Search

ABSTRACTNeural entrainment to the low-frequency modulations of speech might contribute significantly to reading acquisition. Still, no previous study has actually attempted to establish a longitudinal link between them. The present study tested Basque-speaking children twice: once before reading was formally instructed (t1; 5–6 years old) and once after they had received a full school year of reading instruction (t2; 6–7 years old). At t1, speech-brain coherence was recorded via EEG. At t2, in addition to the coherence measure, reading performance was assessed. Our results show that children with larger pre-reading delta-band (<1 Hz) speech-brain coherence at right sites of the scalp performed better in the reading tasks one year later. Overall, our results provide preliminary support for a relevant contribution of right-hemisphere speech-brain coherence to successful reading development and point towards pre-reading neural coherence indexes as useful tools for the early detection of developmental reading disorders.

Speech comprehension has been proposed to critically rely on oscillatory cortical tracking, that is, phase alignment of neural oscillations to the slow temporal modulations (envelope) of speech. Speech-brain entrainment is readjusted over time as transient events (edges) in speech lead to speech-brain phase realignment. Auditory behavioral research suggests that phonological deficits in dyslexia are linked to difficulty in discriminating speech edges. Importantly, research to date has not specifically examined neural responses to speech edges in dyslexia. In the present study, we used MEG to record brain activity from normal and dyslexic readers while they listened to speech. We computed phase locking values (PLVs) to evaluate phase entrainment between neural oscillations and the speech envelope time-locked to edge onsets. In both groups, we observed that edge onsets induced phase resets in the auditory oscillations tracking speech, thereby enhancing their entrainment to speech. Importantly, dyslexic readers showed weaker PLVs compared to normal readers in left auditory regions from ~.15 sec to ~.65 sec after edge onset. Our results indicate that the neural mechanism that adapts cortical entrainment to the speech envelope is impaired in dyslexia. These findings here are consistent with the temporal sampling theory of developmental dyslexia.

Numbers and letters are culturally created symbols which are learned through repeated training. This experience leads to a functional specialization of the perceptual system of our brain. Recent evidence suggests a neural dissociation between these two symbols. While previous literature has shown that letters elicit a left lateralized neural response, new studies suggest that numbers elicit preferentially a bilateral or right lateralized response. However, the time course of the neural patterns that characterize this dissociation is still underspecified. In the present study, we investigated with magnetoencephalography (MEG) the spatio-temporal dynamics of the neural response generated by numbers, letters and perceptually matched false fonts presented visually. Twenty-five healthy adults were recorded while participants performed a dot detection task. By including two experiments, we were able to study the effects of single characters as well as those of strings of characters. The signal analysis was focused on the event related fields (ERF) of the MEG signal in the sensors and in the source space. The main results of our study showed an early (<200 msec) preferential dissociation between single numbers and single letters on occipito-temporal sensors. When comparing strings of numbers and pseudowords, they differed also over prefrontal regions of the brain. These data offer a new example of acquired category-specific responses in the human brain.

Agreement is a syntactic relation involving a controller (e.g., a noun) and a target with matching inflectional morphology (e.g., a verb). Across languages, electrophysiological studies consistently report that the presence of a mismatch yields late positive effects (P600), often preceded by early negativities. The current study focuses on person agreement in Basque to investigate whether online processing routines are modulated by the relative semantic prominence of nominal and verbal person features. In an event-related potentials experiment in Basque, we manipulated the semantic markedness of nominal and verbal person features, creating 1st (marked) and 3rd (unmarked) person correct and incorrect sentences [Japoniarrok1pleuskara ikasi dugu1pl/*dute3plgustora(We Japanese have1pl/*3pllearn Basque with pleasure); Japoniarrek3pleuskara ikasi dute3pl/*dugu1plgustora (The Japanese have3pl/*1pllearnt Basque with pleasure)]. Both mismatches elicited an N400 effect, but only marked 1st person mismatches (Japoniarrok1pl… *dute3.pl) generated a P600, suggesting that (i) mismatches with unmarked 3rd person subjects (Japoniarrek3.pl… *dugu1.pl) are not treated as outright violations; (ii) the emergence of late positive effects is sensitive to fine-grained discourse information. Overall, these results call for a revision in the analysis of agreement relations from a theoretical and a processing perspective.

Despite the growing literature on anticipatory language processing, the brain dynamics of this high-level predictive process are still unclear. In the present MEG study, we analyzed pre- and post-stimulus oscillatory activity time-locked to the reading of a target word. We experimentally contrasted the processing of the same target word following two highly constraining sentence contexts, in which the constraint was driven either by the semantic content or by the lexical association between words. Previous research suggests the presence of sensory facilitation for expected words in the latter condition but not in the former. We observed a dissociation between beta (∼20 Hz) and gamma (>50 Hz) band activity in pre- and post-stimulus time intervals respectively. Both the beta and gamma effects were evident in occipital brain regions, and only the pre-stimulus beta effect additionally involved left pre-articulatory motor regions. Lexically constrained (vs. semantically constrained) words elicited reduced beta power around 400 msec before the target word in motor regions and a functionally related gamma enhancement in occipital regions around 200 msec post-target. The present findings highlight the role of the motor network in word-form prediction and support proposals claiming that low-level perceptual representations can be pre-activated during language prediction.

