1
|
Hu F, Chen AA, Horng H, Bashyam V, Davatzikos C, Alexander-Bloch A, Li M, Shou H, Satterthwaite TD, Yu M, Shinohara RT. Image harmonization: A review of statistical and deep learning methods for removing batch effects and evaluation metrics for effective harmonization. Neuroimage 2023; 274:120125. [PMID: 37084926 PMCID: PMC10257347 DOI: 10.1016/j.neuroimage.2023.120125] [Citation(s) in RCA: 18] [Impact Index Per Article: 18.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/14/2023] [Revised: 04/12/2023] [Accepted: 04/19/2023] [Indexed: 04/23/2023] Open
Abstract
Magnetic resonance imaging and computed tomography from multiple batches (e.g. sites, scanners, datasets, etc.) are increasingly used alongside complex downstream analyses to obtain new insights into the human brain. However, significant confounding due to batch-related technical variation, called batch effects, is present in this data; direct application of downstream analyses to the data may lead to biased results. Image harmonization methods seek to remove these batch effects and enable increased generalizability and reproducibility of downstream results. In this review, we describe and categorize current approaches in statistical and deep learning harmonization methods. We also describe current evaluation metrics used to assess harmonization methods and provide a standardized framework to evaluate newly-proposed methods for effective harmonization and preservation of biological information. Finally, we provide recommendations to end-users to advocate for more effective use of current methods and to methodologists to direct future efforts and accelerate development of the field.
Collapse
Affiliation(s)
- Fengling Hu
- Penn Statistics in Imaging and Visualization Endeavor (PennSIVE), Department of Biostatistics, Epidemiology, and Informatics, Perelman School of Medicine, University of Pennsylvania, 423 Guardian Dr, Philadelphia, PA 19104, United States.
| | - Andrew A Chen
- Penn Statistics in Imaging and Visualization Endeavor (PennSIVE), Department of Biostatistics, Epidemiology, and Informatics, Perelman School of Medicine, University of Pennsylvania, 423 Guardian Dr, Philadelphia, PA 19104, United States
| | - Hannah Horng
- Penn Statistics in Imaging and Visualization Endeavor (PennSIVE), Department of Biostatistics, Epidemiology, and Informatics, Perelman School of Medicine, University of Pennsylvania, 423 Guardian Dr, Philadelphia, PA 19104, United States
| | - Vishnu Bashyam
- Center for Biomedical Image Computing and Analytics (CBICA), Perelman School of Medicine, United States
| | - Christos Davatzikos
- Center for Biomedical Image Computing and Analytics (CBICA), Perelman School of Medicine, United States
| | - Aaron Alexander-Bloch
- Department of Psychiatry, Perelman School of Medicine, University of Pennsylvania, United States; Penn-CHOP Lifespan Brain Institute, United States; Department of Child and Adolescent Psychiatry and Behavioral Science, Children's Hospital of Philadelphia, United States
| | - Mingyao Li
- Statistical Center for Single-Cell and Spatial Genomics, Perelman School of Medicine, University of Pennsylvania, United States
| | - Haochang Shou
- Penn Statistics in Imaging and Visualization Endeavor (PennSIVE), Department of Biostatistics, Epidemiology, and Informatics, Perelman School of Medicine, University of Pennsylvania, 423 Guardian Dr, Philadelphia, PA 19104, United States; Center for Biomedical Image Computing and Analytics (CBICA), Perelman School of Medicine, United States
| | - Theodore D Satterthwaite
- Department of Psychiatry, Perelman School of Medicine, University of Pennsylvania, United States; Penn-CHOP Lifespan Brain Institute, United States; The Penn Lifespan Informatics and Neuroimaging Center, Department of Psychiatry, Perelman School of Medicine, University of Pennsylvania, United States
| | - Meichen Yu
- Indiana Alzheimer's Disease Research Center, Indiana University School of Medicine, United States
| | - Russell T Shinohara
- Penn Statistics in Imaging and Visualization Endeavor (PennSIVE), Department of Biostatistics, Epidemiology, and Informatics, Perelman School of Medicine, University of Pennsylvania, 423 Guardian Dr, Philadelphia, PA 19104, United States; Center for Biomedical Image Computing and Analytics (CBICA), Perelman School of Medicine, United States
| |
Collapse
|
2
|
Hanson JL, Adkins DJ, Nacewicz BM, Barry KR. Impact of Socioeconomic Status on Amygdala and Hippocampus Subdivisions in Children and Adolescents. BIORXIV : THE PREPRINT SERVER FOR BIOLOGY 2023:2023.03.10.532071. [PMID: 36993362 PMCID: PMC10054998 DOI: 10.1101/2023.03.10.532071] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Indexed: 06/19/2023]
Abstract
