Search

Your search keyword '"Morris, John A"' showing total 16,161 results
Start Over Author "Morris, John A"

Refine your results

more »

more »

more »

more »

more »

more »
16,161 results on '"Morris, John A"'

1. Robust, fully-automated assessment of cerebral perivascular spaces and white matter lesions: a multicentre MRI longitudinal study of their evolution and association with risk of dementia and accelerated brain atrophy

Author

Barisano, Giuseppe, Iv, Michael, Choupan, Jeiran, Hayden-Gephart, Melanie, Weiner, Michael, Aisen, Paul, Petersen, Ronald, Jack, Clifford R, Jagust, William, Trojanowki, John Q, Toga, Arthur W, Beckett, Laurel, Green, Robert C, Saykin, Andrew J, Morris, John, Shaw, Leslie M, Liu, Enchi, Montine, Tom, Thomas, Ronald G, Donohue, Michael, Walter, Sarah, Gessert, Devon, Sather, Tamie, Jiminez, Gus, Harvey, Danielle, Bernstein, Matthew, Fox, Nick, Thompson, Paul, Schuff, Norbert, DeCarli, Charles, Borowski, Bret, Gunter, Jeff, Senjem, Matt, Vemuri, Prashanthi, Jones, David, Kantarci, Kejal, Ward, Chad, Koeppe, Robert A, Foster, Norm, Reiman, Eric M, Chen, Kewei, Mathis, Chet, Landau, Susan, Cairns, Nigel J, Householder, Erin, Reinwald, Lisa Taylor, Lee, Virginia, Korecka, Magdalena, Figurski, Michal, Crawford, Karen, Neu, Scott, Foroud, Tatiana M, Potkin, Steven, Shen, Li, Kelley, Faber, Kim, Sungeun, Nho, Kwangsik, Kachaturian, Zaven, Frank, Richard, Snyder, Peter J, Molchan, Susan, Kaye, Jeffrey, Quinn, Joseph, Lind, Betty, Carter, Raina, Dolen, Sara, Schneider, Lon S, Pawluczyk, Sonia, Beccera, Mauricio, Teodoro, Liberty, Spann, Bryan M, Brewer, James, Vanderswag, Helen, Fleisher, Adam, Heidebrink, Judith L, Lord, Joanne L, Mason, Sara S, Albers, Colleen S, Knopman, David, Johnson, Kris, Doody, Rachelle S, Meyer, Javier Villanueva, Chowdhury, Munir, Rountree, Susan, Dang, Mimi, Stern, Yaakov, Honig, Lawrence S, Bell, Karen L, Ances, Beau, Morris, John C, Carroll, Maria, Leon, Sue, Mintun, Mark A, Schneider, Stacy, Oliver, Angela, Marson, Daniel, and Griffith, Randall

Subjects
Health Services and Systems, Health Sciences, Brain Disorders, Aging, Alzheimer's Disease including Alzheimer's Disease Related Dementias (AD/ADRD), Acquired Cognitive Impairment, Alzheimer's Disease, Vascular Cognitive Impairment/Dementia, Biomedical Imaging, Neurosciences, Clinical Research, Neurodegenerative, Dementia, Prevention, Cerebrovascular, Clinical Trials and Supportive Activities, Bioengineering, Alzheimer's Disease Related Dementias (ADRD), 4.1 Discovery and preclinical testing of markers and technologies, 4.2 Evaluation of markers and technologies, Neurological, Humans, Female, Magnetic Resonance Imaging, Male, Atrophy, Aged, Longitudinal Studies, White Matter, Brain, Glymphatic System, Aged, 80 and over, Middle Aged, Reproducibility of Results, Algorithms, Alzheimer’s Disease Neuroimaging Initiative, Alzheimer’s disease, Glymphatic system, Perivascular spaces, Small vessel disease, White matter lesions, Clinical Sciences, Public Health and Health Services, Clinical sciences, Epidemiology
Abstract

BackgroundPerivascular spaces (PVS) on brain MRI are surrogates for small parenchymal blood vessels and their perivascular compartment, and may relate to brain health. However, it is unknown whether PVS can predict dementia risk and brain atrophy trajectories in participants without dementia, as longitudinal studies on PVS are scarce and current methods for PVS assessment lack robustness and inter-scanner reproducibility.MethodsWe developed a robust algorithm to automatically assess PVS count and size on clinical MRI, and investigated 1) their relationship with dementia risk and brain atrophy in participants without dementia, 2) their longitudinal evolution, and 3) their potential use as a screening tool in simulated clinical trials. We analysed 46,478 clinical measurements of cognitive functioning and 20,845 brain MRI scans from 10,004 participants (71.1 ± 9.7 years-old, 56.6% women) from three publicly available observational studies on ageing and dementia (the Alzheimer's Disease Neuroimaging Initiative, the National Alzheimer's Coordinating Centre database, and the Open Access Series of Imaging Studies). Clinical and MRI data collected between 2004 and 2022 were analysed with consistent methods, controlling for confounding factors, and combined using mixed-effects models.FindingsOur fully-automated method for PVS assessment showed excellent inter-scanner reproducibility (intraclass correlation coefficients >0.8). Fewer PVS and larger PVS diameter at baseline predicted higher dementia risk and accelerated brain atrophy. Longitudinal trajectories of PVS markers differed significantly in participants without dementia who converted to dementia compared with non-converters. In simulated placebo-controlled trials for treatments targeting cognitive decline, screening out participants at low risk of dementia based on our PVS markers enhanced the power of the trial independently of Alzheimer's disease biomarkers.InterpretationThese robust cerebrovascular markers predict dementia risk and brain atrophy and may improve risk-stratification of patients, potentially reducing cost and increasing throughput of clinical trials to combat dementia.FundingUS National Institutes of Health.

Published
2025

2. Overview of Alzheimers Disease Neuroimaging Initiative and future clinical trials.

Author

Weiner, Michael, Kanoria, Shaveta, Miller, Melanie, Aisen, Paul, Beckett, Laurel, Conti, Catherine, Diaz, Adam, Flenniken, Derek, Green, Robert, Harvey, Danielle, Jack, Clifford, Jagust, William, Lee, Edward, Morris, John, Nho, Kwangsik, Nosheny, Rachel, Okonkwo, Ozioma, Perrin, Richard, Petersen, Ronald, Rivera-Mindt, Monica, Saykin, Andrew, Shaw, Leslie, Toga, Arthur, Tosun, Duygu, and Veitch, Dallas

Subjects
Alzheimers Disease Neuroimaging Initiative, Alzheimers disease, Alzheimers disease biomarkers, Alzheimers disease clinical trials, Alzheimers disease progression, Laboratory of Neuro Imaging, amyloid, magnetic resonance imaging, neurodegeneration, positron emission tomography, post‐traumatic stress disorder, tau, underrepresented populations, Alzheimer Disease, Humans, Neuroimaging, Clinical Trials as Topic, Positron-Emission Tomography, Biomarkers
Abstract

The overall goal of the Alzheimers Disease Neuroimaging Initiative (ADNI) is to optimize and validate biomarkers for clinical trials while sharing all data and biofluid samples with the global scientific community. ADNI has been instrumental in standardizing and validating amyloid beta (Aβ) and tau positron emission tomography (PET) imaging. ADNI data were used for the US Food and Drug Administration (FDA) approval of the Fujirebio and Roche Elecsys cerebrospinal fluid diagnostic tests. Additionally, ADNI provided data for the trials of the FDA-approved treatments aducanumab, lecanemab, and donanemab. More than 6000 scientific papers have been published using ADNI data, reflecting ADNIs promotion of open science and data sharing. Despite its enormous success, ADNI has some limitations, particularly in generalizing its data and findings to the entire US/Canadian population. This introduction provides a historical overview of ADNI and highlights its significant accomplishments and future vision to pioneer the clinical trial of the future focusing on demographic inclusivity. HIGHLIGHTS: The Alzheimers Disease Neuroimaging Initiative (ADNI) introduced a novel model for public-private partnerships and data sharing. It successfully validated amyloid and Tau PET imaging, as well as CSF and plasma biomarkers, for diagnosing Alzheimers disease. ADNI generated and disseminated vital data for designing AD clinical trials.

Published
2025

3. DIRI: Adversarial Patient Reidentification with Large Language Models for Evaluating Clinical Text Anonymization

Author

Morris, John X., Campion, Thomas R., Nutheti, Sri Laasya, Peng, Yifan, Raj, Akhil, Zabih, Ramin, and Cole, Curtis L.

Subjects
Computer Science - Computation and Language
Abstract

Sharing protected health information (PHI) is critical for furthering biomedical research. Before data can be distributed, practitioners often perform deidentification to remove any PHI contained in the text. Contemporary deidentification methods are evaluated on highly saturated datasets (tools achieve near-perfect accuracy) which may not reflect the full variability or complexity of real-world clinical text and annotating them is resource intensive, which is a barrier to real-world applications. To address this gap, we developed an adversarial approach using a large language model (LLM) to re-identify the patient corresponding to a redacted clinical note and evaluated the performance with a novel De-Identification/Re-Identification (DIRI) method. Our method uses a large language model to reidentify the patient corresponding to a redacted clinical note. We demonstrate our method on medical data from Weill Cornell Medicine anonymized with three deidentification tools: rule-based Philter and two deep-learning-based models, BiLSTM-CRF and ClinicalBERT. Although ClinicalBERT was the most effective, masking all identified PII, our tool still reidentified 9% of clinical notes Our study highlights significant weaknesses in current deidentification technologies while providing a tool for iterative development and improvement.

Published
2024

4. Contextual Document Embeddings

Author

Morris, John X. and Rush, Alexander M.

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

Dense document embeddings are central to neural retrieval. The dominant paradigm is to train and construct embeddings by running encoders directly on individual documents. In this work, we argue that these embeddings, while effective, are implicitly out-of-context for targeted use cases of retrieval, and that a contextualized document embedding should take into account both the document and neighboring documents in context - analogous to contextualized word embeddings. We propose two complementary methods for contextualized document embeddings: first, an alternative contrastive learning objective that explicitly incorporates the document neighbors into the intra-batch contextual loss; second, a new contextual architecture that explicitly encodes neighbor document information into the encoded representation. Results show that both methods achieve better performance than biencoders in several settings, with differences especially pronounced out-of-domain. We achieve state-of-the-art results on the MTEB benchmark with no hard negative mining, score distillation, dataset-specific instructions, intra-GPU example-sharing, or extremely large batch sizes. Our method can be applied to improve performance on any contrastive learning dataset and any biencoder.

Published
2024

6. Functional Analysis of MS-Based Proteomics Data: From Protein Groups to Networks.

Author

Locard-Paulet, Marie, Doncheva, Nadezhda, Morris, John, and Jensen, Lars

Subjects
Bioinformatics, Biological databases, Cytoscape, Functional enrichment, Mass spectrometry, Networks, Protein groups, Proteomics, STRING, Proteomics, Mass Spectrometry, Protein Interaction Maps, Software, Humans, Databases, Protein
Abstract

Mass spectrometry-based proteomics allows the quantification of thousands of proteins, protein variants, and their modifications, in many biological samples. These are derived from the measurement of peptide relative quantities, and it is not always possible to distinguish proteins with similar sequences due to the absence of protein-specific peptides. In such cases, peptide signals are reported in protein groups that can correspond to several genes. Here, we show that multi-gene protein groups have a limited impact on GO-term enrichment, but selecting only one gene per group affects network analysis. We thus present the Cytoscape app Proteo Visualizer (https://apps.cytoscape.org/apps/ProteoVisualizer) that is designed for retrieving protein interaction networks from STRING using protein groups as input and thus allows visualization and network analysis of bottom-up MS-based proteomics data sets.

