Kennedy Eamonn, Vadlamani Shashank, Lindsey Hannah M, Peterson Kelly S, Dams O'Connor Kristen, Agarwal Ronak, Amiri Houshang H, Andersen Raeda K, Babikian Talin, Baron David A, Bigler Erin D, Caeyenberghs Karen, Delano-Wood Lisa, Disner Seth G, Dobryakova Ekaterina, Eapen Blessen C, Edelstein Rachel M, Esopenko Carrie, Genova Helen M, Geuze Elbert, Goodrich-Hunsaker Naomi J, Grafman Jordan, Håberg Asta K, Hodges Cooper B, Hoskinson Kristen R, Hovenden Elizabeth S, Irimia Andrei, Jahanshad Neda, Jha Ruchira M, Keleher Finian, Kenney Kimbra, Koerte Inga K, Liebel Spencer W, Livny Abigail, Løvstad Marianne, Martindale Sarah L, Max Jeffrey E, Mayer Andrew R, Meier Timothy B, Menefee Deleene S, Mohamed Abdalla Z, Mondello Stefania, Monti Martin M, Morey Rajendra A, Newcombe Virginia, Newsome Mary R, Olsen Alexander, Pastorek Nicholas J, Pugh Mary Jo, Razi Adeel, Resch Jacob E, Rowland Jared A, Russell Kelly, Ryan Nicholas P, Scheibel Randall S, Schmidt Adam T, Spitz Gershon, Stephens Jaclyn A, Tal Assaf, Talbert Leah D, Tartaglia Maria Carmela, Taylor Brian A, Thomopoulos Sophia I, Troyanskaya Maya, Valera Eve M, van der Horn Harm Jan, Van Horn John D, Verma Ragini, Wade Benjamin S C, Walker Willian C, Ware Ashley L, Werner J Kent, Yeates Keith Owen, Zafonte Ross D, Zeineh Michael M, Zielinski Brandon, Thompson Paul M, Hillary Frank G, Tate David F, Wilde Elisabeth A, Dennis Emily L
Department of Neurology, University of Utah School of Medicine, Salt Lake City, Utah, USA.
George E. Wahlen Veterans Affairs Medical Center, Salt Lake City, Utah, USA.
J Neurotrauma. 2025 Jun;42(11-12):1008-1020. doi: 10.1089/neu.2024.0301. Epub 2025 Apr 9.
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.
随着时间的推移，已经开发出了大量的症状清单库来测量创伤性脑损伤（TBI）的临床症状，但这种多样性导致了几个长期存在的问题。最明显的是，来自不同环境和研究的结果无法进行比较。这在TBI诊断和结果预测中造成了一个根本性问题，即无法将来自不同工具和症状清单的结果等同起来。在此，我们提出一种使用语义文本相似度（STS）的方法，通过根据概念相似度对项目文本相似度进行排序，将以前不一致的症状清单中的症状和分数联系起来。我们测试了四个预训练的深度学习模型筛选数千对症状描述以查找相关内容的能力——这是一项通常需要专家小组的具有挑战性的任务。模型的任务是预测来自16个国际数据源的6607名参与者在四个不同清单中的症状严重程度。STS方法在五项任务中实现了74.8%的准确率，优于测试的其他模型。相关性和因子分析发现，量表的属性在转换后大致得以保留。这项工作表明，纳入上下文语义信息可以辅助专家决策过程，为TBI评估的协调带来广泛收益。
