Article
Multidisciplinary Sciences
Rafsanjany Kushol, Pedram Parnianpour, Alan H. Wilman, Sanjay Kalra, Yee-Hong Yang
Summary: This study examines the impact of different scanner manufacturers on the performance of deep learning models for disease classification using multi-center MRI data. The results demonstrate a significant decline in classification performance when models trained on one type of scanner are tested with data from different manufacturers. Applying harmonization techniques does not improve the classification performance of the models.
SCIENTIFIC REPORTS
(2023)
Article
Neurosciences
Mahbaneh Eshaghzadeh Torbati, Davneet S. Minhas, Ghasan Ahmad, Erin E. O'Connor, John Muschelli, Charles M. Laymon, Zixi Yang, Ann D. Cohen, Howard J. Aizenstein, William E. Klunk, Bradley T. Christian, Seong Jae Hwang, Ciprian M. Crainiceanu, Dana L. Tudorascu
Summary: This study evaluated the impact of different data analysis methods on data collected from multiple scanners and experimental conditions. The results showed that RAVEL substantially improved the reproducibility of image intensities, while ComBat was preferred for regional level harmonization. RAVEL and ComBat also significantly reduced bias compared to analysis of RAW images.
Article
Multidisciplinary Sciences
Colin Hawco, Erin W. Dickie, Gabrielle Herman, Jessica A. Turner, Miklos Argyelan, Anil K. Malhotra, Robert W. Buchanan, Aristotle N. Voineskos
Summary: Human neuroimaging research has generated a significant amount of knowledge about brain function in both healthy and clinical populations. Understanding the limitations of small sample studies has led to an increase in multi-site, multi-scanner protocols to further explore the human brain. In a specific multi-site project focused on social cognition in schizophrenia, a group of travelling human phantoms underwent annual structural T1, diffusion, and resting-state functional MRIs at three different sites. The scan protocols were carefully harmonized to maintain consistency across sites. Despite scanner upgrades and participant changes, a total of 30 MRI scans across 4 individuals, 6 MRIs, and 4 years were obtained. This dataset offers multiple neuroimaging modalities and repeated scans, enabling the evaluation of scanner differences, pipeline consistency, and multi-scanner harmonization approaches.
Article
Neurosciences
Robel K. Gebre, Matthew L. Senjem, Sheelakumari Raghavan, Christopher G. Schwarz, Jeffery L. Gunter, Ekaterina I. Hofrenning, Robert I. Reid, Kejal Kantarci, Jonathan Graff-Radford, David S. Knopman, Ronald C. Petersen, Clifford R. Jack Jr, Prashanthi Vemuri
Summary: The clinical usefulness of MRI biomarkers for aging and dementia studies relies on accurate brain morphological measurements. However, variations in scanners and protocols may introduce noise or bias. This study evaluates different methods, including deep learning and statistical methods, for post-acquisition scan harmonization to address this issue.
Article
Radiology, Nuclear Medicine & Medical Imaging
Simon Lennartz, Anushri Parakh, Jinjin Cao, David Zopfs, Nils Grosse Hokamp, Avinash Kambadakone
Summary: The study investigates the inter-scan and inter-scanner variation of iodine concentration and attenuation in different DECT scanners, finding that the iodine concentration shows higher variation compared to attenuation values. This variation depends on the scanner pairs and organs assessed, highlighting the importance of acknowledging these differences in clinical and scientific DECT applications.
EUROPEAN RADIOLOGY
(2021)
Article
Neurosciences
Andrew A. Chen, Dhivya Srinivasan, Raymond Pomponio, Yong Fan, Ilya M. Nasrallah, Susan M. Resnick, Lori L. Beason-Held, Christos Davatzikos, Theodore D. Satterthwaite, Dani S. Bassett, Russell T. Shinohara, Haochang Shou
Summary: Community detection on graphs constructed from functional magnetic resonance imaging (fMRI) data has provided important insights into brain functional organization. However, differences in scanners can introduce variability into the results, known as scanner effects. In this study, new methodology for harmonizing functional connectivity is proposed to reduce these effects.