Published
2024

7. The Alzheimers Disease Neuroimaging Initiative Neuropathology Core: An update.

Author

Perrin, Richard, Franklin, Erin, Bernhardt, Haley, Burns, Aime, Schwetye, Katherine, Cairns, Nigel, Baxter, Michael, Weiner, Michael, and Morris, John

Subjects
Alzheimer, Alzheimers Disease Neuroimaging Initiative, Lewy body, TAR DNA binding protein 43, biomarker, neuropathology, Humans, Alzheimer Disease, Neuroimaging, Aged, Brain, Female, Biomarkers, Male, Aged, 80 and over, Neuropathology, Cohort Studies, Sensitivity and Specificity
Abstract

INTRODUCTION: Biomarkers for Alzheimers disease neuropathologic change (ADNC) have been instrumental in developing effective disease-modifying therapeutics. However, to prevent/treat dementia effectively, we require biomarkers for non-AD neuropathologies; for this, neuropathologic examinations and annotated tissue samples are essential. METHODS: We conducted clinicopathologic correlation for the first 100 Alzheimers Disease Neuroimaging Initiative (ADNI) Neuropathology Core (NPC) cases. RESULTS: Clinical syndromes in this cohort showed 95% sensitivity and 79% specificity for predicting high/intermediate ADNC, a 21% false positive rate, and a ∼44% false negative rate. In addition, 60% with high/intermediate ADNC harbored additional potentially dementing co-pathologies. DISCUSSION: These results suggest that clinical presentation imperfectly predicts ADNC and that accurate prediction of high/intermediate ADNC does not exclude co-pathology that may modify presentation, biomarkers, and therapeutic responses. Therefore, new biomarkers are needed for non-AD neuropathologies. The ADNI NPC supports this mission with well-characterized tissue samples (available through ADNI and the National Institute on Aging) and gold-standard diagnostic information (soon to include digital histology). HIGHLIGHTS: The Alzheimers Disease Neuroimaging Initiative (ADNI) Neuropathology Core (NPC) brain donation cohort now exceeds 200 cases. ADNI NPC data in National Alzheimers Coordinating Center format are available through the Laboratory of Neuro Imaging. Digitized slide files from the ADNI NPC will be available in 2025. Requests for ADNI brain tissue samples can be submitted online for ADNI/National Institute on Aging evaluation. Clinical diagnoses of Alzheimers disease (AD)/AD and related dementias (ADRD) do not always predict post mortem neuropathology. Neuropathology is essential for the development of novel AD/ADRD biomarkers.

Published
2024

8. Brain aging patterns in a large and diverse cohort of 49,482 individuals.

Author

Yang, Zhijian, Wen, Junhao, Erus, Guray, Govindarajan, Sindhuja, Melhem, Randa, Mamourian, Elizabeth, Cui, Yuhan, Srinivasan, Dhivya, Abdulkadir, Ahmed, Parmpi, Paraskevi, Wittfeld, Katharina, Grabe, Hans, Bülow, Robin, Frenzel, Stefan, Tosun, Duygu, Bilgel, Murat, An, Yang, Yi, Dahyun, Marcus, Daniel, LaMontagne, Pamela, Benzinger, Tammie, Heckbert, Susan, Austin, Thomas, Waldstein, Shari, Evans, Michele, Zonderman, Alan, Launer, Lenore, Sotiras, Aristeidis, Espeland, Mark, Masters, Colin, Maruff, Paul, Fripp, Jurgen, Toga, Arthur, OBryant, Sid, Chakravarty, Mallar, Villeneuve, Sylvia, Johnson, Sterling, Morris, John, Albert, Marilyn, Yaffe, Kristine, Völzke, Henry, Ferrucci, Luigi, Nick Bryan, R, Shinohara, Russell, Fan, Yong, Habes, Mohamad, Lalousis, Paris, Koutsouleris, Nikolaos, Wolk, David, Resnick, Susan, Shou, Haochang, Nasrallah, Ilya, and Davatzikos, Christos

Subjects
Humans, Brain, Aging, Magnetic Resonance Imaging, Male, Female, Aged, Cohort Studies, Middle Aged, Atrophy, Life Style, Adult, Aged, 80 and over
Abstract

Brain aging process is influenced by various lifestyle, environmental and genetic factors, as well as by age-related and often coexisting pathologies. Magnetic resonance imaging and artificial intelligence methods have been instrumental in understanding neuroanatomical changes that occur during aging. Large, diverse population studies enable identifying comprehensive and representative brain change patterns resulting from distinct but overlapping pathological and biological factors, revealing intersections and heterogeneity in affected brain regions and clinical phenotypes. Herein, we leverage a state-of-the-art deep-representation learning method, Surreal-GAN, and present methodological advances and extensive experimental results elucidating brain aging heterogeneity in a cohort of 49,482 individuals from 11 studies. Five dominant patterns of brain atrophy were identified and quantified for each individual by respective measures, R-indices. Their associations with biomedical, lifestyle and genetic factors provide insights into the etiology of observed variances, suggesting their potential as brain endophenotypes for genetic and lifestyle risks. Furthermore, baseline R-indices predict disease progression and mortality, capturing early changes as supplementary prognostic markers. These R-indices establish a dimensional approach to measuring aging trajectories and related brain changes. They hold promise for precise diagnostics, especially at preclinical stages, facilitating personalized patient management and targeted clinical trial recruitment based on specific brain endophenotypic expression and prognosis.

Published
2024

9. Self-interpreting Adversarial Images

Author

Zhang, Tingwei, Zhang, Collin, Morris, John X., Bagdasarian, Eugene, and Shmatikov, Vitaly

Subjects
Computer Science - Cryptography and Security, Computer Science - Artificial Intelligence, Computer Science - Machine Learning
Abstract

We introduce a new type of indirect, cross-modal injection attacks against visual language models that enable creation of self-interpreting images. These images contain hidden "meta-instructions" that control how models answer users' questions about the image and steer their outputs to express an adversary-chosen style, sentiment, or point of view. Self-interpreting images act as soft prompts, conditioning the model to satisfy the adversary's (meta-)objective while still producing answers based on the image's visual content. Meta-instructions are thus a stronger form of prompt injection. Adversarial images look natural and the model's answers are coherent and plausible--yet they also follow the adversary-chosen interpretation, e.g., political spin, or even objectives that are not achievable with explicit text instructions. We evaluate the efficacy of self-interpreting images for a variety of models, interpretations, and user prompts. We describe how these attacks could cause harm by enabling creation of self-interpreting content that carries spam, misinformation, or spin. Finally, we discuss defenses.

Published
2024

10. Longitudinal analysis of a dominantly inherited Alzheimer disease mutation carrier protected from dementia

Author

Llibre-Guerra, Jorge J., Fernandez, M. Victoria, Joseph-Mathurin, Nelly, Bian, Shijia, Carter, Kathleen, Li, Yan, Aschenbrenner, Andrew J., Pottier, Cyril, Sigurdson, Wendy, McDade, Eric, Gordon, Brian A., Renton, Alan E., Benzinger, Tammie L. S., Ibañez, Laura, Barthelemy, Nico, Johnson, Matthew, Hassenstab, Jason, Wang, Guoqiao, Goate, Alison M., Western, Dan, Wang, Ciyang, Hobbs, Diana, Daniels, Alisha, Karch, Celeste, Morris, John C., Cruchaga, Carlos, Johnson, Erik C. B., and Bateman, Randall J.

Published
2025
Full Text
View/download PDF

11. Dynamic proportional loss of functional connectivity revealed change of left superior frontal gyrus in subjective cognitive decline: an explanatory study based on Chinese and Western cohorts

Author

Wang, Luyao, Hu, Wenjing, Dong, Fan, Sheng, Can, Wu, Jinglong, Han, Ying, Jiang, Jiehui, Weiner, Michael W., Aisen, Paul, Petersen, Ronald, Jack, Clifford R., Jagust, William, Trojanowski, John Q., Toga, Arthur W., Beckett, Laurel, Green, Robert C., Saykin, Andrew J., Morris, John, Shaw, Leslie M., Khachaturian, Zaven, Sorensen, Greg, Kuller, Lew, Raichle, Marcus, Paul, Steven, Davies, Peter, Fillit, Howard, Hefti, Franz, Holtzman, David, Mesulam, Marek M., Potter, William, Snyder, Peter, Schwartz, Adam, Montine, Tom, Thomas, Ronald G., Donohue, Michael, Walter, Sarah, Gessert, Devon, Sather, Tamie, Jiminez, Gus, Harvey, Danielle, Bernstein, Matthew, Thompson, Paul, Schuff, Norbert, Borowski, Bret, Gunter, Jeff, Senjem, Matt, Vemuri, Prashanthi, Jones, David, Kantarci, Kejal, Ward, Chad, Koeppe, Robert A., Foster, Norm, Reiman, Eric M., Chen, Kewei, Mathis, Chet, Landau, Susan, Cairns, Nigel J., Householder, Erin, Taylor-Reinwald, Lisa, Lee, Virginia, Korecka, Magdalena, Figurski, Michal, Crawford, Karen, Neu, Scott, Foroud, Tatiana M., Potkin, Steven G., Shen, Li, Faber, Kelley, Kim, Sungeun, Nho, Kwangsik, Thal, Leon, Buckholtz, Neil, Albert, Marylyn, Frank, Richard, Hsiao, John, Kaye, Jeffrey, Quinn, Joseph, Lind, Betty, Carter, Raina, Dolen, Sara, Schneider, Lon S., Pawluczyk, Sonia, Beccera, Mauricio, Teodoro, Liberty, Spann, Bryan M., Brewer, James, Vanderswag, Helen, Fleisher, Adam, Heidebrink, Judith L., Lord, Joanne L., Mason, Sara S., Albers, Colleen S., Knopman, David, Johnson, Kris, Doody, Rachelle S., Villanueva-Meyer, Javier, Chowdhury, Munir, Rountree, Susan, Dang, Mimi, Stern, Yaakov, Honig, Lawrence S., Bell, Karen L., Ances, Beau, Carroll, Maria, Leon, Sue, Mintun, Mark A., Schneider, Stacy, Oliver, Angela, Marson, Daniel, Griffith, Randall, Clark, David, Geldmacher, David, Brockington, John, Roberson, Erik, Grossman, Hillel, Mitsis, Effie, de Toledo-Morrell, Leyla, Shah, Raj C., Duara, Ranjan, Varon, Daniel, Greig, Maria T., Roberts, Peggy, Onyike, Chiadi, D’Agostino, Daniel, Kielb, Stephanie, Galvin, James E., Cerbone, Brittany, Michel, Christina A., Rusinek, Henry, de Leon, Mony J., Glodzik, Lidia, De Santi, Susan, Doraiswamy, PMurali, Petrella, Jeffrey R., Wong, Terence Z., Arnold, Steven E., Karlawish, Jason H., Wolk, David, Smith, Charles D., Jicha, Greg, Hardy, Peter, Sinha, Partha, Oates, Elizabeth, Conrad, Gary, Lopez, Oscar L., Oakley, MaryAnn, Simpson, Donna M., Porsteinsson, Anton P., Goldstein, Bonnie S., Martin, Kim, Makino, Kelly M., Ismail, MSaleem, Brand, Connie, Mulnard, Ruth A., Thai, Gaby, McAdams-Ortiz, Catherine, Womack, Kyle, Mathews, Dana, Quiceno, Mary, Diaz-Arrastia, Ramon, King, Richard, Weiner, Myron, Martin-Cook, Kristen, DeVous, Michael, Levey, Allan I., Lah, James J., Cellar, Janet S., Burns, Jeffrey M., Anderson, Heather S., Swerdlow, Russell H., Apostolova, Liana, Tingus, Kathleen, Woo, Ellen, Silverman, Daniel H. S., Lu, Po H., Bartzokis, George, Graff-Radford, Neill R., Parfitt, Francine, Kendall, Tracy, Johnson, Heather, Farlow, Martin R., Hake, Ann Marie, Matthews, Brandy R., Herring, Scott, Hunt, Cynthia, van Dyck, Christopher H., Carson, Richard E., MacAvoy, Martha G., Chertkow, Howard, Bergman, Howard, Hosein, Chris, Hsiung, Ging-Yuek Robin, Feldman, Howard, Mudge, Benita, Assaly, Michele, Bernick, Charles, Munic, Donna, Kertesz, Andrew, Rogers, John, Trost, Dick, Kerwin, Diana, Lipowski, Kristine, Wu, Chuang-Kuo, Johnson, Nancy, Sadowsky, Carl, Martinez, Walter, Villena, Teresa, Turner, Raymond Scott, Johnson, Kathleen, Reynolds, Brigid, Sperling, Reisa A., Johnson, Keith A., Marshall, Gad, Frey, Meghan, Lane, Barton, Rosen, Allyson, Tinklenberg, Jared, Sabbagh, Marwan N., Belden, Christine M., Jacobson, Sandra A., Sirrel, Sherye A., Kowall, Neil, Killiany, Ronald, Budson, Andrew E., Norbash, Alexander, Johnson, Patricia Lynn, Allard, Joanne, Lerner, Alan, Ogrocki, Paula, Hudson, Leon, Fletcher, Evan, Carmichae, Owen, Olichney, John, DeCarli, Charles, Kittur, Smita, Borrie, Michael, Lee, T.-Y., Bartha, Rob, Johnson, Sterling, Asthana, Sanjay, Carlsson, Cynthia M., Preda, Adrian, Nguyen, Dana, Tariot, Pierre, Reeder, Stephanie, Bates, Vernice, Capote, Horacio, Rainka, Michelle, Scharre, Douglas W., Kataki, Maria, Adeli, Anahita, Zimmerman, Earl A., Celmins, Dzintra, Brown, Alice D., Pearlson, Godfrey D., Blank, Karen, Anderson, Karen, Santulli, Robert B., Kitzmiller, Tamar J., Schwartz, Eben S., Sink, Kaycee M., Williamson, Jeff D., Garg, Pradeep, Watkins, Franklin, Ott, Brian R., Querfurth, Henry, Tremont, Geoffrey, Salloway, Stephen, Malloy, Paul, Correia, Stephen, Rosen, Howard J., Miller, Bruce L., Mintzer, Jacobo, Spicer, Kenneth, Bachman, David, Pasternak, Stephen, Rachinsky, Irina, Drost, Dick, Pomara, Nunzio, Hernando, Raymundo, Sarrael, Antero, Schultz, Susan K., Ponto, Laura L. Boles, Shim, Hyungsub, Smith, Karen Elizabeth, Relkin, Norman, Chaing, Gloria, Raudin, Lisa, Smith, Amanda, Fargher, Kristin, Raj, Balebail Ashok, Neylan, Thomas, Grafman, Jordan, Davis, Melissa, Morrison, Rosemary, Hayes, Jacqueline, Finley, Shannon, Friedl, Karl, Fleischman, Debra, Arfanakis, Konstantinos, James, Olga, Massoglia, Dino, Fruehling, JJay, Harding, Sandra, Peskind, Elaine R., Petrie, Eric C., Li, Gail, Yesavage, Jerome A., Taylor, Joy L., and Furst, Ansgar J.

