Neural correlates of pronoun processing: An activation likelihood estimation meta-analysis

Abstract

Pronouns are unique linguistic devices that allow for the expression of referential relationships. Despite their communicative utility, the neural correlates of the operations involved in reference assignment and/or resolution, are not well-understood. The present study synthesized the neuroimaging literature on pronoun processing to test extant theories of pronoun comprehension. Following the PRISMA guidelines and thebest-practice recommendations for neuroimaging meta-analyses, a systematic literature search and record assessment were performed. As a result, 16 fMRI studies were included in the meta-analysis, and were coded in Scribe 3.6 for inclusion in the BrainMap database. The activation coordinates for the contrasts of interest were transformed into Talairach space and submitted to an Activation Likelihood Estimation (ALE) meta-analysis in GingerALE 3.0.1. The results indicated that pronoun processing had functional convergence in the left posterior middle and superior temporal gyri, potentially reflecting the retrieval, prediction and integration roles of these areas for pronoun processing.

Keywords: Activation likelihood estimation; Language; Meta-analysis; Pronoun; Sentence processing; fMRI.

Figure 1.

Summary of the predictions of the three neurobiological models, showing regions that are unique to each model (green, purple and orange) and overlapping across models (blue and yellow).

Figure 2.

PRISMA flowchart summarizing the study selection process.

Figure 3.

Results of the P> B ALE meta-analysis. Notes: G: gyrus; IF: inferior frontal; L: left; MF: middle frontal; P: precentral; MT: middle temporal; ST: superior temporal

Figure 4.

Neural architecture and components of the proposed Retrieval and Integration model of pronoun processing. Notes: Activations that correspond to antecedent retrieval and predictive processing are color-coded in green (in L pMTG). Activations that correspond to antecedent integration are color-coded in blue (in L pSTG). The model accommodates activity in other neural regions such as left IFG, PP, and AG (in dashes) based on the specific linguistic context and task in which the pronoun needs to be comprehended. G: gyrus; A: angular; IF: inferior frontal; L: left; MF: middle frontal; P: precentral; pMT: posterior middle temporal; PP: posterior parietal; pST: Posterior superior.