficient framework, contrastive predictive autoencoder (CPAE), to first learn without labels from the EHR data in the pre-training process, and then fine-tune on the downstream tasks. Our framework comprises of two parts: (i) a contrastive learning process, inherited from contrastive predictive coding (CPC), which aims to extract global slow-varying features, and (ii) a reconstruction process, which forces the encoder to capture local features. We also introduce the attention mechanism in one variant of our framework to balance the above two processes. RESULTS Experiments on real-world EHR dataset verify the effectiveness of our proposed framework on two downstream tasks (i.e., in-hospital mortality prediction and length-of-stay prediction), compared to their supervised counterparts, the CPC model, and other baseline models. CONCLUSIONS By comprising of both contrastive learning components and reconstruction components, CPAE aims to extract both global slow-varying information and local transient information. The best results on two downstream tasks are all achieved by CPAE. The variant AtCPAE is particularly superior when fine-tuned on very small training data. Further work may incorporate techniques of multi-task learning to optimize the pre-training process of CPAEs. Moreover, this work is based on the benchmark MIMIC-III dataset which only includes 17 variables. Future work may extend to a larger number of variables.
Collapse
Affiliation(s)
- Shuying Zhu
- Li Ka Shing Faculty of Medicine, the University of Hong Kong, Hong Kong SAR, China.
| | - Weizhong Zheng
- Li Ka Shing Faculty of Medicine, the University of Hong Kong, Hong Kong SAR, China.
| | - Herbert Pang
- Li Ka Shing Faculty of Medicine, the University of Hong Kong, Hong Kong SAR, China; Department of Biostatistics and Bioinformatics, Duke University School of Medicine, NC, USA.
| |
Collapse
|
49
|
Ke J, Shen Y, Lu Y, Guo Y, Shen D. Mine local homogeneous representation by interaction information clustering with unsupervised learning in histopathology images. Comput Methods Programs Biomed 2023; 235:107520. [PMID: 37031665 DOI: 10.1016/j.cmpb.2023.107520] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Subscribe] [Scholar Register] [Received: 11/10/2022] [Revised: 03/13/2023] [Accepted: 03/28/2023] [Indexed: 05/08/2023]
Abstract
BACKGROUND AND OBJECTIVE The success of data-driven deep learning for histopathology images often depends on high-quality training sets and fine-grained annotations. However, as tumors are heterogeneous and annotations are expensive, unsupervised learning approaches are desirable to obtain full automation. METHODS In this paper, an Interaction Information Clustering (IIC) method is proposed to extract locally homogeneous features in mutually exclusive clusters. Trained in an unsupervised paradigm, the framework learns invariant information from multiple spatially adjacent regions for improved classification. Additionally, an adaptive Conditional Random Field (CRF) model is incorporated to detect spatially adjacent image patches of high morphological homogeneity in an offset-constraint free manner. RESULTS Empirically, the proposed model achieves an observable improvement of 11.4% on the downstream patch-level classification accuracy, compared with state-of-the-art unsupervised learning approaches. CONCLUSION Furthermore, evaluated with our clinically collected histopathology whole-slide images, the proposed model shows high consistency in tissue distribution compared with well-trained supervised learning, which is of important diagnostic significance in clinical practice.
Collapse
Affiliation(s)
- Jing Ke
- School of Electronic Information and Electrical Engineering, Shanghai 200240, China.
| | - Yiqing Shen
- Department of Computer Science, Johns Hopkins University, Baltimore, MD 21218, USA.
| | - Yizhou Lu
- Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, China
| | - Yi Guo
- School of Computer, Data and Mathematical Sciences, Western Sydney University, Penrith, NSW 2751, Australia
| | - Dinggang Shen
- School of Biomedical Engineering, ShanghaiTech University, Shanghai 201210, China; Shanghai United Imaging Intelligence Co., Ltd., Shanghai 200230, China; Shanghai Clinical Research and Trial Center, Shanghai, 201210, China
| |
Collapse
|
50
|
Zhao F, Liu M, Gao Z, Jiang X, Wang R, Zhang L. Dual-scale similarity-guided cycle generative adversarial network for unsupervised low-dose CT denoising. Comput Biol Med 2023; 161:107029. [PMID: 37230021 DOI: 10.1016/j.compbiomed.2023.107029] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/20/2023] [Revised: 04/10/2023] [Accepted: 05/09/2023] [Indexed: 05/27/2023]
Abstract
Removing the noise in low-dose CT (LDCT) is crucial to improving the diagnostic quality. Previously, many supervised or unsupervised deep learning-based LDCT denoising algorithms have been proposed. Unsupervised LDCT denoising algorithms are more practical than supervised ones since they do not need paired samples. However, unsupervised LDCT denoising algorithms are rarely used clinically due to their unsatisfactory denoising ability. In unsupervised LDCT denoising, the lack of paired samples makes the direction of gradient descent full of uncertainty. On the contrary, paired samples used in supervised denoising allow the parameters of networks to have a clear direction of gradient descent. To bridge the gap in performance between unsupervised and supervised LDCT denoising, we propose dual-scale similarity-guided cycle generative adversarial network (DSC-GAN). DSC-GAN uses similarity-based pseudo-pairing to better accomplish unsupervised LDCT denoising. We design a Vision Transformer-based global similarity descriptor and a residual neural network-based local similarity descriptor for DSC-GAN to effectively describe the similarity between two samples. During training, pseudo-pairs, i.e., similar LDCT samples and normal-dose CT (NDCT) samples, dominate parameter updates. Thus, the training can achieve equivalent effect as training with paired samples. Experiments on two datasets demonstrate that DSC-GAN beats the state-of-the-art unsupervised algorithms and reaches a level close to supervised LDCT denoising algorithms.
Collapse
Affiliation(s)
- Feixiang Zhao
- College of Nuclear Technology and Automation Engineering, Chengdu University of Technology, Chengdu, 610000, China.
| | - Mingzhe Liu
- College of Nuclear Technology and Automation Engineering, Chengdu University of Technology, Chengdu, 610000, China; School of Data Science and Artificial Intelligence, Wenzhou University of Technology, Wenzhou, 325000, China.
| | - Zhihong Gao
- Department of Big Data in Health Science, First Affiliated Hospital of Wenzhou Medical University, Wenzhou, 325000, China.
| | - Xin Jiang
- School of Data Science and Artificial Intelligence, Wenzhou University of Technology, Wenzhou, 325000, China.
| | - Ruili Wang
- School of Mathematical and Computational Science, Massey University, Auckland, 0632, New Zealand.
| | - Lejun Zhang
- Cyberspace Institute of Advanced Technology, Guangzhou University, Guangzhou, 510006, China; College of Information Engineering, Yangzhou University, Yangzhou, 225127, China.
| |
Collapse
|