Linking Symptom Inventories Using Semantic Textual Similarity

An extensive library of symptom inventories has been developed over time to measure clinical symptoms of traumatic brain injury (TBI), but this variety has led to several long-standing issues. Most notably, results drawn from different settings and studies are not comparable. This creates a fundamental problem in TBI diagnostics and outcome prediction, namely that it is not possible to equate results drawn from distinct tools and symptom inventories. Here, we present an approach using semantic textual similarity (STS) to link symptoms and scores across previously incongruous symptom inventories by ranking item text similarities according to their conceptual likeness. We tested the ability of four pretrained deep learning models to screen thousands of symptom description pairs for related content-a challenging task typically requiring expert panels. Models were tasked to predict symptom severity across four different inventories for 6,607 participants drawn from 16 international data sources. The STS approach achieved 74.8% accuracy across five tasks, outperforming other models tested. Correlation and factor analysis found the properties of the scales were broadly preserved under conversion. This work suggests that incorporating contextual, semantic information can assist expert decision-making processes, yielding broad gains for the harmonization of TBI assessment.

Authors

Institutions

Publication Details

Journal
Journal of Neurotrauma
Published
2025-04-09
DOI
https://doi.org/10.1089/neu.2024.0301
Citations
1
Primary Topic
Traumatic Brain Injury Research
Type
article
Field-Weighted Citation Impact
1.35
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Linking Symptom Inventories Using Semantic Textual Similarity

Mary R. Newsome, Michael Zeineh, Sarah L. Martindale, Kimbra Kenney et al.
1 citations
Journal of Neurotrauma
Traumatic Brain Injury Research
1.35
article

Linking Symptom Inventories Using Semantic Textual Similarity

Mary R. Newsome, Michael Zeineh, Sarah L. Martindale, Kimbra Kenney, Lisa Delano‐Wood, Jeffrey E. Max, Naomi J. Goodrich‐Hunsaker, Assaf Tal, Kristen R. Hoskinson, Elbert Geuze, Cooper B. Hodges, Ross Zafonte, Maria Carmela Tartaglia, Sophia I. Thomopoulos, Emily L. Dennis, Neda Jahanshad, Elizabeth S Hovenden, Andrew R. Mayer, Nicholas J. Pastorek, Carrie Esopenko, Inga Katharina Koerte, Benjamin Wade, Deleene S. Menefee, Nicholas P. Ryan, Brian Taylor, Eve M. Valera, Karen Caeyenberghs, David F. Tate, Helen M. Genova, Alexander Olsen, David Baron, Adeel Razi, Ruchira Menka Jha, Jaclyn A. Stephens, Ashley L. Ware, Eamonn Kennedy, Maya Troyanskaya, Mary Jo Pugh, Jordan Grafman, Martin M. Monti, Keith Owen Yeates, Hannah M. Lindsey, Asta Kristine Haberg, Timothy B. Meier, Abdalla Z. Mohamed, Andrei Irimia, Shashank Vadlamani, Ronak Agarwal, Ekaterina Dobryakova, Harm J. van der Horn, Jared A. Rowland, Virginia Newcombe, Abigail Livny, Spencer W. Liebel, Marianne Løvstad, Gershon Spitz, Kelly Peterson, William C. Walker, Leah D. Talbert, Frank Gerard Hillary, Finian Keleher, J. Kent Werner, Talin Babikian, Adam Schmidt, Stefania Mondello, Ragini Verma, Elisabeth A. Wilde, John D. Van Horn, Houshang Amiri, Kelly Russell, Brandon A. Zielinski, Seth G. Disner, Paul M. Thompson, Kristen Dams-O’Connor, Randall S. Scheibel, Rajendra A. Morey, R Andersen, Erin D. Bigler, Rachel M. Edelstein, Blessen C. Eapen, Jacob E. Resch
article en
1 citations