Published
2025
Full Text
View/download PDF

13. Evolution of white matter hyperintensity segmentation methods and implementation over the past two decades; an incomplete shift towards deep learning

Author

Rahmani, Maryam, Dierker, Donna, Yaeger, Lauren, Saykin, Andrew, Luckett, Patrick H., Vlassenko, Andrei G., Owens, Christopher, Jafri, Hussain, Womack, Kyle, Fripp, Jurgen, Xia, Ying, Tosun, Duygu, Benzinger, Tammie L. S., Masters, Colin L., Lee, Jin-Moo, Morris, John C., Goyal, Manu S., Strain, Jeremy F., Kukull, Walter, Weiner, Michael, Burnham, Samantha, CoxDoecke, Tim James, Fedyashov, Victor, Fripp, Jurgen, Shishegar, Rosita, Xiong, Chengjie, Marcus, Daniel, Raniga, Parnesh, Li, Shenpeng, Aschenbrenner, Andrew, Hassenstab, Jason, Lim, Yen Ying, Maruff, Paul, Sohrabi, Hamid, Robertson, Jo, Markovic, Shaun, Bourgeat, Pierrick, Doré, Vincent, Mayo, Clifford Jack, Mussoumzadeh, Parinaz, Rowe, Chris, Villemagne, Victor, Bateman, Randy, Fowler, Chris, Li, Qiao-Xin, Martins, Ralph, Schindler, Suzanne, Shaw, Les, Cruchaga, Carlos, Harari, Oscar, Laws, Simon, Porter, Tenielle, O’Brien, Eleanor, Perrin, Richard, Kukull, Walter, Bateman, Randy, McDade, Eric, Jack, Clifford, Morris, John, Yassi, Nawaf, Bourgeat, Pierrick, Perrin, Richard, Roberts, Blaine, Villemagne, Victor, Fedyashov, Victor, and Goudey, Benjamin

Published
2024
Full Text
View/download PDF

14. Crafting Interpretable Embeddings by Asking LLMs Questions

Author

Benara, Vinamra, Singh, Chandan, Morris, John X., Antonello, Richard, Stoica, Ion, Huth, Alexander G., and Gao, Jianfeng

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence, Computer Science - Machine Learning, Quantitative Biology - Neurons and Cognition
Abstract

Large language models (LLMs) have rapidly improved text embeddings for a growing array of natural-language processing tasks. However, their opaqueness and proliferation into scientific domains such as neuroscience have created a growing need for interpretability. Here, we ask whether we can obtain interpretable embeddings through LLM prompting. We introduce question-answering embeddings (QA-Emb), embeddings where each feature represents an answer to a yes/no question asked to an LLM. Training QA-Emb reduces to selecting a set of underlying questions rather than learning model weights. We use QA-Emb to flexibly generate interpretable models for predicting fMRI voxel responses to language stimuli. QA-Emb significantly outperforms an established interpretable baseline, and does so while requiring very few questions. This paves the way towards building flexible feature spaces that can concretize and evaluate our understanding of semantic brain representations. We additionally find that QA-Emb can be effectively approximated with an efficient model, and we explore broader applications in simple NLP tasks.

Published
2024

15. Extracting Prompts by Inverting LLM Outputs

Author

Zhang, Collin, Morris, John X., and Shmatikov, Vitaly

Subjects
Computer Science - Computation and Language, Computer Science - Machine Learning
Abstract

We consider the problem of language model inversion: given outputs of a language model, we seek to extract the prompt that generated these outputs. We develop a new black-box method, output2prompt, that learns to extract prompts without access to the model's logits and without adversarial or jailbreaking queries. In contrast to previous work, output2prompt only needs outputs of normal user queries. To improve memory efficiency, output2prompt employs a new sparse encoding techique. We measure the efficacy of output2prompt on a variety of user and system prompts and demonstrate zero-shot transferability across different LLMs.

Published
2024

16. Analyzing heterogeneity in Alzheimer Disease using multimodal normative modeling on imaging-based ATN biomarkers

Author

Kumar, Sayantan, Earnest, Tom, Yang, Braden, Kothapalli, Deydeep, Aschenbrenner, Andrew J., Hassenstab, Jason, Xiong, Chengie, Ances, Beau, Morris, John, Benzinger, Tammie L. S., Gordon, Brian A., Payne, Philip, and Sotiras, Aristeidis

Subjects
Quantitative Biology - Neurons and Cognition, Computer Science - Machine Learning
Abstract

INTRODUCTION: Previous studies have applied normative modeling on a single neuroimaging modality to investigate Alzheimer Disease (AD) heterogeneity. We employed a deep learning-based multimodal normative framework to analyze individual-level variation across ATN (amyloid-tau-neurodegeneration) imaging biomarkers. METHODS: We selected cross-sectional discovery (n = 665) and replication cohorts (n = 430) with available T1-weighted MRI, amyloid and tau PET. Normative modeling estimated individual-level abnormal deviations in amyloid-positive individuals compared to amyloid-negative controls. Regional abnormality patterns were mapped at different clinical group levels to assess intra-group heterogeneity. An individual-level disease severity index (DSI) was calculated using both the spatial extent and magnitude of abnormal deviations across ATN. RESULTS: Greater intra-group heterogeneity in ATN abnormality patterns was observed in more severe clinical stages of AD. Higher DSI was associated with worse cognitive function and increased risk of disease progression. DISCUSSION: Subject-specific abnormality maps across ATN reveal the heterogeneous impact of AD on the brain., Comment: Under review in Alzheimer's & Dementia

Published
2024

17. X‐chromosome-wide association study for Alzheimer’s disease

Author

Le Borgne, Julie, Gomez, Lissette, Heikkinen, Sami, Amin, Najaf, Ahmad, Shahzad, Choi, Seung Hoan, Bis, Joshua, Grenier-Boley, Benjamin, Rodriguez, Omar Garcia, Kleineidam, Luca, Young, Juan, Tripathi, Kumar Parijat, Wang, Lily, Varma, Achintya, Campos-Martin, Rafael, van der Lee, Sven, Damotte, Vincent, de Rojas, Itziar, Palmal, Sagnik, Lipton, Richard, Reiman, Eric, McKee, Ann, De Jager, Philip, Bush, William, Small, Scott, Levey, Allan, Saykin, Andrew, Foroud, Tatiana, Albert, Marilyn, Hyman, Bradley, Petersen, Ronald, Younkin, Steven, Sano, Mary, Wisniewski, Thomas, Vassar, Robert, Schneider, Julie, Henderson, Victor, Roberson, Erik, DeCarli, Charles, LaFerla, Frank, Brewer, James, Swerdlow, Russell, Van Eldik, Linda, Hamilton-Nelson, Kara, Paulson, Henry, Naj, Adam, Lopez, Oscar, Chui, Helena, Crane, Paul, Grabowski, Thomas, Kukull, Walter, Asthana, Sanjay, Craft, Suzanne, Strittmatter, Stephen, Cruchaga, Carlos, Leverenz, James, Goate, Alison, Kamboh, M. Ilyas, George-Hyslop, Peter St, Valladares, Otto, Kuzma, Amanda, Cantwell, Laura, Riemenschneider, Matthias, Morris, John, Slifer, Susan, Dalmasso, Carolina, Castillo, Atahualpa, Küçükali, Fahri, Peters, Oliver, Schneider, Anja, Dichgans, Martin, Rujescu, Dan, Scherbaum, Norbert, Deckert, Jürgen, Riedel-Heller, Steffi, Hausner, Lucrezia, Molina-Porcel, Laura, Düzel, Emrah, Grimmer, Timo, Wiltfang, Jens, Heilmann-Heimbach, Stefanie, Moebus, Susanne, Tegos, Thomas, Scarmeas, Nikolaos, Dols-Icardo, Oriol, Moreno, Fermin, Pérez-Tur, Jordi, Bullido, María J., Pastor, Pau, Sánchez-Valle, Raquel, Álvarez, Victoria, Boada, Mercè, García-González, Pablo, Puerta, Raquel, Mir, Pablo, Real, Luis M., Piñol-Ripoll, Gerard, García-Alberca, Jose María, Royo, Jose Luís, Rodriguez-Rodriguez, Eloy, Soininen, Hilkka, de Mendonça, Alexandre, Mehrabian, Shima, Traykov, Latchezar, Hort, Jakub, Vyhnalek, Martin, Thomassen, Jesper Qvist, Pijnenburg, Yolande A. L., Holstege, Henne, van Swieten, John, Ramakers, Inez, Verhey, Frans, Scheltens, Philip, Graff, Caroline, Papenberg, Goran, Giedraitis, Vilmantas, Boland, Anne, Deleuze, Jean-François, Nicolas, Gael, Dufouil, Carole, Pasquier, Florence, Hanon, Olivier, Debette, Stéphanie, Grünblatt, Edna, Popp, Julius, Ghidoni, Roberta, Galimberti, Daniela, Arosio, Beatrice, Mecocci, Patrizia, Solfrizzi, Vincenzo, Parnetti, Lucilla, Squassina, Alessio, Tremolizzo, Lucio, Borroni, Barbara, Nacmias, Benedetta, Spallazzi, Marco, Seripa, Davide, Rainero, Innocenzo, Daniele, Antonio, Bossù, Paola, Masullo, Carlo, Rossi, Giacomina, Jessen, Frank, Fernandez, Victoria, Kehoe, Patrick Gavin, Frikke-Schmidt, Ruth, Tsolaki, Magda, Sánchez-Juan, Pascual, Sleegers, Kristel, Ingelsson, Martin, Haines, Jonathan, Farrer, Lindsay, Mayeux, Richard, Wang, Li-San, Sims, Rebecca, DeStefano, Anita, Schellenberg, Gerard D., Seshadri, Sudha, Amouyel, Philippe, Williams, Julie, van der Flier, Wiesje, Ramirez, Alfredo, Pericak-Vance, Margaret, Andreassen, Ole A., Van Duijn, Cornelia, Hiltunen, Mikko, Ruiz, Agustín, Dupuis, Josée, Martin, Eden, Lambert, Jean-Charles, Kunkle, Brian, and Bellenguez, Céline

Published
2024
Full Text
View/download PDF

18. Proteogenomic analysis of human cerebrospinal fluid identifies neurologically relevant regulation and implicates causal proteins for Alzheimer’s disease

Author

Western, Daniel, Timsina, Jigyasha, Wang, Lihua, Wang, Ciyang, Yang, Chengran, Phillips, Bridget, Wang, Yueyao, Liu, Menghan, Ali, Muhammad, Beric, Aleksandra, Gorijala, Priyanka, Kohlfeld, Pat, Budde, John, Levey, Allan I., Morris, John C., Perrin, Richard J., Ruiz, Agustin, Marquié, Marta, Boada, Mercè, de Rojas, Itziar, Rutledge, Jarod, Oh, Hamilton, Wilson, Edward N., Le Guen, Yann, Reus, Lianne M., Tijms, Betty, Visser, Pieter Jelle, van der Lee, Sven J., Pijnenburg, Yolande A. L., Teunissen, Charlotte E., del Campo Milan, Marta, Alvarez, Ignacio, Aguilar, Miquel, Greicius, Michael D., Pastor, Pau, Pulford, David J., Ibanez, Laura, Wyss-Coray, Tony, Sung, Yun Ju, and Cruchaga, Carlos

