William Matchin (UC Irvine) · [LingBrains] Exploring the syntax-brain relationship: two neurolinguistic experiments of syntax and sentence processing
In this talk I will present two fMRI experiments aimed at understanding the neural instantiation of syntactic operations. Experiment 1 addresses the leading hypotheses of the neural localization of syntactic operations in Broca's area by the use of a distance manipulation in backward anaphora (cataphora). Previous research had demonstrated a distance effect (increased activation for increased distance) between filler and gap in movement constructions in the anterior portion of Broca's area, but not for canonical (forward) anaphora, suggesting syntactic specificity of the response profile in this region. However, filler-gap dependencies are processed actively while canonical anaphoric dependencies are not, leaving open the possibility that the processing mechanism accounts for the activation, and not the construction. The results of Experiment 1 revealed a distance effect for backward anaphora in this region, showing that the nature of the processing mechanism (active vs. not) accounts for distance effects in this region, not the syntactic operation involved.
Experiment 2 is a novel approach to understanding syntax and sentence processing in the brain. Given the preponderance of neurolinguistic experiments involving sentence comprehension and syntactic complexity/violations, we investigated the neural correlates of syntactic processing through the use of sentence production and a 'syntactic perturbation' technique. Subjects were instructed to produce active or passive sentences contingent on visual cues. On 20% of trials, the subject's target construction (e.g., active) shifted mid-utterance to the other construction (e.g., passive). We examined the BOLD response of shift trials relative to non-shift trials and to trials in which subjects performed an analogous task without sentence structure (i.e., a list of words). The results revealed a network of regions previously implicated in non-linguistic 'go/no-go' tasks. Interestingly (but perhaps unsurprisingly), the response profile of Broca's area did not pattern with one expected to be involved in syntactic operations. The relationship between the activations obtained and mechanisms involved in sentence production will be discussed.