It has been proposed that atypical neural oscillations in both the auditory and the visual modalities could explain why some individuals fail to learn to read and suffer from developmental dyslexia. However, the role of specific oscillatory mechanisms in reading acquisition is still under debate. In this article, we take a cross-linguistic approach and argue that both the phonological and orthographic specifics of a language (e.g., linguistic rhythm, orthographic depth) shape the oscillatory activity thought to contribute to reading development. The proposed theoretical framework should allow future research to test cross-linguistic hypotheses that will shed light on the heterogeneity of auditory and visual disorders and their underlying brain dysfunction(s) in developmental dyslexia, and inform clinical practice by helping us to diagnose dyslexia across languages.

Sentence comprehension is successfully accomplished by means of a form-to-meaning mapping procedure that relies on the extraction of morphosyntactic information from the input and its mapping to higher-level semantic–discourse representations. In this study, we sought to determine whether neuroanatomically distinct brain regions are involved in the processing of different types of information contained in the propositional meaning of a sentence, namely person and number. While person information indexes the role that an individual has in discourse (i.e., the speaker, the addressee or the entity being talked about by speaker and addressee), number indicates its cardinality (i.e., a single entity vsa multitude of entities). An event-related functional magnetic resonance imaging (fMRI) experiment was run using agreement-Correct and Person- and Number-violated sentences in Spanish, to disentangle the processing mechanisms and neural substrates associated with the building of discourse and cardinality representations. The contrast between Person and Number Violations showed qualitativeand quantitativedifferences. A greater response for person compared to number was found in the left middle temporal gyrus (LMTG). However, critically, a posterior-to-anterior functional gradient emerged within this region. While the posterior portion of the LMTG was sensitive to both Person and Number Violations, the anterior portion of this region showed selective response for Person Violations. These results confirm that the comprehension of the propositional meaning of a sentence results from a composite, feature-sensitive mechanism of form-to-meaning mapping in which the nodes of the language network are differentially involved.

In the domain of written sentence comprehension, the computation of agreement dependencies is generally considered as a form-driven processing routine whose domain is syntactic in nature. In the present review we discuss the main findings emerging in the Event-Related Potential (ERP) literature on sentence comprehension, focusing on the different dimensions of agreement patterns (features, values, constituents involved and language): Agreement mismatches usually evoke a biphasic electrophysiological pattern (Left Anterior Negativity – LAN, 300–450msec and P600 after 500msec). This ERP pattern is assumed to reflect rule-based computations sensitive to formal (inflectional) covariations of related words (trigger–target). Here we claim that agreement processing is sensitive to both the type of feature involved and the constituents that express the agreement dependency. More specifically, LAN could reflect violation of expectancy(elicited by the trigger) for the target functional morphology; later, trigger and target are structurally integrated at the sentence level (earlyP600). However, morphosyntactic information could trigger the activation of higher-level representations that are not strictly syntactic in nature. The recruitment of this additional non-syntactic information (mirrored by N400-like effects) indicates that rule-based computations of agreement dependencies are not blind to non-syntactic information but are often recruited to establish sentence-level relations.

Electrical scalp recordings revealed the brain's sensitivity to both lexical properties of words and their contextual fit with a previous sentence context around 400 ms after word presentation. The so-called N400 component has been suggested to reflect the cost either of target word recognition or of a postlexical process for integrating word meaning into a context. In a sentence comprehension study, we manipulated the potential interference exerted in visual word recognition by target words' orthographic neighbors and the semantic constraints induced by the context in one and the same experiment. Neighbor frequency modulated the N400 only in low-constraint contexts; in high-constraint contexts the largely suppressed N400 did not show this neighbor interference effect. Furthermore, the earlier onset of the ERP effect (about 100 ms) induced by the contextual manipulation compared to the neighbor manipulation suggests distinct neurocognitive processes affecting the N400 component in an interactive manner.

We examined the electrophysiological correlates of one of the most influential orthographic effects the transposed-letter-masked priming effect. Transposed-letter nonword–word pairs (‘jugde–judge’), as well as transposed-letter word–word pairs (‘casual–causal’) were included to investigate the influence of prime's lexicality in the transposed-letter effect. The results showed that when compared with the substituted-letter control conditions (‘jugde–judge’ vs. ‘jupte–judge’), transposed-letter primes produced a lower negativity in the N250 component. In contrast, no differences were obtained between the two word–word priming conditions (‘casual–causal’ vs. ‘carnal–causal’). The influence of lexicality in the transposed-letter effect is discussed according to the models of visual word recognition and previous evidence from event-related potentials.