
Language and vision are two primary source modalities of information for humans. To uncover shared and unique mental representations across linguistic and visual inputs, one approach is to test the extent to which a certain neural signature is shared. In the current EEG study, we tested whether the response profile of posterior alpha power - a neural signature well-studied for visual working memory - generalizes to non-visual linguistic input. Employing a novel imagine-onethen-more paradigm in a large EEG sample (N = 52), we found that increasing the number of imagined objects derived from a non-visual linguistic source - auditory speech during eye closure - led to greater suppression of left posterior alpha power, and that this suppression plateaued beyond similar to 3 objects. This response profile closely mirrors effects previously observed for visuallyderived representations. As posterior alpha power likely reflects the representational load of slots/pointers or spatial attention in moment-to-moment cognition, our study thus provides novel insights into the specific representational facets shared across language and vision.
Spoken narrative comprehension unfolds across words, sentences, and events, yet it is unclear whether compact language-model features predict shared brain responses across narratives and listeners. This study tested low-dimensional cumulative-context GPT-2 predictors in the public Narratives fMRI dataset, comprising 567 motion-quality-controlled subject-story scans from 277 participants across 13 intact narratives. Gentle alignments provided word timings, and principal component analysis reduced autoregressive GPT-2 final-layer features to five components. Encoding models tested whether these components improved prediction of responses shared across listeners in 17 bilateral cortical ROIs beyond a rich lexical, speech-timing, waveform, spectral, pitch, and voicing baseline. In each leave-one-story-out fold, the feature transformation and neural coefficients were fitted to 12 equally weighted narratives and evaluated on the complete omitted narrative. GPT-2 components increased held-out-story prediction correlation by a mean Delta r = 0.0174 (95% CI [0.0055, 0.0293], p = 0.0078; 10/13 stories positive). The gain exceeded all 1000 temporal-shift null means and persisted after each tested listener was removed from all training-story responses (Delta r = 0.0045, p = 0.034). Independently embedded TR chunks did not provide reliable cross-narrative improvement (Delta r = 0.0043, p = 0.293), and cumulative features produced a larger increment (paired Delta r = 0.0131, p = 0.0366). A separate individual-scan analysis of local chunks was null. Compact cumulative-context GPT-2 features thus provided a small, reproducible increment in shared, stimulus-locked responses across narratives and listeners.
While the orthographic expectation effect (OEE) is well-documented in alphabetic languages, its manifestation in logographic systems like Chinese, particularly among children with atypical auditory experience such as those with cochlear implants (CIs), remains unexplored. This study investigated whether Mandarin-speaking children can form orthographic expectations for Chinese characters through auditory exposure alone, comparing 36 prelingually deaf children with CIs and 36 age-matched children with normal hearing (NH). Using a novel-word learning paradigm, participants were auditorily exposed to pseudo-characters. Their reaction times in a subsequent lexical decision task were analyzed using Generalized Linear Mixed-Effects Models (GLMM) to compare auditorily learned versus unlearned phonograms and non-phonograms, as well as predictable versus unpredictable phonetic radicals. Results revealed that both groups exhibited significantly faster reaction times for auditorily learned phonograms, and this expectation effect emerged robustly only when the phonetic radicals were highly predictable. However, children with CIs demonstrated markedly slower responses than their NH peers across all conditions. These findings demonstrate that Mandarin-speaking children can form robust orthographic expectations through auditory input regardless of hearing status. Although children with CIs exhibit slower baseline processing speeds due to underlying perceptual constraints, the magnitude of their orthographic expectation effect remains remarkably intact and comparable to their normal-hearing peers. Ultimately, this highlights the resilient nature of cross-modal reading acquisition, offering crucial neurocognitive insights for designing adaptive literacy interventions in clinical and educational settings.
The present study investigated whether syntactic processing in classical Chinese poetry is modular or interacts with prosodic processing, and whether semantic comprehensibility modulates this effect. Classical Chinese poetry, with its fixed rhythmic framework and flexible syntax, is ideal for examining this issue because it permits a dissociation between syntactic and rhythmic boundaries. Each poetic line had a 4 + 3 or 2 + 5 syntax, read with a conventional 4 + 3 or unconventional 2 + 5 rhythm. Syntactic structure effects were operationalized as differences between the 4 + 3 and 2 + 5 syntax. Participants listened to seven-character quatrains with high or low semantic comprehensibility and completed a word recognition task, with accuracy and reaction times recorded. EEG was recorded during listening, and the 2nd, 4th, and 7th characters of each line served as keywords. Behaviorally, reaction times showed only a Semantic Comprehensibility & times; Syntactic Structure interaction. Cluster-based permutation tests on EEG data revealed a significant three-way interaction (Semantic Comprehensibility & times; Rhythmic Pattern & times; Syntactic Structure) at the 2nd and 4th keywords. Within the conventional 4 + 3 rhythm, syntactic processing varied with semantic comprehensibility. Under high comprehensibility, the 2 + 5 syntax elicited a more negative N100 than the 4 + 3 syntax at the 2nd keyword. Under low comprehensibility, the 2 + 5 syntax elicited a larger P300 at the 2nd keyword, and the 4 + 3 syntax elicited a larger post-CPS negativity at the 4th keyword. These findings indicate that semantic comprehensibility modulates the temporal dynamics of the rhythm-syntax interaction, providing neural evidence from classical Chinese poetry for an interactive, rather than a modular, account of language comprehension.
This study investigated the influence of verb subcategorization preferences on sentence processing in Standard Indonesian. We examined how verb biases affect the processing of active and passive sentences by utilizing event-related potentials. Our findings revealed that passive sentences with active-bias verbs elicited a significant late positive shift (P600) relative to active sentences, which indicated increased processing difficulty. No such effect was detected for passive sentences with passive-bias verbs. The cost was observed only where the verb's preference conflicted with the passive structure, suggesting that verb-specific statistics, rather than word order alone, contribute to the difficulty of passive sentences. We highlight the need for a nuanced understanding of sentence comprehension that incorporates both structural and lexical influences.