Article
Neurosciences
Kurt G. Schilling, Chantal M. W. Tax, Francois Rheault, Colin Hansen, Qi Yang, Fang-Cheng Yeh, Leon Cai, Adam W. Anderson, Bennett A. Landman
Summary: This study investigates the impact of various factors on diffusion tractography bundle segmentation, including scan repeats, scanners, vendors, acquisition resolution, diffusion schemes, and diffusion sensitization. The acquisition protocol, particularly resolution, has the largest effect on reproducibility and features variation, followed by vendor and scanner effects. Different segmentation workflows show varying robustness to sources of variation. The choice of bundle segmentation workflows has a bigger impact on the results than other confounding factors.
Article
Radiology, Nuclear Medicine & Medical Imaging
Kenneth Wengler, Clifford Cassidy, Marieke van Der Pluijm, Jodi J. Weinstein, Anissa Abi-Dargham, Elsmarieke van de Giessen, Guillermo Horga
Summary: The study aimed to develop a method to harmonize NM-MRI data across different scanners and sites, using support vector machine and support vector regression to assess the impact of harmonization on biological variability. After harmonization, the classifier accuracy for scanners dropped to chance level, but there was no significant impact on age prediction results.
JOURNAL OF MAGNETIC RESONANCE IMAGING
(2021)
Article
Food Science & Technology
Niels Demaitre, Koen De Reu, Ellen Francois, Lieven De Zutter, Geertrui Rasschaert, Annemie Geeraerd
Summary: The purpose of this study was to assess the growth potential of Listeria monocytogenes in raw pork products. Pork chops showed limited growth potential, while minced pork exhibited higher growth potential. However, there were variations within and between batches, leading to potential underestimations and posing a risk to public health.
INTERNATIONAL JOURNAL OF FOOD MICROBIOLOGY
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Shilun Zhao, Tianhao Zhang, Wei Zhang, Tingting Pan, Ge Zhang, Shuang Feng, Xiwan Zhang, Binbin Nie, Hua Liu, Baoci Shan
Summary: The study aimed to develop a harmonization method to remove scanner variability and assess the consistency of results in multicenter studies. The results showed that HCOBE effectively reduced scanner variability and improved the consistency of multicenter study results, compared to reference data.
JOURNAL OF MAGNETIC RESONANCE IMAGING
(2023)
Review
Health Care Sciences & Services
Shruti Atul Mali, Abdalla Ibrahim, Henry C. Woodruff, Vincent Andrearczyk, Henning Mueller, Sergey Primakov, Zohaib Salahuddin, Avishek Chatterjee, Philippe Lambin
Summary: Radiomics converts medical images into mineable data for clinical decision support. Various investigations have assessed the reproducibility and validation of radiomic features, leading to different harmonization solutions in image and feature domains to address variability across scanners and protocol settings. Deep learning solutions, such as GANs and NST techniques, are highlighted for multi-centric radiomic studies.
JOURNAL OF PERSONALIZED MEDICINE
(2021)
Article
Neurosciences
Ryo Kurokawa, Kouhei Kamiya, Shinsuke Koike, Moto Nakaya, Akiko Uematsu, Saori C. Tanaka, Koji Kamagata, Naohiro Okada, Kentaro Morita, Kiyoto Kasai, Osamu Abe
Summary: Characterization of brain networks using diffusion MRI has seen rapid advancements, especially in the context of data sharing and multi-center studies. This study highlights the need for harmonization methods to correct bias caused by scanner differences in structural network analyses. The study demonstrates lower reliability of edge weights and graph theory metrics in data from different scanners compared to scan-rescan data from the same scanner, indicating potential systematic differences between scanners and the risk of bias in direct comparison of networks from different scanners. By modeling scanner effects at the level of network matrices using traveling-subject data, it is possible to reduce inter-scanner variabilities while preserving inter-subject differences among healthy individuals.