Published
2024
Full Text
View/download PDF

20. Do language models plan ahead for future tokens?

Author

Wu, Wilson, Morris, John X., and Levine, Lionel

Subjects
Computer Science - Machine Learning, Computer Science - Computation and Language
Abstract

Do transformers "think ahead" during inference at a given position? It is known transformers prepare information in the hidden states of the forward pass at time step $t$ that is then used in future forward passes $t+\tau$. We posit two explanations for this phenomenon: pre-caching, in which off-diagonal gradient terms present during training result in the model computing features at $t$ irrelevant to the present inference task but useful for the future, and breadcrumbs, in which features most relevant to time step $t$ are already the same as those that would most benefit inference at time $t+\tau$. We test these hypotheses by training language models without propagating gradients to past timesteps, a scheme we formalize as myopic training. In a constructed synthetic data setting, we find clear evidence for pre-caching. In the autoregressive language modeling setting, our experiments are more suggestive of the breadcrumbs hypothesis, though pre-caching increases with model scale., Comment: 24 pages, 11 figures. Camera-ready for COLM 2024

Published
2024

21. Cybersecurity as a Service

Author

Morris, John, Tatschner, Stefan, Heinl, Michael P., Heinl, Patrizia, Newe, Thomas, and Plaga, Sven

Subjects
Computer Science - Cryptography and Security
Abstract

With the increasing sophistication and sheer number of cyberattacks, more and more companies come to the conclusion that they have to strengthen their cybersecurity posture. At the same time, well-educated Information technology (IT) security personnel are scarce. Cybersecurity as a service (CSaaS) is one possible solution to tackle this problem by outsourcing security functions to managed security service providers (MSSP). This chapter gives an overview of common CSaaS functions and their providers. Moreover, it provides guidance especially for small- and medium-sized businesses, for asking the appropriate questions when it comes to the selection of a specific MSSP., Comment: First two authors contributed equally to this work. Springer Link: https://link.springer.com/chapter/10.1007/978-3-031-45162-1_9

Published
2024
Full Text
View/download PDF

22. Nomic Embed: Training a Reproducible Long Context Text Embedder

Author

Nussbaum, Zach, Morris, John X., Duderstadt, Brandon, and Mulyar, Andriy

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

This technical report describes the training of nomic-embed-text-v1, the first fully reproducible, open-source, open-weights, open-data, 8192 context length English text embedding model that outperforms both OpenAI Ada-002 and OpenAI text-embedding-3-small on the short-context MTEB benchmark and the long context LoCo benchmark. We release the training code and model weights under an Apache 2.0 license. In contrast with other open-source models, we release the full curated training data and code that allows for full replication of nomic-embed-text-v1. You can find code and data to replicate the model at https://github.com/nomic-ai/contrastors., Comment: Accepted to TMLR https://openreview.net/forum?id=IPmzyQSiQE

Published
2024

24. Molecular neuroimaging in dominantly inherited versus sporadic early-onset Alzheimer’s disease

Author

Iaccarino, Leonardo, Llibre-Guerra, Jorge J, McDade, Eric, Edwards, Lauren, Gordon, Brian, Benzinger, Tammie, Hassenstab, Jason, Kramer, Joel H, Li, Yan, Miller, Bruce L, Miller, Zachary, Morris, John C, Mundada, Nidhi, Perrin, Richard J, Rosen, Howard J, Soleimani-Meigooni, David, Strom, Amelia, Tsoy, Elena, Wang, Guoqiao, Xiong, Chengjie, Allegri, Ricardo, Chrem, Patricio, Vazquez, Silvia, Berman, Sarah B, Chhatwal, Jasmeer, Masters, Colin L, Farlow, Martin R, Jucker, Mathias, Levin, Johannes, Salloway, Stephen, Fox, Nick C, Day, Gregory S, Gorno-Tempini, Maria Luisa, Boxer, Adam L, La Joie, Renaud, Bateman, Randall, and Rabinovici, Gil D

Subjects
Biomedical and Clinical Sciences, Biological Psychology, Clinical Sciences, Neurosciences, Psychology, Clinical Research, Dementia, Aging, Alzheimer's Disease including Alzheimer's Disease Related Dementias (AD/ADRD), Acquired Cognitive Impairment, Neurodegenerative, Alzheimer's Disease, Brain Disorders, Biomedical Imaging, 2.1 Biological and endogenous factors, 4.2 Evaluation of markers and technologies, Neurological, Good Health and Well Being, Alzheimer's disease, amyloid-PET, brain glucose metabolism, FDG-PET, Alzheimer’s disease, Clinical sciences, Biological psychology
Abstract

Approximately 5% of Alzheimer's disease patients develop symptoms before age 65 (early-onset Alzheimer's disease), with either sporadic (sporadic early-onset Alzheimer's disease) or dominantly inherited (dominantly inherited Alzheimer's disease) presentations. Both sporadic early-onset Alzheimer's disease and dominantly inherited Alzheimer's disease are characterized by brain amyloid-β accumulation, tau tangles, hypometabolism and neurodegeneration, but differences in topography and magnitude of these pathological changes are not fully elucidated. In this study, we directly compared patterns of amyloid-β plaque deposition and glucose hypometabolism in sporadic early-onset Alzheimer's disease and dominantly inherited Alzheimer's disease individuals. Our analysis included 134 symptomatic sporadic early-onset Alzheimer's disease amyloid-Positron Emission Tomography (PET)-positive cases from the University of California, San Francisco, Alzheimer's Disease Research Center (mean ± SD age 59.7 ± 5.6 years), 89 symptomatic dominantly inherited Alzheimer's disease cases (age 45.8 ± 9.3 years) and 102 cognitively unimpaired non-mutation carriers from the Dominantly Inherited Alzheimer Network study (age 44.9 ± 9.2). Each group underwent clinical and cognitive examinations, 11C-labelled Pittsburgh Compound B-PET and structural MRI. 18F-Fluorodeoxyglucose-PET was also available for most participants. Positron Emission Tomography scans from both studies were uniformly processed to obtain a standardized uptake value ratio (PIB50-70 cerebellar grey reference and FDG30-60 pons reference) images. Statistical analyses included pairwise global and voxelwise group comparisons and group-independent component analyses. Analyses were performed also adjusting for covariates including age, sex, Mini-Mental State Examination, apolipoprotein ε4 status and average composite cortical of standardized uptake value ratio. Compared with dominantly inherited Alzheimer's disease, sporadic early-onset Alzheimer's disease participants were older at age of onset (mean ± SD, 54.8 ± 8.2 versus 41.9 ± 8.2, Cohen's d = 1.91), with more years of education (16.4 ± 2.8 versus 13.5 ± 3.2, d = 1) and more likely to be apolipoprotein ε4 carriers (54.6% ε4 versus 28.1%, Cramer's V = 0.26), but similar Mini-Mental State Examination (20.6 ± 6.1 versus 21.2 ± 7.4, d = 0.08). Sporadic early-onset Alzheimer's disease had higher global cortical Pittsburgh Compound B-PET binding (mean ± SD standardized uptake value ratio, 1.92 ± 0.29 versus 1.58 ± 0.44, d = 0.96) and greater global cortical 18F-fluorodeoxyglucose-PET hypometabolism (mean ± SD standardized uptake value ratio, 1.32 ± 0.1 versus 1.39 ± 0.19, d = 0.48) compared with dominantly inherited Alzheimer's disease. Fully adjusted comparisons demonstrated relatively higher Pittsburgh Compound B-PET standardized uptake value ratio in the medial occipital, thalami, basal ganglia and medial/dorsal frontal regions in dominantly inherited Alzheimer's disease versus sporadic early-onset Alzheimer's disease. Sporadic early-onset Alzheimer's disease showed relatively greater 18F-fluorodeoxyglucose-PET hypometabolism in Alzheimer's disease signature temporoparietal regions and caudate nuclei, whereas dominantly inherited Alzheimer's disease showed relatively greater hypometabolism in frontal white matter and pericentral regions. Independent component analyses largely replicated these findings by highlighting common and unique Pittsburgh Compound B-PET and 18F-fluorodeoxyglucose-PET binding patterns. In summary, our findings suggest both common and distinct patterns of amyloid and glucose hypometabolism in sporadic and dominantly inherited early-onset Alzheimer's disease.

Published
2024

25. Highly accurate blood test for Alzheimer’s disease is similar or superior to clinical cerebrospinal fluid tests

Author

Barthélemy, Nicolas R, Salvadó, Gemma, Schindler, Suzanne E, He, Yingxin, Janelidze, Shorena, Collij, Lyduine E, Saef, Benjamin, Henson, Rachel L, Chen, Charles D, Gordon, Brian A, Li, Yan, La Joie, Renaud, Benzinger, Tammie LS, Morris, John C, Mattsson-Carlgren, Niklas, Palmqvist, Sebastian, Ossenkoppele, Rik, Rabinovici, Gil D, Stomrud, Erik, Bateman, Randall J, and Hansson, Oskar

Subjects
Biomedical and Clinical Sciences, Health Sciences, Neurodegenerative, Alzheimer's Disease including Alzheimer's Disease Related Dementias (AD/ADRD), Acquired Cognitive Impairment, Aging, Dementia, Neurosciences, Alzheimer's Disease, Brain Disorders, Biomedical Imaging, 4.2 Evaluation of markers and technologies, 4.1 Discovery and preclinical testing of markers and technologies, Neurological, Humans, Alzheimer Disease, tau Proteins, Biomarkers, Amyloid beta-Peptides, Cognitive Dysfunction, Hematologic Tests, Positron-Emission Tomography, Medical and Health Sciences, Immunology, Biomedical and clinical sciences, Health sciences
Abstract

With the emergence of Alzheimer's disease (AD) disease-modifying therapies, identifying patients who could benefit from these treatments becomes critical. In this study, we evaluated whether a precise blood test could perform as well as established cerebrospinal fluid (CSF) tests in detecting amyloid-β (Aβ) plaques and tau tangles. Plasma %p-tau217 (ratio of phosporylated-tau217 to non-phosphorylated tau) was analyzed by mass spectrometry in the Swedish BioFINDER-2 cohort (n = 1,422) and the US Charles F. and Joanne Knight Alzheimer Disease Research Center (Knight ADRC) cohort (n = 337). Matched CSF samples were analyzed with clinically used and FDA-approved automated immunoassays for Aβ42/40 and p-tau181/Aβ42. The primary and secondary outcomes were detection of brain Aβ or tau pathology, respectively, using positron emission tomography (PET) imaging as the reference standard. Main analyses were focused on individuals with cognitive impairment (mild cognitive impairment and mild dementia), which is the target population for available disease-modifying treatments. Plasma %p-tau217 was clinically equivalent to FDA-approved CSF tests in classifying Aβ PET status, with an area under the curve (AUC) for both between 0.95 and 0.97. Plasma %p-tau217 was generally superior to CSF tests in classification of tau-PET with AUCs of 0.95-0.98. In cognitively impaired subcohorts (BioFINDER-2: n = 720; Knight ADRC: n = 50), plasma %p-tau217 had an accuracy, a positive predictive value and a negative predictive value of 89-90% for Aβ PET and 87-88% for tau PET status, which was clinically equivalent to CSF tests, further improving to 95% using a two-cutoffs approach. Blood plasma %p-tau217 demonstrated performance that was clinically equivalent or superior to clinically used FDA-approved CSF tests in the detection of AD pathology. Use of high-performance blood tests in clinical practice can improve access to accurate AD diagnosis and AD-specific treatments.

