The power of neural oscillations to inform sentence comprehension: A linguistic perspective

Abstract

The field of psycholinguistics is currently experiencing an explosion of interest in the analysis of neural oscillations—rhythmic brain activity synchronized at different temporal and spatial levels. Given that language comprehension relies on a myriad of processes, which are carried out in parallel in distributed brain networks, there is hope that this methodology might bring the field closer to understanding some of the more basic (spatially and temporally distributed, yet at the same time often overlapping) neural computations that support language function. In this review, we discuss existing proposals linking oscillatory dynamics in different frequency bands to basic neural computations and review relevant theories suggesting associations between band‐specific oscillations and higher‐level cognitive processes. More or less consistent patterns of oscillatory activity related to certain types of linguistic processing can already be derived from the evidence that has accumulated over the past few decades. The centerpiece of the current review is a synthesis of such patterns grouped by linguistic phenomenon. We restrict our review to evidence linking measures of oscillatory power to the comprehension of sentences, as well as linguistically (and/or pragmatically) more complex structures. For each grouping, we provide a brief summary and a table of associated oscillatory signatures that a psycholinguist might expect to find when employing a particular linguistic task. Summarizing across different paradigms, we conclude that a handful of basic neural oscillatory mechanisms are likely recruited in different ways and at different times for carrying out a variety of linguistic computations.

Sentence processing: How words generate syntactic structures in the brain

“…This is also the case for the directionality of the reported theta-band effects, which were found in two studies testing number [59] and person agreement violations [60] as power decreases and increases, respectively. There are several possible interpretations for these consistent effects in the alpha-beta frequency range (for further discussion, see [54]). On the one hand, they could either exclusively [48] or partially [51] reflect processes related to the interruption of syntactic processing.…”

“…Although this might suggest that alpha-(rather than beta-)band oscillations are specifically involved in syntactic prediction (or the preactivation of syntactic categories in general) there is also some evidence for semantic prediction effects prior to the critical word in the alpha band [76]. Likewise, studies using violation paradigms have reported modulations in the alpha and/or beta bands triggered by both (morpho-)syntactic and semantic errors (see [54]). Taken together, these oscillatory similarities between syntactic and semantic manipulations suggest that alpha-beta effects could reflect a more general mechanism related to sentence-level meaning [54], which could explain why they were observed in sentences but not in word lists lacking any structured meaning [47][48][49][50].…”

“…Following the common ERP approach, these studies (see below) manipulated the presence of words with different types of (morpho)syntactic violations, generally exploring the time-frequency effects (i.e., power changes over time) elicited by these grammatically incorrect words. Prystauka and Lewis [54] recently provided a thorough review of studies reporting timefrequency power analyses in the context of sentence processing, including those implementing (morpho)syntactic violation paradigms. Here we will just briefly review the most representative findings obtained from this paradigm (for further studies and a more extended discussion, see [54]).…”

“…Prystauka and Lewis [54] recently provided a thorough review of studies reporting timefrequency power analyses in the context of sentence processing, including those implementing (morpho)syntactic violation paradigms. Here we will just briefly review the most representative findings obtained from this paradigm (for further studies and a more extended discussion, see [54]).…”

“…Apart from delta-band oscillations, frequency effects in the alpha, beta and high-gamma band have been shown to be modulated during syntactic processing (see Section 4). Although these effects could be related to structure-based operations (e.g., [48]), they might alternatively be involved in additional components relevant for sentence processing [45,51,54]. For example, alpha-band oscillations could play a role in the storage of syntactic information in working memory [45,66].…”