Article
Neurosciences
Fabian Preisner, Rouven Behnisch, Veronique Schwehr, Tim Godel, Daniel Schwarz, Olivia Foesleitner, Philipp Baumer, Sabine Heiland, Martin Bendszus, Moritz Kronlage
Summary: This study evaluated the impact of different MR scanners on MR-neurography (MRN) parameters of the sciatic nerve. The results showed good agreement between different scanners for T2 and quantitative diffusion tensor imaging (DTI) parameters, but moderate agreement for proton spin density (PD). Significant inter-scanner differences were observed for several parameters. The calculated standard error of measurement values were mostly within one standard deviation of the absolute mean values.
FRONTIERS IN NEUROSCIENCE
(2022)
Article
Oncology
Yingping Li, Samy Ammari, Corinne Balleyguier, Nathalie Lassau, Emilie Chouzenoux
Summary: This study investigates how image preprocessing methods and harmonization methods can help remove scanner effects and improve the reproducibility of radiomic features in brain MRI studies. The ComBat method is found to be essential in removing scanner effects, while intensity normalization methods improve the robustness of the harmonized features.
Review
Clinical Neurology
Ashley D. Harris, Houshang Amiri, Mariana Bento, Ronald Cohen, Christopher R. K. Ching, Christina Cudalbu, Emily L. Dennis, Arne Doose, Stefan Ehrlich, Ivan I. Kirov, Ralf Mekle, Georg Oeltzschner, Eric Porges, Roberto Souza, Friederike I. Tam, Brian Taylor, Paul M. Thompson, Yann Quide, Elisabeth A. Wilde, John Williamson, Alexander P. Lin, Brenda Bartnik-Olson
Summary: Magnetic resonance spectroscopy (MRS) is a powerful and non-invasive imaging technique that quantitatively measures brain metabolites and has utility in diagnosing and characterizing neurological diseases. However, its impact has been limited by small sample sizes, methodological variability, and intrinsic limitations. This manuscript provides an overview of MRS data harmonization, including key considerations for retrospective and prospective studies, and various approaches to harmonization. The goal is to provide knowledge for conducting MRS data harmonization studies.
FRONTIERS IN NEUROLOGY
(2023)
Article
Neuroimaging
Inga K. Koerte, Carrie Esopenko, Sidney R. Hinds, Martha E. Shenton, Elena M. Bonke, Jeffrey J. Bazarian, Kevin C. Bickart, Erin D. Bigler, Sylvain Bouix, Thomas A. Buckley, Meeryo C. Choe, Paul S. Echlin, Jessica Gill, Christopher C. Giza, Jasmeet Hayes, Cooper B. Hodges, Andrei Irimia, Paula K. Johnson, Kimbra Kenney, Harvey S. Levin, Alexander P. Lin, Hannah M. Lindsey, Michael L. Lipton, Jeffrey E. Max, Andrew R. Mayer, Timothy B. Meier, Kian Merchant-Borna, Tricia L. Merkley, Brian D. Mills, Mary R. Newsome, Tara Porfido, Jaclyn A. Stephens, Maria Carmela Tartaglia, Ashley L. Ware, Ross D. Zafonte, Michael M. Zeineh, Paul M. Thompson, David F. Tate, Emily L. Dennis, Elisabeth A. Wilde, David Baron
Summary: Research on sport-related brain injury is often limited by small sample sizes and variation in neuroimaging techniques and analysis tools. The ENIGMA Sports Injury working group aims to address these limitations by promoting data sharing and collaborative data analysis, with a focus on reproducibility and enhancing data quality for future prospective studies.
