Elamur Research⌕
☰
home / papers / 2026 / Detecting Primary Progressive Aphasia (PPA) from Text: A Benchmarking Study

Detecting Primary Progressive Aphasia (PPA) from Text: A Benchmarking Study

Paper · 2026 · Ghofrane Merhbene, Fabian Lecron, Philippe Fortemps et al.

Summary

The paper tests how well different NLP approaches can identify primary progressive aphasia (PPA) subtypes from transcribed speech. The three subtypes are logopenic (lvPPA), semantic (svPPA) and nonfluent (nfvPPA), and controls are included as a separate class. This is a hard task because the subtypes share many linguistic markers, and labelled clinical data are scarce. The authors used anonymized interview transcripts first collected by Rezaii et al. (2022) and compared three families of methods:

  • classic classifiers built on a wide range of text representations
  • fine-tuned transformer models
  • prompted large language models

Performance was judged with balanced accuracy against a class-frequency-weighted random baseline. Transformers and LLMs did better than chance, but they did not clearly beat simpler pipelines. The strongest result came from an MLP trained on MentalBERT embeddings. LIME analyses linked influential tokens to known clinical speech patterns of each subtype. The study gives a reproducible comparison point for text-only PPA screening tools meant to support clinicians, not replace them.

Key points

  • Data: transcripts from 79 PPA patients and 53 controls, shared via the Open Science Framework. The data come in two versions: whole transcripts and a sentence-level split in which each sentence inherits its transcript's label.
  • Features: TF-IDF, BoW, Word2Vec, GloVe, FastText, embeddings from BERT, RoBERTa, MentalBERT and ClinicalBERT, 2–4-grams, LSA, LDA and dependency parsing.
  • Classifiers: SVM, Naive Bayes, Logistic Regression, MLP and XGBoost, all run with default hyperparameters to limit overfitting on the small dataset.
  • Best model: MLP with MentalBERT embeddings. LLMs were used only through expert-informed prompts, with no fine-tuning.
  • Limitations: the analysis uses text only, with no acoustic cues. There is no demographic metadata, no systematic hyperparameter tuning, and LLM runs were not repeated.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Harvard UniversityCambridge · United States112 works · 42,023 citations🏥Massachusetts General HospitalBoston · United States4 works · 5,421 citations🎓Bern University of Applied SciencesBern · Switzerland4 works · 154 citations🎓University of MonsMons · Belgium3 works · 67 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Psychology · 1,552 worksSocial Psychology · 306 worksMedicine · 241 worksHealth Professions · 69 worksHealth Information Management · 50 worksPsychiatry and Mental health · 9 works

Cite

Cite as Ghofrane Merhbene, Fabian Lecron, Philippe Fortemps, Bradford C. Dickerson, Mascha Kurpicz-Briki and Neguine Rezaii. 2026. Detecting Primary Progressive Aphasia (PPA) from Text: A Benchmarking Study. Findings of the Association for Computational Linguistics: EACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-eacl.19

@inproceedings{merhbene2026detecting,
  title = {{Detecting Primary Progressive Aphasia (PPA) from Text: A Benchmarking Study}},
  author = {Ghofrane Merhbene, Fabian Lecron, Philippe Fortemps, Bradford C. Dickerson, Mascha Kurpicz-Briki and Neguine Rezaii},
  booktitle = {Findings of the Association for Computational Linguistics: EACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-eacl.19},
  url = {https://doi.org/10.18653/v1/2026.findings-eacl.19}
}

