One Way or Another: Cortical Language Areas Flexibly Adapt Processing Strategies to Perceptual And Contextual Properties of Speech

Abstract

Cortical circuits rely on the temporal regularities of speech to optimize signal parsing for sound-to-meaning mapping. Bottom-up speech analysis is accelerated by top–down predictions about upcoming words. In everyday communications, however, listeners are regularly presented with challenging input—fluctuations of speech rate or semantic content. In this study, we asked how reducing speech temporal regularity affects its processing—parsing, phonological analysis, and ability to generate context-based predictions. To ensure that spoken sentences were natural and approximated semantic constraints of spontaneous speech we built a neural network to select stimuli from large corpora. We analyzed brain activity recorded with magnetoencephalography during sentence listening using evoked responses, speech-to-brain synchronization and representational similarity analysis. For normal speech theta band (6.5–8 Hz) speech-to-brain synchronization was increased and the left fronto-temporal areas generated stronger contextual predictions. The reverse was true for temporally irregular speech—weaker theta synchronization and reduced top–down effects. Interestingly, delta-band (0.5 Hz) speech tracking was greater when contextual/semantic predictions were lower or if speech was temporally jittered. We conclude that speech temporal regularity is relevant for (theta) syllabic tracking and robust semantic predictions while the joint support of temporal and contextual predictability reduces word and phrase-level cortical tracking (delta).

Increased top-down semantic processing in natural speech linked to better reading in dyslexia

“…Although, as expected, performance in reading words and pseudowords was related, the strength of this relation differed between groups: the correlation between Word and Pseudoword reading in the dyslexic sample was weaker ( r = 0.35), compared to a stronger relationship ( r = 0.6) in controls. We ran a hierarchical multiple regression analysis with GFP Energy (area under the curve) for Envelope and Surprisal as dependant variables and Word (WR), Pseudoword (PS) reading, non-verbal IQ (Matrix test) and Age as predictors in each hemisphere separately and selecting theoretically relevant fronto-temporal sensors ( Klimovich-Gray et al, 2021 ). To understand how dyslexic readers differed from controls we asked if adding a group (control = 0, dyslexic = 1) regressor as well as interactions between group, word and pseudoword reading would significantly improve the model fit (F change).…”

“…The 160 unmodified sentences had a varied degree of constraint with respect to target nouns embedded in them (target noun position in sentence M = 9.3, SD = 3, for full details of stimuli design see Klimovich-Gray et al, 2021 ), resulting in 80 more and 80 less constraining sentences. The degree of constraint was measured as the Surprisal (negative log2 of the probability -Hale, 2016 ) associated with the target word (all nouns) embedded in the sentence, given preceding words.…”

“What” and “when” predictions jointly modulate speech processing

“…This entrainment, also known as phase locking, is reflected in phasically increased neural activity in regions involved in auditory processing (Obleser and Kayser, 2019) including the STG (Keitel et al, 2017), although it is debated to what degree increased phase-locking reflects a true oscillatory mechanism rather than a series of evoked responses (Doelling et al, 2019;Oganian et al, 2023). In contrast, jittered speech sequences, which lack a consistent temporal pattern, exhibit weaker neural entrainment (Klimovich-Gray et al, 2021). Research suggests that the brain's ability to entrain to rhythmic patterns plays a role in processing and understanding speech, particularly in situations where the temporal structure of the speech signal is predictable (Kösem et al, 2018;Riecke et al, 2018;Zoefel et al, 2018).…”

Early language experience modulates the cortical tracking of speech

“…In monolingual children, Panda et al (2020) showed that lexico-semantic (vocabulary) knowledge was linked to the synchronization between language-related brain areas through cortical oscillatory activity during continuous speech listening, although they did not assess CTS. With respect to CTS, it has been shown that in adults it is modulated by both context-driven word predictability (Broderick et al, 2021; Klimovich-Gray et al, 2021; Koskinen et al, 2020; Molinaro et al, 2021) and the syntactic structuring of speech (Kaufeld et al, 2020; Meyer et al, 2017; Meyer & Gumbert, 2018). Of particular relevance for our hypotheses, these studies converge on the finding that the predictability of lexico-semantic and syntactic information shapes CTS in adults, possibly as they have developed efficient neurocognitive language models as a function of accumulated language experience.…”

“…Lexico-semantic knowledge, whose development is tightly related to accumulated language experience during childhood (Carbajal & Peperkamp, 2020; Oller et al, 2007; Paradis & Genesee, 1996; Pearson et al, 1997), has been linked to the efficiency of cortical oscillatory mechanisms for speech processing. Recent studies have shown that, in adults, lexico-semantic CTS modulated by both context-driven word predictability (Broderick et al, 2021; Klimovich-Gray et al, 2021; Koskinen et al, 2020; Molinaro et al, 2021). Of particular relevance for our hypotheses, these studies converge on the finding that the predictability of lexico-semantic information shapes CTS in adults, possibly as they have developed efficient neurocognitive language models as a function of accumulated language experience.…”