The present study examined how task demands modulate neural activations during the processing of emotion-label words (words describing emotional states) and emotion-laden words (words triggering emotions through connotations) in a second language (L2). To this end, we directly compared behavioral and electrophysiological responses in late Chinese-English bilinguals when the words’ affective information was task-relevant (emotional categorization task, ECT) versus task-irrelevant (emotional Stroop task, EST). Our results revealed the modulation of emotion word type and valence on L2 emotion word processing. Specifically, negative emotion-laden words exhibited slower response times, and positive emotion-label words elicited larger Late Positive Complex amplitudes. Notably, these effects were observed exclusively within the task where the emotional dimension of the stimuli was task-relevant, that is, when participants explicitly determined the word valence. Taken together, our study highlights the potential task-dependent nature of the effects of emotion word type and valence in late bilinguals’ L2, suggesting that assumptions about the universality of such effects should be evaluated in light of task demands and language context.

Deviations in P300 activity have been implicated in depression and anxiety; however, much of this research has been conducted in adult samples and has primarily examined the association between P300 amplitude and internalizing symptoms between participants. We sought to simultaneously examine the between- and within-subject associations between depression and anxiety symptoms with P300. Self-report and neural data from a flanker task were collected at three timepoints over the course of two years in a large sample of adolescents (n = 490). Blunted P300 was robustly related to elevated between-subject depression. Conversely, elevations in within-subject anxiety were associated with larger P300. Results implicate the P300 as a reliable correlate of between-subjects level depression-related deficits in cognitive functions that is not susceptible to within-subject changes. Additionally, P300 also serves as a correlate of within-subject elevations in youth anxiety symptoms likely reflecting greater hyperarousal at the time of assessment.

The Bergen 4-day treatment (B4DT) for obsessive compulsive disorder (OCD) is a concentrated form of exposure and response prevention that has been implemented at the Icelandic Anxiety Centre. The aim of the present study was to assess 12-month results of 86 participants undertaking the treatment. The Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) was administered at pre-, post-, and 3- and 12-month follow-up, accompanied by measures on generalised anxiety (GAD-7), depression (PHQ-9), and functional impairment (WSAS). The mean pre-treatment score on Y-BOCS was 30.5 (SE=0.5), the post-treatment score was 10.7 (SE=0.4), and the 12-month follow-up score was 11.5 (SE=0.9). There was no significant difference in Y-BOCS scores from post-treatment to 3- and 12-month follow-up, which indicates that treatment gains were maintained over time. Functional impairment, along with symptoms of anxiety and depression, decreased significantly from pre- to post-treatment, with further reductions in functional impairment observed at the 12-month follow-up. All participants completed treatment, and at 12-month follow-up, 83.7% had responded and 67.4% recovered. Among the participants that had responded or remitted at post-treatment, 22% had a worse outcome at 12-month follow-up, but 54% of the participants in the response or no change categories at post-treatment had a better outcome at follow-up. These results support the effectiveness and long-term benefits of the B4DT when implemented outside of its country of origin.

Key learning aims

1. (1) The Bergen 4-day treatment (B4DT) is a highly intensive format of exposure and response prevention (ERP) for OCD.

2. (2) Treatment gains from the B4DT for OCD are sustained for at least 12 months, with 67% of patients remaining recovered and 84% showing improvement.

3. (3) Approximately half of those who do not fully benefit initially experience further improvement by the 12-month follow-up.

4. (4) Drop-out rates are negligible, which is unusual for exposure and response prevention.

5. (5) The B4DT demonstrates effectiveness beyond its country of origin.

This study investigated whether L2 processing of derived words engages biphasic morphological decomposition, comprising morpho-orthographic segmentation followed by morpho-semantic integration, as L1 processing does. Using an overt priming paradigm (SOA = 300 ms), ERP responses were compared across morphological (e.g., farmer–farm), orthographic (e.g., cashew–cash) and semantic (e.g., doctor–nurse) priming conditions in native and L2 speakers. Results revealed that both language groups exhibited distinct priming effects for morphologically related prime–target pairs across the early and late N400 windows, reflecting morpho-orthographic segmentation and morpho-semantic integration, respectively, rather than additive effects of form and meaning overlap. However, the late negativity effect, reflecting intensified lateral inhibition among similar orthographic representations, was observed during orthographic priming only in native speakers, suggesting less efficient inhibitory control in L2 processing. These findings are discussed within the framework of the Shallow Structure Hypothesis, which has provided a theoretical basis for many previous L2 studies of derived-word processing.