Socioeconomic status (SES) in childhood can impact behavioral and brain development. Past work has consistently focused on the amygdala and hippocampus, two brain areas critical for emotion and behavioral responding. While there are SES differences in amygdala and hippocampal volumes, there are many unanswered questions in this domain connected to neurobiological specificity, and for whom these effects may be more pronounced. We may be able to investigate some anatomical subdivisions of these brain areas, as well as if relations with SES vary by participant age and sex. No work to date has however completed these types of analyses. To overcome these limitations, here, we combined multiple, large neuroimaging datasets of children and adolescents with information about neurobiology and SES (N=2,765). We examined subdivisions of the amygdala and hippocampus and found multiple amygdala subdivisions, as well as the head of the hippocampus, were related to SES. Greater volumes in these areas were seen for higher-SES youth participants. Looking at age- and sex-specific subgroups, we tended to see stronger effects in older participants, for both boys and girls. Paralleling effects for the full sample, we see significant positive associations between SES and volumes for the accessory basal amygdala and head of the hippocampus. We more consistently found associations between SES and volumes of the hippocampus and amygdala in boys (compared to girls). We discuss these results in relation to conceptions of "sex-as-a-biological variable" and broad patterns of neurodevelopment across childhood and adolescence. These results fill in important gaps on the impact of SES on neurobiology critical for emotion, memory, and learning.
Collapse
|
3
|
Volumetric MRI Findings in Mild Traumatic Brain Injury (mTBI) and Neuropsychological Outcome. Neuropsychol Rev 2023; 33:5-41. [PMID: 33656702 DOI: 10.1007/s11065-020-09474-0] [Citation(s) in RCA: 15] [Impact Index Per Article: 15.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/17/2019] [Accepted: 12/20/2020] [Indexed: 10/22/2022]
Abstract
Region of interest (ROI) volumetric assessment has become a standard technique in quantitative neuroimaging. ROI volume is thought to represent a coarse proxy for making inferences about the structural integrity of a brain region when compared to normative values representative of a healthy sample, adjusted for age and various demographic factors. This review focuses on structural volumetric analyses that have been performed in the study of neuropathological effects from mild traumatic brain injury (mTBI) in relation to neuropsychological outcome. From a ROI perspective, the probable candidate structures that are most likely affected in mTBI represent the target regions covered in this review. These include the corpus callosum, cingulate, thalamus, pituitary-hypothalamic area, basal ganglia, amygdala, and hippocampus and associated structures including the fornix and mammillary bodies, as well as whole brain and cerebral cortex along with the cerebellum. Ventricular volumetrics are also reviewed as an indirect assessment of parenchymal change in response to injury. This review demonstrates the potential role and limitations of examining structural changes in the ROIs mentioned above in relation to neuropsychological outcome. There is also discussion and review of the role that post-traumatic stress disorder (PTSD) may play in structural outcome in mTBI. As emphasized in the conclusions, structural volumetric findings in mTBI are likely just a single facet of what should be a multimodality approach to image analysis in mTBI, with an emphasis on how the injury damages or disrupts neural network integrity. The review provides an historical context to quantitative neuroimaging in neuropsychology along with commentary about future directions for volumetric neuroimaging research in mTBI.
Collapse
|
4
|
Ali H, Biswas R, Ali F, Shah U, Alamgir A, Mousa O, Shah Z. The role of generative adversarial networks in brain MRI: a scoping review. Insights Imaging 2022; 13:98. [PMID: 35662369 PMCID: PMC9167371 DOI: 10.1186/s13244-022-01237-0] [Citation(s) in RCA: 8] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/13/2022] [Accepted: 05/11/2022] [Indexed: 11/23/2022] Open
Abstract
The performance of artificial intelligence (AI) for brain MRI can improve if enough data are made available. Generative adversarial networks (GANs) showed a lot of potential to generate synthetic MRI data that can capture the distribution of real MRI. Besides, GANs are also popular for segmentation, noise removal, and super-resolution of brain MRI images. This scoping review aims to explore how GANs methods are being used on brain MRI data, as reported in the literature. The review describes the different applications of GANs for brain MRI, presents the most commonly used GANs architectures, and summarizes the publicly available brain MRI datasets for advancing the research and development of GANs-based approaches. This review followed the guidelines of PRISMA-ScR to perform the study search and selection. The search was conducted on five popular scientific databases. The screening and selection of studies were performed by two independent reviewers, followed by validation by a third reviewer. Finally, the data were synthesized using a narrative approach. This review included 139 studies out of 789 search results. The most common use case of GANs was the synthesis of brain MRI images for data augmentation. GANs were also used to segment brain tumors and translate healthy images to diseased images or CT to MRI and vice versa. The included studies showed that GANs could enhance the performance of AI methods used on brain MRI imaging data. However, more efforts are needed to transform the GANs-based methods in clinical applications.