Published
2024

26. Biomedical knowledge graph-optimized prompt generation for large language models

Author

Soman, Karthik, Rose, Peter W, Morris, John H, Akbas, Rabia E, Smith, Brett, Peetoom, Braian, Villouta-Reyes, Catalina, Cerono, Gabriel, Shi, Yongmei, Rizk-Jackson, Angela, Israni, Sharat, Nelson, Charlotte A, Huang, Sui, and Baranzini, Sergio E

Subjects
Computer Science - Computation and Language
Abstract

Large Language Models (LLMs) are being adopted at an unprecedented rate, yet still face challenges in knowledge-intensive domains like biomedicine. Solutions such as pre-training and domain-specific fine-tuning add substantial computational overhead, requiring further domain expertise. Here, we introduce a token-optimized and robust Knowledge Graph-based Retrieval Augmented Generation (KG-RAG) framework by leveraging a massive biomedical KG (SPOKE) with LLMs such as Llama-2-13b, GPT-3.5-Turbo and GPT-4, to generate meaningful biomedical text rooted in established knowledge. Compared to the existing RAG technique for Knowledge Graphs, the proposed method utilizes minimal graph schema for context extraction and uses embedding methods for context pruning. This optimization in context extraction results in more than 50% reduction in token consumption without compromising the accuracy, making a cost-effective and robust RAG implementation on proprietary LLMs. KG-RAG consistently enhanced the performance of LLMs across diverse biomedical prompts by generating responses rooted in established knowledge, accompanied by accurate provenance and statistical evidence (if available) to substantiate the claims. Further benchmarking on human curated datasets, such as biomedical true/false and multiple-choice questions (MCQ), showed a remarkable 71% boost in the performance of the Llama-2 model on the challenging MCQ dataset, demonstrating the framework's capacity to empower open-source models with fewer parameters for domain specific questions. Furthermore, KG-RAG enhanced the performance of proprietary GPT models, such as GPT-3.5 and GPT-4. In summary, the proposed framework combines explicit and implicit knowledge of KG and LLM in a token optimized fashion, thus enhancing the adaptability of general-purpose LLMs to tackle domain-specific questions in a cost-effective fashion., Comment: 29 pages, 5 figures, 1 table, 1 supplementary file

Published
2023

27. Language Model Inversion

Author

Morris, John X., Zhao, Wenting, Chiu, Justin T., Shmatikov, Vitaly, and Rush, Alexander M.

Subjects
Computer Science - Computation and Language, Computer Science - Machine Learning
Abstract

Language models produce a distribution over the next token; can we use this information to recover the prompt tokens? We consider the problem of language model inversion and show that next-token probabilities contain a surprising amount of information about the preceding text. Often we can recover the text in cases where it is hidden from the user, motivating a method for recovering unknown prompts given only the model's current distribution output. We consider a variety of model access scenarios, and show how even without predictions for every token in the vocabulary we can recover the probability vector through search. On Llama-2 7b, our inversion method reconstructs prompts with a BLEU of $59$ and token-level F1 of $78$ and recovers $27\%$ of prompts exactly. Code for reproducing all experiments is available at http://github.com/jxmorris12/vec2text.

Published
2023

28. Tree Prompting: Efficient Task Adaptation without Fine-Tuning

Author

Morris, John X., Singh, Chandan, Rush, Alexander M., Gao, Jianfeng, and Deng, Yuntian

Subjects
Computer Science - Computation and Language, Computer Science - Machine Learning
Abstract

Prompting language models (LMs) is the main interface for applying them to new tasks. However, for smaller LMs, prompting provides low accuracy compared to gradient-based finetuning. Tree Prompting is an approach to prompting which builds a decision tree of prompts, linking multiple LM calls together to solve a task. At inference time, each call to the LM is determined by efficiently routing the outcome of the previous call using the tree. Experiments on classification datasets show that Tree Prompting improves accuracy over competing methods and is competitive with fine-tuning. We also show that variants of Tree Prompting allow inspection of a model's decision-making process., Comment: Both first authors contributed equally; accepted to EMNLP 2023

Published
2023

29. Text Embeddings Reveal (Almost) As Much As Text

Author

Morris, John X., Kuleshov, Volodymyr, Shmatikov, Vitaly, and Rush, Alexander M.

Subjects
Computer Science - Computation and Language, Computer Science - Machine Learning
Abstract

How much private information do text embeddings reveal about the original text? We investigate the problem of embedding \textit{inversion}, reconstructing the full text represented in dense text embeddings. We frame the problem as controlled generation: generating text that, when reembedded, is close to a fixed point in latent space. We find that although a na\"ive model conditioned on the embedding performs poorly, a multi-step method that iteratively corrects and re-embeds text is able to recover $92\%$ of $32\text{-token}$ text inputs exactly. We train our model to decode text embeddings from two state-of-the-art embedding models, and also show that our model can recover important personal information (full names) from a dataset of clinical notes. Our code is available on Github: \href{https://github.com/jxmorris12/vec2text}{github.com/jxmorris12/vec2text}., Comment: Accepted at EMNLP 2023

Published
2023

30. Accuracy of TrUE-Net in comparison to established white matter hyperintensity segmentation methods: An independent validation study.

Author

Strain, Jeremy, Rahmani, Maryam, Dierker, Donna, Owen, Christopher, Jafri, Hussain, Vlassenko, Andrei, Womack, Kyle, Fripp, Jurgen, Tosun, Duygu, Benzinger, Tammie, Weiner, Michael, Masters, Colin, Lee, Jin-Moo, Morris, John, and Goyal, Manu

Subjects
Aging, LST, Segmentation Tools, TrUE-Net, WMH, Humans, White Matter, Magnetic Resonance Imaging, Brain, Algorithms, Aging, Leukoaraiosis
Abstract

White matter hyperintensities (WMH) are nearly ubiquitous in the aging brain, and their topography and overall burden are associated with cognitive decline. Given their numerosity, accurate methods to automatically segment WMH are needed. Recent developments, including the availability of challenge data sets and improved deep learning algorithms, have led to a new promising deep-learning based automated segmentation model called TrUE-Net, which has yet to undergo rigorous independent validation. Here, we compare TrUE-Net to six established automated WMH segmentation tools, including a semi-manual method. We evaluated the techniques at both global and regional level to compare their ability to detect the established relationship between WMH burden and age. We found that TrUE-Net was highly reliable at identifying WMH regions with low false positive rates, when compared to semi-manual segmentation as the reference standard. TrUE-Net performed similarly or favorably when compared to the other automated techniques. Moreover, TrUE-Net was able to detect relationships between WMH and age to a similar degree as the reference standard semi-manual segmentation at both the global and regional level. These results support the use of TrUE-Net for identifying WMH at the global or regional level, including in large, combined datasets.

Published
2024

31. Biomedical knowledge graph-optimized prompt generation for large language models

Author

Soman, Karthik, Rose, Peter W, Morris, John H, Akbas, Rabia E, Smith, Brett, Peetoom, Braian, Villouta-Reyes, Catalina, Cerono, Gabriel, Shi, Yongmei, Rizk-Jackson, Angela, Israni, Sharat, Nelson, Charlotte A, Huang, Sui, and Baranzini, Sergio E

Subjects
Data Management and Data Science, Information and Computing Sciences, Networking and Information Technology R&D (NITRD), Mathematical Sciences, Biological Sciences, Bioinformatics, Biological sciences, Information and computing sciences, Mathematical sciences
Abstract

MotivationLarge Language Models (LLMs) are being adopted at an unprecedented rate, yet still face challenges in knowledge-intensive domains like biomedicine. Solutions such as pre-training and domain-specific fine-tuning add substantial computational overhead, requiring further domain-expertise. Here, we introduce a token-optimized and robust Knowledge Graph-based Retrieval Augmented Generation (KG-RAG) framework by leveraging a massive biomedical KG (SPOKE) with LLMs such as Llama-2-13b, GPT-3.5-Turbo and GPT-4, to generate meaningful biomedical text rooted in established knowledge.ResultsCompared to the existing RAG technique for Knowledge Graphs, the proposed method utilizes minimal graph schema for context extraction and uses embedding methods for context pruning. This optimization in context extraction results in more than 50% reduction in token consumption without compromising the accuracy, making a cost-effective and robust RAG implementation on proprietary LLMs. KG-RAG consistently enhanced the performance of LLMs across diverse biomedical prompts by generating responses rooted in established knowledge, accompanied by accurate provenance and statistical evidence (if available) to substantiate the claims. Further benchmarking on human curated datasets, such as biomedical true/false and multiple-choice questions (MCQ), showed a remarkable 71% boost in the performance of the Llama-2 model on the challenging MCQ dataset, demonstrating the framework's capacity to empower open-source models with fewer parameters for domain-specific questions. Furthermore, KG-RAG enhanced the performance of proprietary GPT models, such as GPT-3.5 and GPT-4. In summary, the proposed framework combines explicit and implicit knowledge of KG and LLM in a token optimized fashion, thus enhancing the adaptability of general-purpose LLMs to tackle domain-specific questions in a cost-effective fashion.Availability and implementationSPOKE KG can be accessed at https://spoke.rbvi.ucsf.edu/neighborhood.html. It can also be accessed using REST-API (https://spoke.rbvi.ucsf.edu/swagger/). KG-RAG code is made available at https://github.com/BaranziniLab/KG_RAG. Biomedical benchmark datasets used in this study are made available to the research community in the same GitHub repository.Supplementary informationSupplementary data are available at Bioinformatics online.

Published
2024

32. Robust, fully-automated assessment of cerebral perivascular spaces and white matter lesions: a multicentre MRI longitudinal study of their evolution and association with risk of dementia and accelerated brain atrophy