References

43 references, 12 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Ankit Aich, Avery Quynh, Varsha Badal, Amy Pinkham, Philip Harvey, Colin Depp, and Natalie Parde. 2022. Towards intelligent clinically-informed language analyses of people with bipolar disorder and schizophrenia. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 2871–28… find
  2. Emily Alsentzer, John Murphy, William Boag, Wei-Hung Weng, Di Jindi, Tristan Naumann, and Matthew McDermott. 2019. Publicly available clinical BERT embeddings. In Proceedings of the 2nd Clinical Natural Language Processing Workshop, pages 72–78, Minneapolis, Minnesota, USA. Association for Computati…
  3. David M Blei, Andrew Y Ng, and Michael I Jordan. 2003. Latent dirichlet allocation. Journal of machine Learning research, 3:993–1022.
  4. Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2017. Enriching word vectors with subword information. Transactions of the Association for Computational Linguistics, 5:135–146.
  5. Peter F. Brown, Stephen A. Della Pietra, Vincent J. Della Pietra, and Robert L. Mercer. 1992. An estimate of an upper bound for the entropy of english. Computational Linguistics, 18(4):467–472.
  6. Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter,… — Language Models are Few-Shot Learners
  7. Heather M. Clark, Rene L. Utianski, Joseph R. Duffy, Edythe A. Strand, Herholtz Botha, Keith A. Josephs, and Jennifer L. Whitwell. 2020. Western aphasia battery-revised profiles in primary progressive aphasia and primary progressive apraxia of speech. find
  8. American Journal of Speech-Language Pathology, 29(1S):498–510. find
  9. Shan Cong, Hang Wang, Yang Zhou, Zheng Wang, Xiaohui Yao, and Chunsheng Yang. 2024a. Comprehensive review of transformer-based models in neuroscience, neurology, and psychiatry. Brain-X, 2(2):e57. find
  10. Yan Cong, Jiyeon Lee, and Arianna LaCroix. 2024b. Leveraging pre-trained large language models for aphasia detection in English and Chinese speakers. In Proceedings of the 6th Clinical Natural Language Processing Workshop, pages 238–245, Mexico City, Mexico. Association for Computational Linguistics… find
  11. Scott Deerwester, Susan T. Dumais, George W. Furnas, Thomas K. Landauer, and Richard Harshman. 1990. Indexing by latent semantic analysis. Journal of the American Society for Information Science, 41(6):391– 407. find
  12. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,…
  13. Miriam Faust, Elisheva Ben-Artzi, and Nili Vardi. 2012. Semantic processing in native and second language: Evidence from hemispheric differences in fine and coarse semantic coding. Brain and Language, 123(3):228–233. find
  14. Kathleen C. Fraser, Jed A. Meltzer, and Frank Rudzicz. 2014. Comparison of different feature sets for identification of variants in primary progressive aphasia. In Proceedings of the Workshop on Computational Linguistics and Clinical Psychology: From Linguistic Signal to Clinical Reality, pages 17–2… find
  15. Maria Luisa Gorno-Tempini, Argye E Hillis, Sandra Weintraub, Andrew Kertesz, Mario Mendez, Stefano F Cappa, Jennifer M Ogar, Jonathan D Rohrer, Sandra Black, Bradley F Boeve, et al. 2011. Classification of primary progressive aphasia and its variants. Neurology, 76(11):1006–1014. find
  16. Zellig S. Harris. 1954. Distributional structure. WORD, 10(2-3):146–162. find
  17. Harvard Health Publishing. 2022. Progressive aphasia involves many losses—here’s what you need to know. Accessed: 2025-02-11. find
  18. Maya L. Henry, Stephen M. Wilson, Mary C. Babiak, Maria L. Mandelli, Pélagie M. Beeson, Zoe A. Miller, and Maria Luisa Gorno-Tempini. 2016. Phonological processing in primary progressive aphasia. Journal of Cognitive Neuroscience, 28(2):210–222. find
  19. Sarah Itani, Fabian Lecron, and Philippe Fortemps. 2019. Specifics of medical data mining for diagnosis aid: A survey. Expert Systems with Applications, 118:300–314. find
  20. Shaoxiong Ji, Tianlin Zhang, Luna Ansari, Jie Fu, Prayag Tiwari, and Erik Cambria. 2022. Mental-BERT: Publicly available pretrained language models for mental healthcare. In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 7184–7190, Marseille, France. European Langu… find
  21. Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, Lélio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix,… find
  22. Sylvia Josephy-Hernandez, Negar Rezaii, Amanda Jones, Elise Loyer, David Hochberg, Meaghan Quimby, Bruce Wong, and Bradford C. Dickerson. 2023. Automated analysis of written language in the three variants of primary progressive aphasia. Brain Communications, 5(4):fcad202. find