Abstract

An extensive library of symptom inventories has been developed over time to measure clinical symptoms of traumatic brain injury (TBI), but this variety has led to several long-standing issues. Most notably, results drawn from different settings and studies are not comparable. This creates a fundamental problem in TBI diagnostics and outcome prediction, namely that it is not possible to equate results drawn from distinct tools and symptom inventories. Here, we present an approach using semantic textual similarity (STS) to link symptoms and scores across previously incongruous symptom inventories by ranking item text similarities according to their conceptual likeness. We tested the ability of four pretrained deep learning models to screen thousands of symptom description pairs for related content-a challenging task typically requiring expert panels. Models were tasked to predict symptom severity across four different inventories for 6,607 participants drawn from 16 international data sources. The STS approach achieved 74.8% accuracy across five tasks, outperforming other models tested. Correlation and factor analysis found the properties of the scales were broadly preserved under conversion. This work suggests that incorporating contextual, semantic information can assist expert decision-making processes, yielding broad gains for the harmonization of TBI assessment.

Journal of NeurotraumaVol. 42(11-12)
Brigham Young University (US), Rutgers, The State University of New Jersey (US), Canadian Institute for Advanced Research (CA), Northwestern University (US), University of Messina (IT), University of Southern California (US), Texas Tech University (US), Brigham and Women's Hospital (US), Michael E. DeBakey VA Medical Center (US), University of Minnesota (US), Pennsylvania State University (US), Virginia Department of Health (US), Nationwide Children's Hospital (US), United States Department of Veterans Affairs (US), Shirley Ryan AbilityLab (US), Barrow Neurological Institute (US), University Health Network (CA), Norwegian Institute of Public Health (NO), Mind Research Network (US), LAC+USC Medical Center (US), The University of Texas MD Anderson Cancer Center (US), Harvard University (US), Uniformed Services University of the Health Sciences (US), Deakin University (AU), University of California, Los Angeles (US), Tel Aviv University (IL), Wellcome Centre for Human Neuroimaging (GB), Duke University (US), University of the Sunshine Coast (AU), Baylor College of Medicine (US), Georgia State University (US), Virginia Commonwealth University (US), University of Oslo (NO), University of Toronto (CA), Medical College of Wisconsin (US), Norwegian University of Science and Technology (NO), University of Utah (US), Kerman University of Medical Sciences (IR), University of Cambridge (GB), Kessler Foundation (US), Ben Taub Hospital (US), Spaulding Rehabilitation Hospital (US), Sheba Medical Center (IL), Sunnaas sykehus (NO), Walter Reed National Military Medical Center (US), Shepherd Center (US), Marina Del Rey Hospital (US), Alberta Children's Hospital (CA), Brain Mapping Foundation (US), University of Utah Hospital (US), University of Florida (US), University of California San Diego (US), Massachusetts General Hospital (US), VA Salt Lake City Healthcare System (US), Children's Hospital Research Institute of Manitoba (CA), Ministry of Defence (NL), VA San Diego Healthcare System (US), Cohen Veterans Bioscience (US), Sunshine Coast University Hospital (AU), St Olav's University Hospital (NO), Amsterdam Neuroscience (NL), Occupational Cancer Research Centre (CA), W. G. (Bill) Hefner VA Medical Center (US), Minneapolis VA Health Care System (US), Translational Therapeutics (United States) (US), Murdoch Children's Research Institute (AU), VA Greater Los Angeles Healthcare System (US), VA Mid-Atlantic Health Care Network (US), George E. Wahlen Department of VA Medical Center (US), Imaging Center (US), Rutgers New Jersey Medical School (US), University College London (GB), University of Manitoba (CA), Wake Forest University (US), University of Virginia (US), The Ohio State University (US), Monash University (AU), University of Pennsylvania (US), Ludwig-Maximilians-Universität München (DE), Penn State Milton S. Hershey Medical Center (US), Vrije Universiteit Amsterdam (NL), Colorado State University (US), Western University of Health Sciences (US), Stanford University (US), Icahn School of Medicine at Mount Sinai (US)
Openalex Percentile: Top 21%
Traumatic Brain Injury Research
1.35
DOISourceOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.