Author

Weiner, Michael, Aisen, Paul, Petersen, Ronald, Jack, Clifford R., Jagust, William, Trojanowki, John Q., Toga, Arthur W., Beckett, Laurel, Green, Robert C., Saykin, Andrew J., Morris, John, Shaw, Leslie M., Liu, Enchi, Montine, Tom, Thomas, Ronald G., Donohue, Michael, Walter, Sarah, Gessert, Devon, Sather, Tamie, Jiminez, Gus, Harvey, Danielle, Bernstein, Matthew, Fox, Nick, Thompson, Paul, Schuff, Norbert, DeCarli, Charles, Borowski, Bret, Gunter, Jeff, Senjem, Matt, Vemuri, Prashanthi, Jones, David, Kantarci, Kejal, Ward, Chad, Koeppe, Robert A., Foster, Norm, Reiman, Eric M., Chen, Kewei, Mathis, Chet, Landau, Susan, Cairns, Nigel J., Householder, Erin, Reinwald, Lisa Taylor, Lee, Virginia, Korecka, Magdalena, Figurski, Michal, Crawford, Karen, Neu, Scott, Foroud, Tatiana M., Potkin, Steven, Shen, Li, Kelley, Faber, Kim, Sungeun, Nho, Kwangsik, Kachaturian, Zaven, Frank, Richard, Snyder, Peter J., Molchan, Susan, Kaye, Jeffrey, Quinn, Joseph, Lind, Betty, Carter, Raina, Dolen, Sara, Schneider, Lon S., Pawluczyk, Sonia, Beccera, Mauricio, Teodoro, Liberty, Spann, Bryan M., Brewer, James, Vanderswag, Helen, Fleisher, Adam, Heidebrink, Judith L., Lord, Joanne L., Mason, Sara S., Albers, Colleen S., Knopman, David, Johnson, Kris, Doody, Rachelle S., Meyer, Javier Villanueva, Chowdhury, Munir, Rountree, Susan, Dang, Mimi, Stern, Yaakov, Honig, Lawrence S., Bell, Karen L., Ances, Beau, Morris, John C., Carroll, Maria, Leon, Sue, Mintun, Mark A., Schneider, Stacy, Oliver, Angela, Marson, Daniel, Griffith, Randall, Clark, David, Geld-macher, David, Brockington, John, Roberson, Erik, Grossman, Hillel, Mitsis, Effie, Leyla deToledo-Morrell, Shah, Raj C., Duara, Ranjan, Varon, Daniel, Greig, Maria T., Roberts, Peggy, Albert, Marilyn, Onyike, Chiadi, D’Agostino, Daniel, Kielb, Stephanie, Galvin, James E., Pogorelec, Dana M., Cerbone, Brittany, Michel, Christina A., Rusinek, Henry, de Leon, Mony J., Glodzik, Lidia, De Santi, Susan, Doraiswamy, P. Murali, Petrella, Jeffrey R., Wong, Terence Z., Arnold, Steven E., Karlawish, Jason H., Wolk, David, Smith, Charles D., Jicha, Greg, Hardy, Peter, Sinha, Partha, Oates, Elizabeth, Conrad, Gary, Lopez, Oscar L., Oakley, MaryAnn, Simpson, Donna M., Porsteinsson, Anton P., Goldstein, Bonnie S., Martin, Kim, Makino, Kelly M., Ismail, M. Saleem, Brand, Connie, Mulnard, Ruth A., Thai, Gaby, Adams Ortiz, Catherine Mc, Womack, Kyle, Mathews, Dana, Quiceno, Mary, Arrastia, Ramon Diaz, King, Richard, Weiner, Myron, Cook, Kristen Martin, DeVous, Michael, Levey, Allan I., Lah, James J., Cellar, Janet S., Burns, Jeffrey M., Anderson, Heather S., Swerdlow, Russell H., Apostolova, Liana, Tingus, Kathleen, Woo, Ellen, Silverman, Daniel H.S., Lu, Po H., Bartzokis, George, Graff Radford, Neill R., Parfitt, Francine, Kendall, Tracy, Johnson, Heather, Farlow, Martin R., Hake, AnnMarie, Matthews, Brandy R., Herring, Scott, Hunt, Cynthia, van Dyck, Christopher H., Carson, Richard E., MacAvoy, Martha G., Chertkow, Howard, Bergman, Howard, Hosein, Chris, Black, Sandra, Stefanovic, Bojana, Caldwell, Curtis, Robin Hsiung, Ging-Yuek, Feldman, Howard, Mudge, Benita, Assaly, Michele, Kertesz, Andrew, Rogers, John, Trost, Dick, Bernick, Charles, Munic, Donna, Kerwin, Diana, Mesulam, Marek Marsel, Lipowski, Kristine, Wu, Chuang Kuo, Johnson, Nancy, Sadowsky, Carl, Martinez, Walter, Villena, Teresa, Turner, Raymond Scott, Johnson, Kathleen, Reynolds, Brigid, Sperling, Reisa A., Johnson, Keith A., Marshall, Gad, Frey, Meghan, Yesavage, Jerome, Taylor, Joy L., Lane, Barton, Rosen, Allyson, Tinklenberg, Jared, Sabbagh, Marwan N., Belden, Christine M., Jacobson, Sandra A., Sirrel, Sherye A., Kowall, Neil, Killiany, Ronald, Budson, Andrew E., Norbash, Alexander, Johnson, Patricia Lynn, Obisesan, Thomas O., Wolday, Saba, Allard, Joanne, Lerner, Alan, Ogrocki, Paula, Hudson, Leon, Fletcher, Evan, Carmichael, Owen, Olichney, John, Kittur, Smita, Borrie, Michael, Lee, T.Y., Bartha, Rob, Johnson, Sterling, Asthana, Sanjay, Carlsson, Cynthia M., Potkin, Steven G., Preda, Adrian, Nguyen, Dana, Tariot, Pierre, Reeder, Stephanie, Bates, Vernice, Capote, Horacio, Rainka, Michelle, Scharre, Douglas W., Kataki, Maria, Adeli, Anahita, Zimmerman, Earl A., Celmins, Dzintra, Brown, Alice D., Pearlson, Godfrey D., Blank, Karen, Anderson, Karen, Santulli, Robert B., Kitzmiller, Tamar J., Schwartz, Eben S., Sink, Kaycee M., Williamson, Jeff D., Garg, Pradeep, Watkins, Franklin, Ott, Brian R., Querfurth, Henry, Tremont, Geoffrey, Salloway, Stephen, Malloy, Paul, Correia, Stephen, Rosen, Howard J., Miller, Bruce L., Mintzer, Jacobo, Spicer, Kenneth, Bachman, David, Finger, Elizabeth, Pasternak, Stephen, Rachinsky, Irina, Drost, Dick, Pomara, Nunzio, Hernando, Raymundo, Sarrael, Antero, Schultz, Susan K., Boles Ponto, Laura L., Shim, Hyungsub, Smith, Karen Elizabeth, Relkin, Norman, Chaing, Gloria, Raudin, Lisa, Smith, Amanda, Fargher, Kristin, Raj, Balebail Ashok, Barisano, Giuseppe, Iv, Michael, Choupan, Jeiran, and Hayden-Gephart, Melanie

Published
2025
Full Text
View/download PDF

34. Genetic and clinical correlates of two neuroanatomical AI dimensions in the Alzheimer’s disease continuum

Author

Wen, Junhao, Yang, Zhijian, Nasrallah, Ilya M., Cui, Yuhan, Erus, Guray, Srinivasan, Dhivya, Abdulkadir, Ahmed, Mamourian, Elizabeth, Hwang, Gyujoon, Singh, Ashish, Bergman, Mark, Bao, Jingxuan, Varol, Erdem, Zhou, Zhen, Boquet-Pujadas, Aleix, Chen, Jiong, Toga, Arthur W., Saykin, Andrew J., Hohman, Timothy J., Thompson, Paul M., Villeneuve, Sylvia, Gollub, Randy, Sotiras, Aristeidis, Wittfeld, Katharina, Grabe, Hans J., Tosun, Duygu, Bilgel, Murat, An, Yang, Marcus, Daniel S., LaMontagne, Pamela, Benzinger, Tammie L., Heckbert, Susan R., Austin, Thomas R., Launer, Lenore J., Espeland, Mark, Masters, Colin L., Maruff, Paul, Fripp, Jurgen, Johnson, Sterling C., Morris, John C., Albert, Marilyn S., Bryan, R. Nick, Resnick, Susan M., Ferrucci, Luigi, Fan, Yong, Habes, Mohamad, Wolk, David, Shen, Li, Shou, Haochang, and Davatzikos, Christos

Published
2024
Full Text
View/download PDF

35. Genetic and multi-omic resources for Alzheimer disease and related dementia from the Knight Alzheimer Disease Research Center

Author

Fernandez, Maria Victoria, Liu, Menghan, Beric, Aleksandra, Johnson, Matt, Cetin, Arda, Patel, Maulik, Budde, John, Kohlfeld, Pat, Bergmann, Kristy, Lowery, Joseph, Flynn, Allison, Brock, William, Sanchez Montejo, Brenda, Gentsch, Jen, Sykora, Nicholas, Norton, Joanne, Gentsch, Jen, Valdez, Olga, Gorijala, Priyanka, Sanford, Jessie, Sun, Yichen, Wang, Ciyang, Western, Dan, Timsina, Jigyasha, Mangetti Goncalves, Tassia, Do, Anh N., Sung, Yun Ju, Zhao, Guoyan, Morris, John C., Moulder, Krista, Holtzman, David M., Bateman, Randall J., Karch, Celeste, Hassenstab, Jason, Xiong, Chengjie, Schindler, Suzanne E., Balls-Berry, Joyce (Joy), Benzinger, Tammie L. S., Perrin, Richard J., Denny, Andrea, Snider, B. Joy, Stark, Susan L., Ibanez, Laura, and Cruchaga, Carlos

Published
2024
Full Text
View/download PDF

36. Baseline levels and longitudinal changes in plasma Aβ42/40 among Black and white individuals

Author

Xiong, Chengjie, Luo, Jingqin, Wolk, David A., Shaw, Leslie M., Roberson, Erik D., Murchison, Charles F., Henson, Rachel L., Benzinger, Tammie L. S., Bui, Quoc, Agboola, Folasade, Grant, Elizabeth, Gremminger, Emily N., Moulder, Krista L., Geldmacher, David S., Clay, Olivio J., Babulal, Ganesh, Cruchaga, Carlos, Holtzman, David M., Bateman, Randall J., Morris, John C., and Schindler, Suzanne E.

Published
2024
Full Text
View/download PDF

38. Gene-SGAN: discovering disease subtypes with imaging and genetic signatures via multi-view weakly-supervised deep clustering

Author

Yang, Zhijian, Wen, Junhao, Abdulkadir, Ahmed, Cui, Yuhan, Erus, Guray, Mamourian, Elizabeth, Melhem, Randa, Srinivasan, Dhivya, Govindarajan, Sindhuja T., Chen, Jiong, Habes, Mohamad, Masters, Colin L., Maruff, Paul, Fripp, Jurgen, Ferrucci, Luigi, Albert, Marilyn S., Johnson, Sterling C., Morris, John C., LaMontagne, Pamela, Marcus, Daniel S., Benzinger, Tammie L. S., Wolk, David A., Shen, Li, Bao, Jingxuan, Resnick, Susan M., Shou, Haochang, Nasrallah, Ilya M., and Davatzikos, Christos

Published
2024
Full Text
View/download PDF

39. Personality predictors of dementia diagnosis and neuropathological burden: An individual participant data meta‐analysis

Author

Beck, Emorie D, Yoneda, Tomiko, James, Bryan D, Bennett, David A, Hassenstab, Jason, Katz, Mindy J, Lipton, Richard B, Morris, John, Mroczek, Daniel K, and Graham, Eileen K

Subjects
Social and Personality Psychology, Psychology, Aging, Dementia, Behavioral and Social Science, Prevention, Neurodegenerative, Acquired Cognitive Impairment, Clinical Research, Neurosciences, Brain Disorders, Neurological, agreeableness, arteriosclerosis, Braak stage, CERAD, cerebral amyloid angiopathy, cerebral atherosclerosis, extraversion, gross cerebral infarcts, gross cerebral microinfarcts, hippocampal sclerosis, individual participant data meta-analysis, Lewy body disease, openness, positive affect, satisfaction with life, TDP-43, Clinical Sciences, Geriatrics, Clinical sciences, Biological psychology
Abstract

IntroductionThe extent to which the Big Five personality traits and subjective well-being (SWB) are discriminatory predictors of clinical manifestation of dementia versus dementia-related neuropathology is unclear.MethodsUsing data from eight independent studies (Ntotal = 44,531; Ndementia = 1703; baseline Mage = 49 to 81 years, 26 to 61% female; Mfollow-up range = 3.53 to 21.00 years), Bayesian multilevel models tested whether personality traits and SWB differentially predicted neuropsychological and neuropathological characteristics of dementia.ResultsSynthesized and individual study results indicate that high neuroticism and negative affect and low conscientiousness, extraversion, and positive affect were associated with increased risk of long-term dementia diagnosis. There were no consistent associations with neuropathology.DiscussionThis multistudy project provides robust, conceptually replicated and extended evidence that psychosocial factors are strong predictors of dementia diagnosis but not consistently associated with neuropathology at autopsy.HighlightsN(+), C(-), E(-), PA(-), and NA(+) were associated with incident diagnosis. Results were consistent despite self-report versus clinical diagnosis of dementia. Psychological factors were not associated with neuropathology at autopsy. Individuals with higher conscientiousness and no diagnosis had less neuropathology. High C individuals may withstand neuropathology for longer before death.

Published
2023

40. Longitudinal clinical, cognitive and biomarker profiles in dominantly inherited versus sporadic early-onset Alzheimer’s disease

Author

Llibre-Guerra, Jorge J, Iaccarino, Leonardo, Coble, Dean, Edwards, Lauren, Li, Yan, McDade, Eric, Strom, Amelia, Gordon, Brian, Mundada, Nidhi, Schindler, Suzanne E, Tsoy, Elena, Ma, Yinjiao, Lu, Ruijin, Fagan, Anne M, Benzinger, Tammie LS, Soleimani-Meigooni, David, Aschenbrenner, Andrew J, Miller, Zachary, Wang, Guoqiao, Kramer, Joel H, Hassenstab, Jason, Rosen, Howard J, Morris, John C, Miller, Bruce L, Xiong, Chengjie, Perrin, Richard J, Allegri, Ricardo, Chrem, Patricio, Surace, Ezequiel, Berman, Sarah B, Chhatwal, Jasmeer, Masters, Colin L, Farlow, Martin R, Jucker, Mathias, Levin, Johannes, Fox, Nick C, Day, Gregory, Gorno-Tempini, Maria Luisa, Boxer, Adam L, La Joie, Renaud, Rabinovici, Gil D, and Bateman, Randall

Subjects
Biomedical and Clinical Sciences, Biological Psychology, Clinical Sciences, Neurosciences, Psychology, Brain Disorders, Alzheimer's Disease including Alzheimer's Disease Related Dementias (AD/ADRD), Dementia, Neurodegenerative, Genetics, Acquired Cognitive Impairment, Alzheimer's Disease, Aging, Clinical Research, 2.1 Biological and endogenous factors, Neurological, early-onset Alzheimer's disease, sporadic, dominantly inherited, early-onset Alzheimer’s disease, Clinical sciences, Biological psychology
Abstract

Approximately 5% of Alzheimer's disease cases have an early age at onset (

Published
2023

41. Circular-SWAT for deep learning based diagnostic classification of Alzheimer's disease: application to metabolome data