BRAIN IMAGING AND BEHAVIOR
(2021)
Article
Neuroimaging
R. Jarrett Rushmore, Sylvain Bouix, Marek Kubicki, Yogesh Rathi, Douglas L. Rosene, Edward H. Yeterian, Nikos Makris
Summary: Research on the rhesus monkey brain has provided insights into primate brain function and structure, offering valuable references for studying the human brain. The macaque Harvard-Oxford Atlas (mHOA) is a novel parcellation system based on the human Harvard-Oxford Atlas (HOA), utilizing an ontological and taxonomic framework to categorize anatomical features. This system of parcellation provides a foundation for interpreting results from various experimental studies, such as functional magnetic resonance imaging (fMRI), physiology, connectivity, and graph theory.
BRAIN IMAGING AND BEHAVIOR
(2021)
Article
Psychiatry
Elisabetta C. del Re, William S. Stone, Sylvain Bouix, Johanna Seitz, Victor Zeng, Anthony Guliano, Nathaniel Somes, Tianhong Zhang, Benjamin Reid, Amanda Lyall, Monica Lyons, Huijun Li, Susan Whitfield-Gabrieli, Matcheri Keshavan, Larry J. Seidman, Robert W. McCarley, Jijun Wang, Yingying Tang, Martha E. Shenton, Margaret A. Niznikiewicz
Summary: The study found that cortical thickness was significantly reduced in CHR compared to the control group, and cortical thickness abnormalities in converters relative to CHR-NC and HC were found in core auditory and language processing regions.
SCHIZOPHRENIA BULLETIN
(2021)
Article
Neurosciences
David Kaufmann, Nico Sollmann, Elisabeth Kaufmann, Rosanna Veggeberg, Yorghos Tripodis, Pawel P. Wrobel, Janna Kochsiek, Brett M. Martin, Alexander P. Lin, Michael J. Coleman, Michael L. Alosco, Ofer Pasternak, Sylvain Bouix, Robert A. Stern, Martha E. Shenton, Inga K. Koerte
Summary: The study found that a younger age at first exposure to repetitive head impacts is associated with decreased cortical thickness in the brains of American football players, which in turn is linked to worse neuropsychological performance.
Article
Psychiatry
Martin Jani, Zora Kikinis, Jan Losak, Ofer Pasternak, Filip Szczepankiewicz, Carina Heller, Sophia Swago, Annelise Silva, Sylvain Bouix, Marek Kubicki, Libor Ustohal, Petr Kudlicka, Lubomir Vojtisek, Carl-Frederik Westin, Tomas Kasparek
Summary: Individuals with schizophrenia showed decreased emotional awareness compared to healthy controls, but there were no statistically significant between-group differences in gray matter volumes of the emotional awareness network. Performance on LEAS Other correlated negatively with right precuneus gray matter volume only in the schizophrenia group.
FRONTIERS IN PSYCHIATRY
(2021)
Article
Neurosciences
Doron Elad, Suheyla Cetin-Karayumak, Fan Zhang, Kang Ik K. Cho, Amanda E. Lyall, Johanna Seitz-Holland, Rami Ben-Ari, Godfrey D. Pearlson, Carol A. Tamminga, John A. Sweeney, Brett A. Clementz, David J. Schretlen, Petra Verena Viher, Katharina Stegmayer, Sebastian Walther, Jungsun Lee, Tim J. Crow, Anthony James, Aristotle N. Voineskos, Robert W. Buchanan, Philip R. Szeszko, Anil K. Malhotra, Matcheri S. Keshavan, Martha E. Shenton, Yogesh Rathi, Sylvain Bouix, Nir Sochen, Marek R. Kubicki, Ofer Pasternak
Summary: Diffusion MRI studies have shown differences in white matter between individuals with schizophrenia and healthy controls, but these abnormalities are often not observable at the individual level. Normative modeling analysis, which identifies extreme deviations in affected brain locations based on a normative range, has limitations in predictive accuracy. Combining information from multiple white matter tracts and imaging measures can improve prediction performance for individual-level classification in schizophrenia.
