Biology subjects

Koelbl, N.

Publications and source records attributed to Koelbl, N..

4 recordsLinked to original sources

bioRxiv2025

Different Multiword Verb Categories are Processed Differentially in the Brain: An Evidence from EEG Analysis and Decoding

The mental representation of multiword verb constructions is a central question in neurolinguistics: are they stored as single lexical units or completely compositional items? This study investigates the neurocognitive processing of two multiword verb types: phrasal verbs (e.g., look up) and prepositional verbs (e.g., decide on). Taking in consideration verb-to-infinitive constructions (e.g., want to go) as a control group. We analyse event-related potentials from eleven native English speakers who completed a listening task while EEG data were recorded. Grand-averaged waveforms and root-mean-square amplitudes were analysed across four-time windows. Therefore, statistical comparisons showed a significantly larger N400 amplitudes for prepositional verbs compared to phrasal verbs, while no significant differences were found between prepositional and to-infinitive constructions. Multivariate pattern analyses confirmed neural discriminability between phrasal and prepositional verbs, but not between prepositional and verb-to-infinitive structures. These results confirm that prepositional and verb-to-infinitive constructions are processed compositionally via valency-based integration, whereas phrasal verbs are stored as lexicalized units. The findings support a theoretical model in which multiword verb constructions differ in their degree of lexicalization, with measurable consequences for real-time neural processing.

neuroscience↗

bioRxiv2025

Prediction, Syntax and Semantic Grounding in the Brain and Large Language Models

Language comprehension involves continuous prediction of upcoming words, with syntactic structure and semantic meaning intertwined in the human brain. To date, few studies have used combined magnetoencephalography (MEG) and electroen-cephalography (EEG) measurements to investigate how syntactic processing, predictive coding, and semantic grounding interact in real time. Here we present the first combined MEG-EEG investigation of syntactic processing and semantic grounding under naturalistic conditions. Twenty-nine healthy participants listened to a German audio book while their neural responses were recorded. Event-related fields and event-related potentials for four word classes - nouns, verbs, adjectives, and proper nouns - showed highly reproducible, characteristic spatio-temporal signatures, including significant pre-onset activity for nouns, suggesting enhanced predictability of this word class. Source-space analyses revealed pronounced activation in the pre- and post-central gyri for nouns, suggesting a deeper semantic grounding of nouns in e.g. sensory experiences than verbs. To further investigate predictive mechanisms, we analyzed the hidden representations of the large language model Llama. By comparing the transformer-based representations to neural responses, we explored the relationship between computational language models and human brain activity, offering new insights into syntactic and semantic prediction. These findings highlight the power of simultaneous MEG-EEG recordings in unraveling the predictive, syntactic, and semantic mechanisms that underlie the comprehension of natural language.

neuroscience↗

bioRxiv2024

Analyzing Differences in Processing Nouns and Verbs in the Human Brain using Combined EEG and MEG Measurements

Language and consequently the ability to transmit and spread complex information is unique to the human species. The disruptive event of the introduction of large language models has shown that the ability to process language alone leads to incredible abilities and, to some extent, to intelligence. However, how language is processed in the human brain remains elusive. Many insights originate from fMRI studies, as the high spatial resolution of fMRI devices provides valid information about where things happen. Nevertheless, the limited temporal resolution prevents us from gaining a deep understanding on the underlying mechanisms. In this study, we performed combined EEG and MEG measurements in 29 healthy right-handed human subjects during the presentation of continuous speech. We compared the evoked potentials (ERPs and ERFs) for different word types in source space and sensor space across the whole brain. We found characteristic spatio-temporal patterns for different word types (nouns, verbs) especially at latencies of 300ms to 1 s. This is further emphasized by the fact that we observe these effects in two pre-defined sub-samples of the data set (exploration and validation sample). We expect this study to be the starting point for further evaluations of semantic and syntactic processing in the brain.

neuroscience↗

bioRxiv2024

Temporal and Hemispheric Dynamics in Neural Processing of Auditory and Speech Stimuli Across Linguistic Complexity: A MEG Source Space Study

In this study we investigated the neural processing of auditory stimuli of varying complexity: a non-linguistic (pure tone), a simple linguistic (phoneme) and a complex linguistic (word) stimulus. We recorded brain activity of 30 healthy, righthanded participants using magnetoencephalography (MEG), and compared the resulting evoked fields (ERFs) in source space in three different time intervals, i.e. early (0-250ms), mid (250-500ms) and late (500-750ms) responses. Our results reveal a bilateral activation during early response and rightlateralized activation in the mid-phase for all stimuli. Hoewever, the late response exhibited lateralization variations. The pure tone predominantly activated the right hemisphere, consistent with pitch processing theories. The phoneme primarily engaged the left hemisphere, supporting its role in phonemic processing. Notably, the word elicited activation in both hemispheres, reflecting phonemic processing on the left and stress patterns on the right. These findings highlight the intricate interplay between temporal processing and hemispheric lateralization in speech perception, emphasizing the importance of stimulus complexity and temporal dynamics in understanding auditory and speech processing.

neuroscience↗