  23. Eliyahu Kiperwasser and Yoav Goldberg. 2016. Simple and accurate dependency parsing using bidirectional lstm feature representations. Transactions of the Association for Computational Linguistics, 4:313–327.
  24. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. ArXiv, abs/1907.11692.
  25. Sladjana Lukic, Abigail E. Licata, Elizabeth Weis, Rian Bogley, Buddhika Ratnasiri, Ariane E. Welch, Leighton B. N. Hinkley, Z. Miller, Adolfo M. Garcia, John F. Houde, Srikantan S. Nagarajan, Maria Luisa Gorno-Tempini, and Valentina Borghesani. 2022. Auditory verb generation performance patterns di… find
  26. Tomas Mikolov, Kai Chen, Gregory S. Corrado, and Jeffrey Dean. 2013. Efficient estimation of word representations in vector space. In International Conference on Learning Representations.
  27. Kevin P. Murphy. 2012. Machine learning - a probabilistic perspective. In Adaptive computation and machine learning series. find
  28. OpenAI. 2023. Gpt-4o-mini: Advancing cost-efficient intelligence. https://openai.com/index/gpt-4 o-mini-advancing-cost-efficient-intellige nce/. Accessed: 2025-02-06. ↗
  29. Jeffrey Pennington, Richard Socher, and Christopher D. Manning. 2014. Glove: Global vectors for word representation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1532–1543. Association for Computational Linguistics.
  30. Neguine Rezaii, Daisy Hochberg, Megan Quimby, Bonnie Wong, Michael Brickhouse, Alexandra Tourouto-glou, Bradford C. Dickerson, and Phillip Wolff. 2024. Artificial intelligence classifies primary progressive aphasia from connected speech. Brain, 147(9):3070– 3082. find
  31. Neguine Rezaii, Kyle Mahowald, Rachel Ryskin, Bradford Dickerson, and Edward Gibson. 2022. A syntax-lexicon trade-off in language production. Proceedings of the National Academy of Sciences, 119(25):e2120203119. find
  32. Neguine Rezaii, Boyu Ren, Megan Quimby, Daisy Hochberg, and Bradford C. Dickerson. 2023. Less is more in language production: an information-theoretic analysis of agrammatism in primary progressive aphasia. Brain Communications, 5. find
  33. Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2016. "why should i trust you?": Explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’16, page 1135–1144, New York, NY, USA. Association for Co…
  34. C. Robertson, N. Rezaii, D. Hochberg, M. Quimby, P. Wolff, and B. C. Dickerson. 2024. Using explainable artificial intelligence to identify linguistic biomarkers of amyloid pathology in primary progressive aphasia. medRxiv. Preprint. find
  35. Gerard Salton and Christopher Buckley. 1988. Term-weighting approaches in automatic text retrieval. Information processing & management, 24(5):513– 523. find
  36. Yamanki Santander-Cruz, Sebastián Salazar-Colores, Wilfrido Jacobo Paredes-García, Humberto Guendulain-Arenas, and Saúl Tovar-Arriaga. 2022. Semantic feature extraction using sbert for dementia detection. Brain Sciences, 12(2):270. find
  37. Charalambos Themistocleous, Bronte Ficek, Kimberly Webster, Dirk-Bart den Ouden, Argye E. Hillis, and Kyrana Tsapkini. 2021. Automatic subtyping of individuals with primary progressive aphasia. Journal of Alzheimer’s Disease, 79(3):1185–1194. find
  38. Donna C. Tippett. 2020. Classification of primary progressive aphasia: challenges and complexities. F1000Research, 9:F1000 Faculty Rev–64. find
  39. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Awanish Batra, Daniel Haziza, et al. 2023. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2307.09288.
  40. Lisa Wauters, Karen Croot, Heather Dial, Joseph R. Duffy, Stephanie M Grasso, Esther Kim, Kristin M. Schaffer, Kirrie J. Ballard, Heather M. Clark, Leeah Kohley, Laura L. Murray, Emily Rogalskı, Mathieu Figeys, Lisa H. Milman, and Maya L. Henry. 2023. Behavioral treatment for speech and language in… find
  41. Stephen M. Wilson, Maya L. Henry, Max Besbris, Jennifer M. Ogar, Nina F. Dronkers, William Jarrold, Bruce L. Miller, and Maria Luisa Gorno-Tempini. 2010. Connected speech production in three variants of primary progressive aphasia. Brain, 133(7):2069– 2088. find
  42. Borui Yang, Md Afif Al Mamun, Jie M. Zhang, and Gias Uddin. 2025. Hallucination detection in large language models with metamorphic relations. Preprint, arXiv:2502.15844. — [arXiv:2502.15844](https://arxiv.org/abs/2502.15844) find
  43. Tianlin Zhang, Annika M Schoene, Shaoxiong Ji, and Sophia Ananiadou. 2022. Natural language processing applied to mental illness detection: a narrative review. NPJ digital medicine, 5(1):1–13. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/merhbene-2026-detecting-primary-progressive-aphasia-ppa/