HUMAN BRAIN MAPPING
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Janna Kochsiek, Lauren J. O'Donnell, Fan Zhang, Elena M. Bonke, Nico Sollmann, Yorghos Tripodis, Tim L. T. Wiegand, David Kaufmann, Lisa Umminger, Maria A. Di Biase, Elisabeth Kaufmann, Vivian Schultz, Michael L. Alosco, Brett M. Martin, Alexander P. Lin, Michael J. Coleman, Yogesh Rathi, Ofer Pasternak, Sylvain Bouix, Robert A. Stern, Martha E. Shenton, Inga K. Koerte
Summary: This retrospective cohort study found an association between repetitive head impacts and later-life corpus callosum microstructure, plasma total tau, and clinical functioning in former professional American football players.
JOURNAL OF MAGNETIC RESONANCE IMAGING
(2021)
Article
Psychiatry
Carina Heller, Thomas Weiss, Elisabetta C. del Re, Sophia Swago, Ioana L. Coman, Kevin M. Antshel, Wanda Fremont, Sylvain Bouix, Wendy R. Kates, Marek R. Kubicki, Zora Kikinis
Summary: The study found that lateral ventricular volumes are negatively correlated with subcortical volumes in individuals with 22q11DS, while in patients with both PS, Global Assessment of Functioning is positively correlated with lateral ventricular volumes and negatively correlated with subcortical volumes.
PSYCHIATRY RESEARCH
(2021)
Article
Clinical Neurology
Michael L. Alosco, Megan L. Mariani, Charles H. Adler, Laura J. Balcer, Charles Bernick, Rhoda Au, Sarah J. Banks, William B. Barr, Sylvain Bouix, Robert C. Cantu, Michael J. Coleman, David W. Dodick, Lindsay A. Farrer, Yonas E. Geda, Douglas I. Katz, Inga K. Koerte, Neil W. Kowall, Alexander P. Lin, Daniel S. Marcus, Kenneth L. Marek, Michael D. McClean, Ann C. McKee, Jesse Mez, Joseph N. Palmisano, Elaine R. Peskind, Yorghos Tripodis, Robert W. Turner, Jennifer V. Wethe, Jeffrey L. Cummings, Eric M. Reiman, Martha E. Shenton, Robert A. Stern
Summary: The DIAGNOSE CTE Research Project aims to develop biomarkers for CTE, characterize its clinical presentation, validate diagnostic criteria, study risk factors, and share resources. With 240 male participants enrolled and baseline evaluations completed, the study is expected to contribute significantly to CTE research.
ALZHEIMERS RESEARCH & THERAPY
(2021)
Article
Neurosciences
Eimear M. Foley, Yorghos Tripodis, Eukyung Yhang, Inga K. Koerte, Brett M. Martin, Joseph Palmisano, Nikos Makris, Vivian Schultz, Chris Lepage, Marc Muehlmann, Pawel P. Wrobel, Jeffrey P. Guenette, Robert C. Cantu, Alexander P. Lin, Michael Coleman, Jesse Mez, Sylvain Bouix, Martha E. Shenton, Robert A. Stern, Michael L. Alosco
Summary: This study examines the association between reserve metrics and cognitive and neuropsychiatric functioning in former professional football players. The study finds that reading ability is associated with attention, episodic memory, fluency, and behavioral regulation. The residual variance in episodic memory and executive functioning is also associated with cognitive performance. Traditional reserve proxies may not be suitable for elite athlete samples, and alternative approaches are needed to quantify reserve.
JOURNAL OF ALZHEIMERS DISEASE
(2022)
Article
Clinical Neurology
Madeline Uretsky, Sylvain Bouix, Ronald J. Killiany, Yorghos Tripodis, Brett Martin, Joseph Palmisano, Asim Z. Mian, Karen Buch, Chad Farris, Daniel H. Daneshvar, Brigid Dwyer, Lee Goldstein, Douglas Katz, Christopher Nowinski, Robert Cantu, Neil Kowall, Bertrand Russell Huber, Robert A. Stern, Victor E. Alvarez, Thor D. Stein, Ann McKee, Jesse Mez, Michael L. Alosco
Summary: This study investigated the neuropathologic correlates of white matter hyperintensities (WMH) in brain donors exposed to repetitive head impacts. The results suggest that WMH may capture long-term white matter pathologies from repetitive head impacts, including white matter rarefaction and p-tau accumulation, in addition to microvascular disease. Prospective imaging-pathologic correlation studies are needed to further understand these associations.
