Waves and Words: Oscillatory activity and language processing

Successful language comprehension depends not only on the involvement of different domain-specific linguistic processes, but also on their respective time-courses. Both aspects of the comprehension process can be examined by means of event-related brain potentials (ERPs), which not only provide a di...

Full description

Saved in:
Bibliographic Details
Main Author: Roehm, Dietmar
Contributors: Schlesewsky, Matthias HD Dr. (Thesis advisor)
Format: Dissertation
Language:English
Published: Philipps-Universität Marburg 2004
Germanistische Sprachwissenschaft
Subjects:
EEG
Online Access:PDF Full Text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Successful language comprehension depends not only on the involvement of different domain-specific linguistic processes, but also on their respective time-courses. Both aspects of the comprehension process can be examined by means of event-related brain potentials (ERPs), which not only provide a direct reflection of human brain activity within the millisecond range, but also allow for a qualitative dissociation between different language-related processing domains. However, recent ERP findings indicate that the desired one-to-one mapping between ERP components and linguistic processes cannot be upheld, thus leading to an interpretative uncertainty. This thesis presents a fundamentally new analysis technique for language-based ERP components, which aims to address the ambiguity associated with traditional language-related ERP effects. It is argued that this new method, which supplements ERP measures with corresponding frequency-based analyses, not only allows for a differentiation of ERP components on the basis of activity in distinct frequency bands and underlying dynamic behaviour (in terms of power changes and/or phase locking), but also provides further insights into the functional organisation of the language comprehension system and its inherent complexity. On the basis of 5 EEG experiments, I show (1) that it is possible to dissociate two superficially indistinguishable language-related ERP components on the basis of their respective underlying frequency characteristics (Experiment 1), thereby resolving the vagueness of interpretation inherent to the ERP components themselves; (2) that the processing nature of the classical semantic N400 effect can be unambiguously specified in terms of its underlying frequency characteristics, i.e. in terms of (evoked and whole) power and phase-locking differences in specific frequency bands, thereby allowing for a first interpretative categorisation of the N400 effect with respect to its underlying neuronal processing dynamics; and (3) that frequency-based analyses may be employed to distinguish the semantic N400 effect from N400-like effects that appear in contexts which cannot readily be characterised as semantic-interpretative processes. Experiments 2 5 investigated the processing of antonym relations under different task conditions. Whereas in Experiment 2, the processing of antonym pairs (black white) was compared to that of related (black yellow) and non-related (black nice) word pairs in a sentence context, Experiments 3 to 5 presented isolated word pairs. The frequency-based analysis showed that the observed N400 effects were not uniform in nature, but rather resulted from the superposition of functionally different frequency components. Task-relevant targets elicited a specific frequency modulation, which showed up as a P300-like positivity in terms of ERP measures. In addition, lexical-semantic processing elicited a pronounced increase in a different frequency range that was independent of the experimental context. For antonyms (Experiments 2 and 3), the task-related positive component appeared almost simultaneously to the N400 deflection for non-related words, thereby giving rise to a substantial N400 effect. In contrast, for pseudowords (Experiment 5), this positivity appeared in temporal succession to the N400. In sum, in the present results provide converging evidence that N400 effects should not be regarded as functionally uniform. Depending on the respective task and stimulus manipulations, the N400 effect appears as a result of the superposition of functionally different activities, which can be clearly distinguished in terms of their underlying frequency characteristics. In this way, the proposed frequency-based methods directly bear upon the interpretation of language-related ERP effects and thus have straightforward consequences for psycholinguistic theory. In view of the phenomenon that language-related processes have, in a number of cases, been directly attributed to the lexical-semantic processing domain on account of the observation of an N400, these results not only call for a reinterpretation of previous findings but also for a reinterpretation of their theoretical consequences.
DOI:https://doi.org/10.17192/z2005.0110