As people increasingly interact with large language models (LLMs), a critical question emerges: do humans process language differently when communicating with an LLM versus another human? While there is good evidence that people adapt comprehension based on their expectations toward their interlocutor in human–human interaction, human–computer interaction research suggests the adaptation to machines is often suspended until expectation violation occurs. We conducted two event-related potential experiments examining Chinese sentence comprehension, measuring neural responses to semantic and syntactic anomalies attributed to an LLM or a human. Experiment 1 revealed reduced N400 but larger P600 responses to semantic anomalies in LLM-attributed text than human-attributed one, suggesting participants anticipated semantic errors yet required increased composition/integration efforts. Experiment 2 showed enhanced P600 responses to LLM-attributed than human-attributed syntactic anomalies, reflecting greater reanalysis or integration difficulty in the former than in the latter. Notably, neural responses to LLM-attributed semantic anomalies (but not syntactic anomalies) were further modulated by participants’ belief about humanlike knowledge in LLMs, with a larger N400 and a smaller P600 in participants with stronger belief of humanlike knowledge in LLMs. These findings provide the first neurocognitive evidence that people develop mental models of LLM capabilities and adapt neural processing accordingly, offering theoretical insights aligned with multidisciplinary frameworks and practical implications for designing effective human–AI communication systems.

Subject relative (SR) clauses have a reliable processing advantage in VO languages like English in which relative clauses (RCs) follow the head noun. The question is whether this is also routinely true in OV languages like Japanese and Korean, in which RCs precede the head noun. We conducted an event-related brain potential (ERP) study of Korean RCs to test whether the SR advantage manifests in brain responses, and to tease apart the typological factors that might contribute to these responses. Our results suggest that brain responses to RCs are remarkably similar in VO and OV languages. Our results also suggest that the marking of the right edge of the RC in Chinese (Yang et al. 2010) and Korean and the absence of such marking in Japanese (Ueno & Garnsey 2008) affect the response to the following head noun. The consistent SR advantage found in ERP studies lends further support to a universal subject preference in the processing of relative clauses.

Individuals differ in a range of processes related to reading comprehension, including working memory capacity, decoding skills, inference making and main idea identification. In this exploratory study, we examined evoked potential N400 amplitude during reading comprehension tasks and focused on identifying the main idea in the text, modulated by working memory capacity. Participants included monolinguals or bilinguals who were either typical readers (n = 33) or had been diagnosed with dyslexia (n = 19). Analyses revealed significant group differences for main idea conditions. Participants with dyslexia showed greater N400 amplitude than typical readers, particularly in the right hemisphere, when the main idea was in the last position in the paragraph. There were no significant differences in performance between bilinguals and monolinguals, which does not support the idea of a cognitive advantage for bilingualism. It was noteworthy that, if they had dyslexia, they were similarly negatively impacted by their reading disability. Findings highlight the processing advantages typical readers have relative to dyslexia.

Although bilinguals use both auditory and visual cues, the cognitive cost of language switching in audiovisual contexts is unclear. We investigated the cost in Tibetan–Chinese bilinguals using a task with audiovisual, visual and auditory modalities. In Study 1, the audiovisual modality yielded the fastest reaction times, reflecting improved processing efficiency. ERP data revealed smaller positive amplitudes in the early window (200–350 ms) for audiovisual modality, indicating reduced neural demand, while only auditory modality showed significant divergence in the later window (350–700 ms). Moreover, audiovisual context, L2-to-L1 switching and early neural responses predicted switching behavior. Study 2 replicated the behavioral and ERP findings of Study 1 and demonstrated that auditory input and second-language processing exacerbated switch costs. These findings shed light on multisensory integration in language switching by demonstrating that audiovisual cues reduce switch costs, whereas auditory input and second-language processing exacerbate them, with implications for language education and cognitive interventions.

Understanding high-variability speech is particularly challenging for second-language (L2) learners due to difficulties with extrinsic normalization, a perceptual strategy utilizing contextual cues to overcome speech variability. This study investigates the neural correlates of these difficulties among Mandarin speakers learning Cantonese, using EEG. Behaviorally, Mandarin learners demonstrated a significant yet considerably reduced ability to normalize Cantonese tone variability with contexts compared to native Cantonese speakers. EEG analysis showed that while native speakers engage multiple neural components (N1, P2, and LPC) for acoustic, phonetic/phonological, and cognitive adjustments in extrinsic normalization, Mandarin learners only activated P2, focusing on phonetic/phonological adjustments. This discrepancy underscores the multi-faceted nature of successful extrinsic normalization, which L2 learners fail to fully engage. L2 immersion significantly improves extrinsic normalization, particularly at the cognitive-adjustment stage. Overall, this study illuminates the intricate nature of poor extrinsic normalization in L2 learners and the importance of L2 immersion for effective L2 speech perception.

A central question in foreign language (LX) learning is how vocabulary acquisition is affected by using image versus orthographic referents. According to the picture superiority effect (PSE) and bilingual/dual coding theory (b/DCT), images should lead to better novel word encoding and retrieval. We tested this prediction using behavioral and event-related potential (ERP) measures. Thirty Polish native speakers learned 40 LX (artificial language) words using either image or L1/orthographic referents. After 24 hours, participants were tested using a translational priming paradigm in congruent and incongruent training-testing modalities. Behavioral results showed higher accuracy and faster responses for LX words learned and tested with images, in line with the PSE and b/DCT. ERP results revealed smaller Late Positive Complex (LPC) amplitudes for words preceded by image compared to lexical primes, likely reflecting less cognitively demanding lexical retrieval. These results provide converging evidence that visual referents provide a more salient modality for L2 learning.