Article
Psychiatry
Johanna Seitz-Holland, Monica Lyons, Leila Kushan, Amy Lin, Julio E. Villalon-Reina, Kang Ik Kevin Cho, Fan Zhang, Tashrif Billah, Sylvain Bouix, Marek Kubicki, Carrie E. Bearden, Ofer Pasternak
Summary: Deletions and duplications at the 22q11.2 locus show opposing effects on white matter microstructure, with deletions possibly linked to enlarged CSF space and denser WM, while duplications exhibit evidence of demyelination similar to neuropsychiatric disorders.
TRANSLATIONAL PSYCHIATRY
(2021)
Article
Medicine, General & Internal
Philine Rojczyk, Johanna Seitz-Holland, Elisabeth Kaufmann, Valerie J. Sydnor, Cara L. Kim, Lisa F. Umminger, Tim L. T. Wiegand, Jeffrey P. Guenette, Fan Zhang, Yogesh Rathi, Sylvain Bouix, Ofer Pasternak, Catherine B. Fortier, David Salat, Sidney R. Hinds, Florian Heinen, Lauren J. O'Donnell, William P. Milberg, Regina E. McGlinchey, Martha E. Shenton, Inga K. Koerte
Summary: Sleep disturbances are strongly associated with mild traumatic brain injury (mTBI) and post-traumatic stress disorder (PTSD), and poor sleep quality has a compounding effect on white matter (WM) microstructure in veterans with comorbid PTSD+mTBI. Veterans with PTSD and comorbid PTSD+mTBI reported poorer sleep quality than those with mTBI or no history of PTSD or mTBI. Poor sleep quality fully mediated the association between greater PTSD symptom severity and impaired WM microstructure. Sleep-targeted interventions are necessary for improving brain health in veterans with PTSD+mTBI.
JOURNAL OF CLINICAL MEDICINE
(2023)
Article
Medicine, General & Internal
Lara Pankatz, Philine Rojczyk, Johanna Seitz-Holland, Sylvain Bouix, Leonard B. Jung, Tim L. T. Wiegand, Elena M. Bonke, Nico Sollmann, Elisabeth Kaufmann, Holly Carrington, Twishi Puri, Yogesh Rathi, Michael J. Coleman, Ofer Pasternak, Mark S. George, Thomas W. Mcallister, Ross Zafonte, Murray B. Stein, Christine E. Marx, Martha E. Shenton, Inga K. Koerte
Summary: This study analyzed diffusion and structural MRI data of 278 participants with and without military background, and found microstructural alterations at the gray matter/white matter boundary of the brain after mild traumatic brain injury (mTBI) that were associated with post-concussive symptom severity, functional, and cognitive impairment. These findings suggest that microstructural changes at the gray matter/white matter boundary may be sensitive markers of adverse long-term outcomes following mTBI.
JOURNAL OF CLINICAL MEDICINE
(2023)
Article
Neurosciences
Maria Geisler, Elizabeth Rizzoni, Nikolaos Makris, Ofer Pasternak, Yogesh Rathi, Sylvain Bouix, Marco Herbsleb, Karl-Jurgen Bar, Thomas Weiss, Zora Kikinis
Summary: This study found that variation in the microstructure of the medial forebrain bundle (MFB) is associated with interindividual pain sensitivity. Specifically, lower ratings of interindividual pain intensity were correlated with higher FA(t) and lower RDt of the MFB.
HUMAN BRAIN MAPPING
(2021)