Show simple item record

dc.contributor.authorKauf, Carina
dc.contributor.authorTuckute, Greta
dc.contributor.authorLevy, Roger
dc.contributor.authorAndreas, Jacob
dc.contributor.authorFedorenko, Evelina
dc.date.accessioned2024-02-12T21:19:48Z
dc.date.available2024-02-12T21:19:48Z
dc.date.issued2023-09-21
dc.identifier.issn2641-4368
dc.identifier.urihttps://hdl.handle.net/1721.1/153506
dc.description.abstractRepresentations from artificial neural network (ANN) language models have been shown to predict human brain activity in the language network. To understand what aspects of linguistic stimuli contribute to ANN-to-brain similarity, we used an fMRI data set of responses to n = 627 naturalistic English sentences (Pereira et al., 2018) and systematically manipulated the stimuli for which ANN representations were extracted. In particular, we (i) perturbed sentences’ word order, (ii) removed different subsets of words, or (iii) replaced sentences with other sentences of varying semantic similarity. We found that the lexical-semantic content of the sentence (largely carried by content words) rather than the sentence’s syntactic form (conveyed via word order or function words) is primarily responsible for the ANN-to-brain similarity. In follow-up analyses, we found that perturbation manipulations that adversely affect brain predictivity also lead to more divergent representations in the ANN’s embedding space and decrease the ANN’s ability to predict upcoming tokens in those stimuli. Further, results are robust as to whether the mapping model is trained on intact or perturbed stimuli and whether the ANN sentence representations are conditioned on the same linguistic context that humans saw. The critical result—that lexical-semantic content is the main contributor to the similarity between ANN representations and neural ones—aligns with the idea that the goal of the human language system is to extract meaning from linguistic strings. Finally, this work highlights the strength of systematic experimental manipulations for evaluating how close we are to accurate and generalizable models of the human language network.en_US
dc.language.isoen_US
dc.publisherMIT Pressen_US
dc.relation.isversionof10.1162/nol_a_00116en_US
dc.rightsCreative Commons Attributionen_US
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/en_US
dc.sourceThe MIT Pressen_US
dc.subjectNeurologyen_US
dc.subjectLinguistics and Languageen_US
dc.titleLexical-Semantic Content, Not Syntactic Structure, Is the Main Contributor to ANN-Brain Similarity of fMRI Responses in the Language Networken_US
dc.typeArticleen_US
dc.identifier.citationKauf, C., Tuckute, G., Levy, R., Andreas, J., & Fedorenko, E. (2023).Lexical-semantic content, not syntactic structure, is the main contributor to ANN-brain similarity of fMRI responses in the language network. Neurobiology of Language. Advance publication.en_US
dc.relation.journalNeurobiology of Languageen_US
dc.eprint.versionFinal published versionen_US
dc.type.urihttp://purl.org/eprint/type/JournalArticleen_US
eprint.statushttp://purl.org/eprint/status/PeerRevieweden_US
dspace.date.submission2024-02-12T21:16:29Z
mit.licensePUBLISHER_CC
mit.metadata.statusAuthority Work and Publication Information Neededen_US


Files in this item

Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record