Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Elements of the neurobiology of language: a systematic review of 220 neurostimulation studies and an open database for science and surgery
Poster Session E, Friday, October 2, 11:00 am - 1:00 pm, Wangari Maathai
T R Williamson1,2, Gesa Hartwigsen3,4, Kris Kinsey1,5, Neil U Barua1,2, Naomi Heffer5, Philipp Kuhnke3,4, Sonia Mariotti1,2, Eimear McKnight1,6, Lydia Wiernik1,7,8, Jemma Sedgemond5, Antonia Vogt9, Anna E Piasecki1,2,5; 1Brain, Language, and Behaviour Laboratory, UWE Bristol, UK, 2Southmead Hospital, North Bristol NHS Trust, UK, 3Wilhelm Wundt Institute for Psychology, Leipzig University, Germany, 4Cognition and Plasticity, Max Planck Institute for Human Cognitive and Brain Sciences, Germany, 5School of Social Sciences, UWE Bristol, UK, 6Dept. of Linguistics, Queen Mary University of London, UK, 7SignLab, Dept. of German Philology, University of Göttingen, Germany, 8Neuropsychology, Max Planck Institute for Human Cognitive and Brain Sciences, Germany, 9Dept. of Medicine, University of Cambridge, UK
Introduction. Models of the neurobiology of language predominantly rely on correlational neuroimaging and broad constructs inherited from linguistic theory (e.g., semantics, syntax) which often fail to map cleanly onto neural architecture (Fedorenko et al., 2024; Reilly et al., 2025). The brain's cytoarchitecture demarcates functionally relevant regions at far greater precision (Palomero-Gallagher & Zilles, 2019), and neurostimulation — TMS, transcranial electrical stimulation (tES), and direct electrical stimulation (DES) — provides the capacity to causally perturb specific linguistic processes. This advantage has not been systematically exploited to construct a comprehensive causal model of language in the brain. Methods. We present elementalism: a neuromodulation-driven framework grounded in high process specificity and whole-brain inference. We conducted a preregistered systematic review (PROSPERO CRD42024602006) across PubMed, Scopus, Embase, and PsychInfo, returning 12,763 records; 220 papers (1999–2025) met inclusion criteria. Each significant outcome was qualitatively analysed (Braun & Clarke, 2006) to infer the most specific linguistic subprocess causally modulated — termed an element — and organised within a seven-level multidimensional hierarchy of process specificity, from discipline-level descriptors (e.g., morphosyntax) to fine-grained activities (e.g., non-adjacent dependency processing, articulatory control of specific speech units). Results. The 220 papers reported 608 significant outcomes across 38 languages and approximately 6,300 participants, distributed across TMS (115 papers, 311 outcomes), tES (78 papers, 173 outcomes), and DES (18 papers, 113 outcomes). Traditional linguistic constructs were frequently recoverable — semantics: 111 papers / 351 outcomes; phonology: 52 / 188; syntax: 29 / 101; phonetics: 14 / 93; pragmatics: 9 / 54; language acquisition: 15 / 46; morphology: 1 / 9 — but rarely delineated discrete functional territories. Bottom-up clustering of coordinate-level evidence yields region-specific element inventories that cross-cut traditional categories. The IFG (74 papers, N = 2,036) supports 8 elements including communicative output planning, competing candidate suppression, hierarchical syntactic assembly, sublexical phonological computation, and prosody-driven structural interpretation. Comparable inventories were derived for STG (5 elements), MTG (4), SMG (7), and MFG (5) — 29 elements across the five regions, drawn from 185 paper-region pairings. The construct semantics alone returned eight elements across 117 papers (N = 3,239), including semantic-to-lexical mapping (42 studies), semantic feature knowledge retrieval (35), and controlled semantic selection (23). Conclusion. Elementalism reconciles older modelling traditions by combining causal neuromodulation evidence, task heterogeneity, and whole-brain generalisation. It addresses three limitations of current models: that broad linguistic terms overgeneralise across functionally heterogeneous regions; that single-laboratory or single-paradigm evidence under-supports cross-brain inference; and that correlational tools cannot adjudicate causal contribution (Logothetis, 2008). The output of this work constitutes the largest ever review of neurostimulation and the most specific model of language in the brain to-date. The accompanying open database (language-elements.org) provides searchable, coordinate-resolved access to all 608 outcomes, with bottom-up element profiles per region and AI-assisted construct grouping; it is built to extend continuously as new neurostimulation evidence accumulates. Implications follow for theory-building in language neuroscience and for evidence-based intraoperative language-mapping protocols (De Witt Hamer et al., 2012).
Topic Areas: Computational Approaches, Development of Resources, Software, Educational Materials, etc.