Collapse
Affiliation(s)
- Hazrat Ali
- College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, 34110, Doha, Qatar.
| | - Rafiul Biswas
- College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, 34110, Doha, Qatar
| | - Farida Ali
- College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, 34110, Doha, Qatar
| | - Uzair Shah
- College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, 34110, Doha, Qatar
| | - Asma Alamgir
- College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, 34110, Doha, Qatar
| | - Osama Mousa
- College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, 34110, Doha, Qatar
| | - Zubair Shah
- College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, 34110, Doha, Qatar.
| |
Collapse
|
5
|
Generative Adversarial Networks in Brain Imaging: A Narrative Review. J Imaging 2022; 8:jimaging8040083. [PMID: 35448210 PMCID: PMC9028488 DOI: 10.3390/jimaging8040083] [Citation(s) in RCA: 4] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/21/2021] [Revised: 03/08/2022] [Accepted: 03/15/2022] [Indexed: 02/04/2023] Open
Abstract
Artificial intelligence (AI) is expected to have a major effect on radiology as it demonstrated remarkable progress in many clinical tasks, mostly regarding the detection, segmentation, classification, monitoring, and prediction of diseases. Generative Adversarial Networks have been proposed as one of the most exciting applications of deep learning in radiology. GANs are a new approach to deep learning that leverages adversarial learning to tackle a wide array of computer vision challenges. Brain radiology was one of the first fields where GANs found their application. In neuroradiology, indeed, GANs open unexplored scenarios, allowing new processes such as image-to-image and cross-modality synthesis, image reconstruction, image segmentation, image synthesis, data augmentation, disease progression models, and brain decoding. In this narrative review, we will provide an introduction to GANs in brain imaging, discussing the clinical potential of GANs, future clinical applications, as well as pitfalls that radiologists should be aware of.
Collapse
|
6
|
Hanson JL, Nacewicz BM. Amygdala Allostasis and Early Life Adversity: Considering Excitotoxicity and Inescapability in the Sequelae of Stress. Front Hum Neurosci 2021; 15:624705. [PMID: 34140882 PMCID: PMC8203824 DOI: 10.3389/fnhum.2021.624705] [Citation(s) in RCA: 9] [Impact Index Per Article: 3.0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/01/2020] [Accepted: 04/23/2021] [Indexed: 12/19/2022] Open
Abstract
Early life adversity (ELA), such as child maltreatment or child poverty, engenders problems with emotional and behavioral regulation. In the quest to understand the neurobiological sequelae and mechanisms of risk, the amygdala has been of major focus. While the basic functions of this region make it a strong candidate for understanding the multiple mental health issues common after ELA, extant literature is marked by profound inconsistencies, with reports of larger, smaller, and no differences in regional volumes of this area. We believe integrative models of stress neurodevelopment, grounded in "allostatic load," will help resolve inconsistencies in the impact of ELA on the amygdala. In this review, we attempt to connect past research studies to new findings with animal models of cellular and neurotransmitter mediators of stress buffering to extreme fear generalization onto testable research and clinical concepts. Drawing on the greater impact of inescapability over unpredictability in animal models, we propose a mechanism by which ELA aggravates an exhaustive cycle of amygdala expansion and subsequent toxic-metabolic damage. We connect this neurobiological sequela to psychosocial mal/adaptation after ELA, bridging to behavioral studies of attachment, emotion processing, and social functioning. Lastly, we conclude this review by proposing a multitude of future directions in preclinical work and studies of humans that suffered ELA.
Collapse
Affiliation(s)
- Jamie L. Hanson
- Department of Psychology, University of Pittsburgh, Pittsburgh, PA, United States
| | - Brendon M. Nacewicz
- Department of Psychiatry, University of Wisconsin-Madison, Madison, WI, United States
| |
Collapse
|