Author

Jo, Taeho, Kim, Junpyo, Bice, Paula, Huynh, Kevin, Wang, Tingting, Arnold, Matthias, Meikle, Peter J, Giles, Corey, Kaddurah-Daouk, Rima, Saykin, Andrew J, Nho, Kwangsik, Kueider-Paisley, Alexandra, Doraiswamy, P Murali, Blach, Colette, Moseley, Arthur, Thompson, Will, St John-Williams, Lisa, Mahmoudiandehkhordi, Siamak, Tenenbaum, Jessica, Welsh-Balmer, Kathleen, Plassman, Brenda, Risacher, Shannon L, Kastenmüller, Gabi, Han, Xianlin, Baillie, Rebecca, Knight, Rob, Dorrestein, Pieter, Brewer, James, Mayer, Emeran, Labus, Jennifer, Baldi, Pierre, Gupta, Arpana, Fiehn, Oliver, Barupal, Dinesh, Meikle, Peter, Mazmanian, Sarkis, Rader, Dan, Kling, Mitchel, Shaw, Leslie, Trojanowski, John, van Duijin, Cornelia, Nevado-Holgado, Alejo, Bennett, David, Krishnan, Ranga, Keshavarzian, Ali, Vogt, Robin, Ikram, Arfan, Hankemeier, Thomas, Thiele, Ines, Price, Nathan, Funk, Cory, Baloni, Priyanka, Jia, Wei, Wishart, David, Brinton, Roberta, Chang, Rui, Farrer, Lindsay, Au, Rhoda, Qiu, Wendy, Würtz, Peter, Koal, Therese, Mangravite, Lara, Krumsiek, Jan, Suhre, Karsten, Newman, John, Moreno, Herman, Foroud, Tatania, Sacks, Frank, Jansson, Janet, Weiner, Michael W, Aisen, Paul, Petersen, Ronald, Jack, Clifford R, Jagust, William, Trojanowki, John Q, Toga, Arthur W, Beckett, Laurel, Green, Robert C, Morris, John C, Perrin, Richard J, Shaw, Leslie M, Khachaturian, Zaven, Carrillo, Maria, Potter, William, Barnes, Lisa, Bernard, Marie, Gonzalez, Hector, Ho, Carole, Hsiao, John K, Jackson, Jonathan, Masliah, Eliezer, Masterman, Donna, Okonkwo, Ozioma, Perrin, Richard, and Ryan, Laurie

Subjects
Medical Biochemistry and Metabolomics, Biomedical and Clinical Sciences, Neurosciences, Alzheimer's Disease including Alzheimer's Disease Related Dementias (AD/ADRD), Aging, Acquired Cognitive Impairment, Neurodegenerative, Networking and Information Technology R&D (NITRD), Machine Learning and Artificial Intelligence, Dementia, Alzheimer's Disease, Brain Disorders, Neurological, Humans, Aged, Magnetic Resonance Imaging, Deep Learning, Alzheimer Disease, Neuroimaging, Metabolome, Lipids, Cognitive Dysfunction, Alzheimer's Disease Metabolomics Consortium, Alzheimer's Disease Neuroimaging Initiative, Alzheimer's disease, Deep learning, Lipidomics, Machine learning, Metabolomics, Clinical Sciences, Public Health and Health Services, Clinical sciences, Epidemiology
Abstract

BackgroundDeep learning has shown potential in various scientific domains but faces challenges when applied to complex, high-dimensional multi-omics data. Alzheimer's Disease (AD) is a neurodegenerative disorder that lacks targeted therapeutic options. This study introduces the Circular-Sliding Window Association Test (c-SWAT) to improve the classification accuracy in predicting AD using serum-based metabolomics data, specifically lipidomics.MethodsThe c-SWAT methodology builds upon the existing Sliding Window Association Test (SWAT) and utilizes a three-step approach: feature correlation analysis, feature selection, and classification. Data from 997 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI) served as the basis for model training and validation. Feature correlations were analyzed using Weighted Gene Co-expression Network Analysis (WGCNA), and Convolutional Neural Networks (CNN) were employed for feature selection. Random Forest was used for the final classification.FindingsThe application of c-SWAT resulted in a classification accuracy of up to 80.8% and an AUC of 0.808 for distinguishing AD from cognitively normal older adults. This marks a 9.4% improvement in accuracy and a 0.169 increase in AUC compared to methods without c-SWAT. These results were statistically significant, with a p-value of 1.04 × 10ˆ-4. The approach also identified key lipids associated with AD, such as Cer(d16:1/22:0) and PI(37:6).InterpretationOur results indicate that c-SWAT is effective in improving classification accuracy and in identifying potential lipid biomarkers for AD. These identified lipids offer new avenues for understanding AD and warrant further investigation.FundingThe specific funding of this article is provided in the acknowledgements section.

Published
2023

42. A Quic(k) Security Overview: A Literature Research on Implemented Security Recommendations

Author

Tatschner, Stefan, Peters, Sebastian N., Emeis, David, Morris, John, and Newe, Thomas

Subjects
Computer Science - Cryptography and Security
Abstract

Built on top of UDP, the relatively new QUIC protocol serves as the baseline for modern web protocol stacks. Equipped with a rich feature set, the protocol is defined by a 151 pages strong IETF standard complemented by several additional documents. Enabling fast updates and feature iteration, most QUIC implementations are implemented as user space libraries leading to a large and fragmented ecosystem. This work addresses the research question, "if a complex standard with a large number of different implementations leads to an insecure ecosystem?". The relevant RFC documents were studied and "Security Consideration" items describing conceptional problems were extracted. During the research, 13 popular production ready QUIC implementations were compared by evaluating 10 security considerations from RFC9000. While related studies mostly focused on the functional part of QUIC, this study confirms that available QUIC implementations are not yet mature enough from a security point of view., Comment: to be published in: ARES 2023: The 18th International Conference on Availability, Reliability and Security, Benevento, Italy, August 2023

Published
2023
Full Text
View/download PDF

43. A data‐driven examination of apathy and depressive symptoms in dementia with independent replication

Author

Da Silva, Miguel Vasconcelos, Melendez‐Torres, Gerardo Javier, Ismail, Zahinoor, Testad, Ingelin, Ballard, Clive, Creese, Byron, Weiner, Michael, Aisen, Paul, Petersen, Ronald, Jack, Clifford R, Jagust, William, Trojanowki, John Q, Toga, Arthur W, Beckett, Laurel, Green, Robert C, Saykin, Andrew J, Morris, John, Shaw, Leslie M, Liu, Enchi, Montine, Tom, Thomas, Ronald G, Donohue, Michael, Walter, Sarah, Gessert, Devon, Sather, Tamie, Jiminez, Gus, Harvey, Danielle, Bernstein, Matthew, Fox, Nick, Thompson, Paul, Schuff, Norbert, DeCArli, Charles, Borowski, Bret, Gunter, Jeff, Senjem, Matt, Vemuri, Prashanthi, Jones, David, Kantarci, Kejal, Ward, Chad, Koeppe, Robert A, Foster, Norm, Reiman, Eric M, Chen, Kewei, Mathis, Chet, Landau, Susan, Cairns, Nigel J, Householder, Erin, Reinwald, Lisa Taylor, Lee, Virginia, Korecka, Magdalena, Figurski, Michal, Crawford, Karen, Neu, Scott, Foroud, Tatiana M, Potkin, Steven, Shen, Li, Kelley, Faber, Kim, Sungeun, Nho, Kwangsik, Kachaturian, Zaven, Frank, Richard, Snyder, Peter J, Molchan, Susan, Kaye, Jeffrey, Quinn, Joseph, Lind, Betty, Carter, Raina, Dolen, Sara, Schneider, Lon S, Pawluczyk, Sonia, Beccera, Mauricio, Teodoro, Liberty, Spann, Bryan M, Brewer, James, Van der Swag, Helen, Fleisher, Adam, Heidebrink, Judith L, Lord, Joanne L, Mason, Sara S, Albers, Colleen S, Knopman, David, Johnson, Kris, Doody, Rachelle S, Meyer, Javier Villanueva, Chowdhury, Munir, Rountree, Susan, Dang, Mimi, Stern, Yaakov, Honig, Lawrence S, Bell, Karen L, Ances, Beau, Morris, John C, Carroll, Maria, Leon, Sue, Mintun, Mark A, Schneider, Stacy, and Oliver, Angela

Subjects
Clinical and Health Psychology, Biomedical and Clinical Sciences, Psychology, Depression, Mental Health, Brain Disorders, Mental Illness, Neurological, Mental health, Alzheimer's Disease Neuroimaging Initiative, apathy, dementia, depression, latent class analysis, Genetics, Neurosciences, Biological psychology
Abstract

Apathy is one of the most common neuropsychiatric symptoms (NPS) and is associated with poor clinical outcomes. Research that helps define the apathy phenotype is urgently needed, particularly for clinical and biomarker studies. We used latent class analysis (LCA) with two independent cohorts to understand how apathy and depression symptoms co-occur statistically. We further explored the relationship between latent class membership, demographics, and the presence of other NPS. The LCA identified a four-class solution (no symptoms, apathy, depression, and combined apathy/depression), reproducible over both cohorts, providing robust support for an apathy syndrome distinct from depression and confirming that an apathy/depression syndrome exists, supported by the model fit test with the four-class solution scores evidencing better fitting (Bayesian information criterion adjusted and entropy R 2). Using a data-driven method, we show distinct and statistically meaningful co-occurrence of apathy and depressive symptoms. There was evidence that these classes have different clinical associations, which may help inform diagnostic categories for research studies and clinical practice.HighlightsWe found four classes: no symptoms, apathy, depression and apathy/depression.Apathy conferred a higher probability for agitation.Apathy diagnostic criteria should include accompanying neuropsychiatric symptoms.

Published
2023

44. Gene-SGAN: a method for discovering disease subtypes with imaging and genetic signatures via multi-view weakly-supervised deep clustering

Author

Yang, Zhijian, Wen, Junhao, Abdulkadir, Ahmed, Cui, Yuhan, Erus, Guray, Mamourian, Elizabeth, Melhem, Randa, Srinivasan, Dhivya, Govindarajan, Sindhuja T., Chen, Jiong, Habes, Mohamad, Masters, Colin L., Maruff, Paul, Fripp, Jurgen, Ferrucci, Luigi, Albert, Marilyn S., Johnson, Sterling C., Morris, John C., LaMontagne, Pamela, Marcus, Daniel S., Benzinger, Tammie L. S., Wolk, David A., Shen, Li, Bao, Jingxuan, Resnick, Susan M., Shou, Haochang, Nasrallah, Ilya M., and Davatzikos, Christos

Subjects
Quantitative Biology - Quantitative Methods, Computer Science - Machine Learning, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

Disease heterogeneity has been a critical challenge for precision diagnosis and treatment, especially in neurologic and neuropsychiatric diseases. Many diseases can display multiple distinct brain phenotypes across individuals, potentially reflecting disease subtypes that can be captured using MRI and machine learning methods. However, biological interpretability and treatment relevance are limited if the derived subtypes are not associated with genetic drivers or susceptibility factors. Herein, we describe Gene-SGAN - a multi-view, weakly-supervised deep clustering method - which dissects disease heterogeneity by jointly considering phenotypic and genetic data, thereby conferring genetic correlations to the disease subtypes and associated endophenotypic signatures. We first validate the generalizability, interpretability, and robustness of Gene-SGAN in semi-synthetic experiments. We then demonstrate its application to real multi-site datasets from 28,858 individuals, deriving subtypes of Alzheimer's disease and brain endophenotypes associated with hypertension, from MRI and SNP data. Derived brain phenotypes displayed significant differences in neuroanatomical patterns, genetic determinants, biological and clinical biomarkers, indicating potentially distinct underlying neuropathologic processes, genetic drivers, and susceptibility factors. Overall, Gene-SGAN is broadly applicable to disease subtyping and endophenotype discovery, and is herein tested on disease-related, genetically-driven neuroimaging phenotypes.

Published
2023

45. UCSF ChimeraX: Tools for Structure Building and Analysis

Author

Meng, Elaine C, Goddard, Thomas D, Pettersen, Eric F, Couch, Greg S, Pearson, Zach J, Morris, John H, and Ferrin, Thomas E

Subjects
Biological Sciences, Bioinformatics and Computational Biology, Networking and Information Technology R&D (NITRD), AlphaFold, ChimeraX, atomic model building, cryo-electron microscopy, protein structure prediction, refinement, Biochemistry and Cell Biology, Computation Theory and Mathematics, Other Information and Computing Sciences, Biophysics, Biochemistry and cell biology, Medicinal and biomolecular chemistry
Abstract

Advances in computational tools for atomic model building are leading to accurate models of large molecular assemblies seen in electron microscopy, often at challenging resolutions of 3-4 Å. We describe new methods in the UCSF ChimeraX molecular modeling package that take advantage of machine-learning structure predictions, provide likelihood-based fitting in maps, and compute per-residue scores to identify modeling errors. Additional model-building tools assist analysis of mutations, post-translational modifications, and interactions with ligands. We present the latest ChimeraX model-building capabilities, including several community-developed extensions. ChimeraX is available free of charge for noncommercial use at https://www.rbvi.ucsf.edu/chimerax. This article is protected by copyright. All rights reserved.

