Slide Slam

< Slide Slam Sessions

Slide Slam L9

Neural correlates of left and right branching

Slide Slam Session L, Thursday, October 7, 2021, 6:00 - 8:00 am PDT Log In to set Timezone

Simona Mancini1, Manuel Carreiras2, César Caballero-Gaudes3, Maite Termenon4, Pedro M. Paz Alonso5; 1BCBL, Basque Center on Cognition Brain and Language

In all human languages the building of hierarchical syntactic structures is thought to rely on the recursive application of a basic binary operation called merge1. While the underlying hierarchical structure created by Merge is constant, the superficial arrangement of elements in a sentence – i.e. linearization- varies cross-linguistically: in right-branching languages like Spanish the object follows the verb [Hemos comprado un libro/(we)Have bought a book], but in left-branching ones like Basque the object comes first [Liburua erosi dugu/book bought have (we)/(we)Have bought a book]. Several studies have investigated the neurobiological bases of merge and linearization operations in right-branching languages. Using functional magnetic resonance imaging (fMRI), Pallier and colleagues parametrically varied the size of syntactic constituents in a visual stream of 12 words and pseudo-words in French. While the pars orbitalis in the left inferior frontal gyrus (IFG) and the posterior portion of the superior temporal sulcus (pSTS) showed increased sensitivity to constituent size both in real- and pseudo-words, the anterior portion of the temporal lobe (TP) mainly responded to real-word stimuli. Further studies in German3 also indicated the pars opercularis of the left IFG as critically involved in basic syntactic structure building. Critically, it is still unclear whether Merge and linearization mechanisms recruit similar neuro-anatomical bases in left-branching languages such as Basque. The comparison between Spanish and Basque can therefore shed light onto this fundamental property of human languages. Fifty-one right-handed young adults, all simultaneous Spanish-Basque bilinguals, participated in an fMRI experiment. Participants were exposed to blocks of 24 visually presented real- and pseudo-words in Spanish and Basque. Similarly to Pallier and colleagues, we manipulated the size of the syntactic phrase(s) by merging pseudo- and real words parametrically: from a determiner phrase [Level1: determiner+noun (Spanish)/noun+determiner (Basque)] to a verb phrase [Level 2: non-finite verb+determiner phrase /determiner phrase+non-finite verb] to a full-fledged sentence [Level3: auxiliary+verb+determiner phrase/ determiner phrase+verb+auxiliary] Results reveal a main effect of syntactic Level: as the stimuli become more sentence-like, a significant increase in activation emerges (Level1<Level2<Level3) in the left IFG (opercularis/triangularis), pSTS, and TP. A dissociation within the left fronto-temporo-parietal network is also found, with left IFG, superior parietal and pSTS regions showing phrase-level sensitivity for both real- and pseudo-word conditions in Spanish and Basque, as opposed to anterior temporal regions being specifically responsive to real-word conditions. A significant Language x Level interaction is observed in the pars orbitalis: while greater activation for Level3 stimuli compared to Level2 and 1 is found in Basque, the reverse pattern arises in Spanish (Level1>Level2>Level3). These results suggest that the neural substrates supporting left branching largely overlap with those supporting right branching in Spanish (and other languages, see2,3). The two linear orders differ, however, in how they engage left IFG regions generally associated with controlled retrieval of semantic information. Further functional connectivity analyses are underway to better characterize the interaction between different syntactic levels and languages in real- and pseudo-word conditions among brain networks. 1 Chomsky (1995) 2 Pallier et al. (2011, PNAS) 3see review in Zaccarella et al.(2017 Neurosci Biobehav Rev.)

< Slide Slam Sessions

SNL Account Login

Forgot Password?
Create an Account