Published
2023

46. A Biomedical Open Knowledge Network Harnesses the Power of AI to Understand Deep Human Biology

Author

Baranzini, Sergio, Börner, Katy, Morris, John, Nelson, Charlotte, Soman, Karthik, Schleimer, Erica, Keiser, Michael, Musen, Mark, Pearce, Roger, Reza, Tahsin, Smith, Brett, Herr II, Bruce, Oskotsky, Boris, Rizk-Jackson, Angela, Rankin, Katherine, Sanders, Stephan, Bove, Riley, Rose, Peter, Israni, Sharat, and Huang, Sui

Subjects
Information and Computing Sciences, Artificial Intelligence, Neurodegenerative, Generic health relevance, Good Health and Well Being, Artificial Intelligence and Image Processing, Cognitive Sciences, Artificial Intelligence & Image Processing, Artificial intelligence, Machine learning
Abstract

Knowledge representation and reasoning (KR&R) has been successfully implemented in many fields to enable computers to solve complex problems with AI methods. However, its application to biomedicine has been lagging in part due to the daunting complexity of molecular and cellular pathways that govern human physiology and pathology. In this article, we describe concrete uses of Scalable PrecisiOn Medicine Knowledge Engine (SPOKE), an open knowledge network that connects curated information from thirty-seven specialized and human-curated databases into a single property graph, with 3 million nodes and 15 million edges to date. Applications discussed in this article include drug discovery, COVID-19 research and chronic disease diagnosis, and management.

Published
2023

47. Determinants of cognitive and brain resilience to tau pathology: a longitudinal analysis

Author

Bocancea, Diana I, Svenningsson, Anna L, van Loenhoud, Anna C, Groot, Colin, Barkhof, Frederik, Strandberg, Olof, Smith, Ruben, Weiner, Michael W, Aisen, Paul, Petersen, Ronald, Jack, Clifford R, Jagust, William, Trojanowki, John Q, Toga, Arthur W, Beckett, Laurel, Green, Robert C, Saykin, Andrew J, Morris, John C, Perrin, Richard J, Shaw, Leslie M, Khachaturian, Zaven, Carrillo, Maria, Potter, William, Barnes, Lisa, Bernard, Marie, González, Hector, Ho, Carole, Hsiao, John K, Jackson, Jonathan, Masliah, Eliezer, Masterman, Donna, Okonkwo, Ozioma, Ryan, Laurie, Silverberg, Nina, Fleisher, Adam, Sacrey, Diana Truran, Fockler, Juliet, Conti, Cat, Veitch, Dallas, Neuhaus, John, Jin, Chengshi, Nosheny, Rachel, Ashford, Miriam, Flenniken, Derek, Kormo, Adrienne, Montine, Tom, Conti, Cat B, Rafii, Michael, Raman, Rema, Jimenez, Gustavo, Donohue, Michael, Gessert, Devon, Salazar, Jennifer, Zimmerman, Caileigh, Cabrera, Yuliana, Walter, Sarah, Miller, Garrett, Coker, Godfrey, Clanton, Taylor, Hergesheimer, Lindsey, Smith, Stephanie, Adegoke, Olusegun, Mahboubi, Payam, Moore, Shelley, Pizzola, Jeremy, Shaffer, Elizabeth, Harvey, Danielle, Forghanian-Arani, Arvin, Borowski, Bret, Ward, Chad, Schwarz, Christopher, Jones, David, Gunter, Jeff, Kantarci, Kejal, Senjem, Matthew, Vemuri, Prashanthi, Reid, Robert, Fox, Nick C, Malone, Ian, Thompson, Paul, Thomopoulos, Sophia I, Nir, Talia M, Jahanshad, Neda, DeCarli, Charles, Knaack, Alexander, Fletcher, Evan, Tosun-Turgut, Duygu, Chen, Stephanie Rossi, Choe, Mark, and Crawfor, Karen

Subjects
Health Sciences, Dementia, Brain Disorders, Cerebrovascular, Aging, Neurosciences, Alzheimer's Disease Related Dementias (ADRD), Neurodegenerative, Biomedical Imaging, Vascular Cognitive Impairment/Dementia, Alzheimer's Disease including Alzheimer's Disease Related Dementias (AD/ADRD), Alzheimer's Disease, Behavioral and Social Science, Clinical Research, Acquired Cognitive Impairment, 2.1 Biological and endogenous factors, 1.1 Normal biological development and functioning, Mental health, Neurological, Humans, Alzheimer Disease, Longitudinal Studies, tau Proteins, Cross-Sectional Studies, Cerebral Cortical Thinning, Positron-Emission Tomography, Brain, Cognition, Apolipoproteins E, Alzheimer's disease, tau, resilience, cognition, PET, MRI, Alzheimer’s Disease Neuroimaging Initiative, Alzheimer’s disease, Medical and Health Sciences, Psychology and Cognitive Sciences, Neurology & Neurosurgery, Biomedical and clinical sciences, Health sciences, Psychology
Abstract

Mechanisms of resilience against tau pathology in individuals across the Alzheimer's disease spectrum are insufficiently understood. Longitudinal data are necessary to reveal which factors relate to preserved cognition (i.e. cognitive resilience) and brain structure (i.e. brain resilience) despite abundant tau pathology, and to clarify whether these associations are cross-sectional or longitudinal. We used a longitudinal study design to investigate the role of several demographic, biological and brain structural factors in yielding cognitive and brain resilience to tau pathology as measured with PET. In this multicentre study, we included 366 amyloid-β-positive individuals with mild cognitive impairment or Alzheimer's disease dementia with baseline 18F-flortaucipir-PET and longitudinal cognitive assessments. A subset (n = 200) additionally underwent longitudinal structural MRI. We used linear mixed-effects models with global cognition and cortical thickness as dependent variables to investigate determinants of cognitive resilience and brain resilience, respectively. Models assessed whether age, sex, years of education, APOE-ε4 status, intracranial volume (and cortical thickness for cognitive resilience models) modified the association of tau pathology with cognitive decline or cortical thinning. We found that the association between higher baseline tau-PET levels (quantified in a temporal meta-region of interest) and rate of cognitive decline (measured with repeated Mini-Mental State Examination) was adversely modified by older age (Stβinteraction = -0.062, P = 0.032), higher education level (Stβinteraction = -0.072, P = 0.011) and higher intracranial volume (Stβinteraction = -0.07, P = 0.016). Younger age, higher education and greater cortical thickness were associated with better cognitive performance at baseline. Greater cortical thickness was furthermore associated with slower cognitive decline independent of tau burden. Higher education also modified the negative impact of tau-PET on cortical thinning, while older age was associated with higher baseline cortical thickness and slower rate of cortical thinning independent of tau. Our analyses revealed no (cross-sectional or longitudinal) associations for sex and APOE-ε4 status on cognition and cortical thickness. In this longitudinal study of clinically impaired individuals with underlying Alzheimer's disease neuropathological changes, we identified education as the most robust determinant of both cognitive and brain resilience against tau pathology. The observed interaction with tau burden on cognitive decline suggests that education may be protective against cognitive decline and brain atrophy at lower levels of tau pathology, with a potential depletion of resilience resources with advancing pathology. Finally, we did not find major contributions of sex to brain nor cognitive resilience, suggesting that previous links between sex and resilience might be mainly driven by cross-sectional differences.

Published
2023

48. Evaluation of methods to assign cell type labels to cell clusters from single-cell RNA-sequencing data

Author

Diaz-Mejia, J Javier, Meng, Elaine C, Pico, Alexander R, MacParland, Sonya A, Ketela, Troy, Pugh, Trevor J, Bader, Gary D, and Morris, John H

Subjects
Networking and Information Technology R&D (NITRD), Genetics, Prevention, Good Health and Well Being, Biochemistry and Cell Biology, Clinical Sciences, Oncology and Carcinogenesis
Abstract

Background: Identification of cell type subpopulations from complex cell mixtures using single-cell RNA-sequencing (scRNA-seq) data includes automated computational steps like data normalization, dimensionality reduction and cell clustering. However, assigning cell type labels to cell clusters is still conducted manually by most researchers, resulting in limited documentation, low reproducibility and uncontrolled vocabularies. Two bottlenecks to automating this task are the scarcity of reference cell type gene expression signatures and the fact that some dedicated methods are available only as web servers with limited cell type gene expression signatures. Methods: In this study, we benchmarked four methods (CIBERSORT, GSEA, GSVA, and ORA) for the task of assigning cell type labels to cell clusters from scRNA-seq data. We used scRNA-seq datasets from liver, peripheral blood mononuclear cells and retinal neurons for which reference cell type gene expression signatures were available. Results: Our results show that, in general, all four methods show a high performance in the task as evaluated by receiver operating characteristic curve analysis (average area under the curve (AUC) = 0.94, sd = 0.036), whereas precision-recall curve analyses show a wide variation depending on the method and dataset (average AUC = 0.53, sd = 0.24). Conclusions: CIBERSORT and GSVA were the top two performers. Additionally, GSVA was the fastest of the four methods and was more robust in cell type gene expression signature subsampling simulations. We provide an extensible framework to evaluate other methods and datasets at https://github.com/jdime/scRNAseq_cell_cluster_labeling.

Published
2023

49. Cerebrospinal fluid proteomics define the natural history of autosomal dominant Alzheimer’s disease

Author

Johnson, Erik CB, Bian, Shijia, Haque, Rafi U, Carter, E Kathleen, Watson, Caroline M, Gordon, Brian A, Ping, Lingyan, Duong, Duc M, Epstein, Michael P, McDade, Eric, Barthélemy, Nicolas R, Karch, Celeste M, Xiong, Chengjie, Cruchaga, Carlos, Perrin, Richard J, Wingo, Aliza P, Wingo, Thomas S, Chhatwal, Jasmeer P, Day, Gregory S, Noble, James M, Berman, Sarah B, Martins, Ralph, Graff-Radford, Neill R, Schofield, Peter R, Ikeuchi, Takeshi, Mori, Hiroshi, Levin, Johannes, Farlow, Martin, Lah, James J, Haass, Christian, Jucker, Mathias, Morris, John C, Benzinger, Tammie LS, Roberts, Blaine R, Bateman, Randall J, Fagan, Anne M, Seyfried, Nicholas T, and Levey, Allan I

Subjects
Biomedical and Clinical Sciences, Health Sciences, Brain Disorders, Dementia, Alzheimer's Disease, Acquired Cognitive Impairment, Aging, Neurodegenerative, Biotechnology, Neurosciences, Alzheimer's Disease including Alzheimer's Disease Related Dementias (AD/ADRD), 2.1 Biological and endogenous factors, 4.2 Evaluation of markers and technologies, Neurological, Humans, Alzheimer Disease, Biomarkers, Proteomics, Male, Female, Adult, Middle Aged, Mutation, Age of Onset, Dominantly Inherited Alzheimer Network, Medical and Health Sciences, Immunology, Biomedical and clinical sciences, Health sciences
Abstract

Alzheimer's disease (AD) pathology develops many years before the onset of cognitive symptoms. Two pathological processes-aggregation of the amyloid-β (Aβ) peptide into plaques and the microtubule protein tau into neurofibrillary tangles (NFTs)-are hallmarks of the disease. However, other pathological brain processes are thought to be key disease mediators of Aβ plaque and NFT pathology. How these additional pathologies evolve over the course of the disease is currently unknown. Here we show that proteomic measurements in autosomal dominant AD cerebrospinal fluid (CSF) linked to brain protein coexpression can be used to characterize the evolution of AD pathology over a timescale spanning six decades. SMOC1 and SPON1 proteins associated with Aβ plaques were elevated in AD CSF nearly 30 years before the onset of symptoms, followed by changes in synaptic proteins, metabolic proteins, axonal proteins, inflammatory proteins and finally decreases in neurosecretory proteins. The proteome discriminated mutation carriers from noncarriers before symptom onset as well or better than Aβ and tau measures. Our results highlight the multifaceted landscape of AD pathophysiology and its temporal evolution. Such knowledge will be critical for developing precision therapeutic interventions and biomarkers for AD beyond those associated with Aβ and tau.

Published
2023

Catalog

Books, media, physical & digital resources
See catalog results