Functional cortical networks during language comprehension

Main purpose of this project is the examination of the neuronal basis of language processing at the word and sentence level. Electric brain activity (EEG) shall be recorded during different language comprehension tasks and spectral analysis shall be applied. Particular emphasis is laid on the investigation of properties and meaning of coherence patterns within special frequency bands of the EEG. Previous results showed that coherence accompanying language processing considerably changes with respect to the frequency bands investigated. Coherence is a frequency-domain measure of similarity between EEG signals and reflects the degree of functional cooperation between neuronal substrates underlying the generation of those signals. In the light of current neurobiological theories knowledge about cortical synchronisation and information transfer is necessary for an adequate characterisation of cognitive brain functioning. The chosen methods for spectral analysis of the data are Fourier transform and the application of an adaptive fit of bivariate autoregressive moving average (ARMA) models which complement the data obtained by Fourier transform due to improved time and frequency resolution. The first planned experiment deals with the investigation of neuronal correlates of different verb categories. In recent years several studies supported the notion of different functional representation of lexical categories within the brain. With the computation of EEG coherence it is possible to monitor cortical cooperation and the participation of different neural networks accompanying processing of certain categories. The second experiment shall be performed in order to monitor coherence changes during German relative clauses dependent on the imageability of the contents of these sentences. The influence of imageability on neuronal processes during sentence comprehension is not yet investigated. The third experiment deals with the role of coherence during processing congruous and incongruous German sentences and the selection of an appropriate baseline for language tasks. The fourth main aim of the proposal deals with the continuation of spectral analysis of English relative clauses. In close cooperation with Bärbel Schack (University of Jena, Germany), Marta Kutas (University of California, San Diego, USA), John King (University of Missouri-Columbia, USA) and Horst M. Müller (University of Bielefeld, Germany) an adaptive fit of bivariate autoregressive moving average (ARMA) models shall be applied to the data in order to obtain improved temporal and frequency resolution. Beside basic research on the relation between brain function, cognition and language processing the results of these experiments can be applied in different fields of Medicine, Linguistics and Cognitive Science. As an example results can be applied to the treatment of disturbed language acquisition in children (L1) and to the improvement of second language acquisition (L2). Furthermore, aphasic syndromes may be better understood and therapy may be improved. Moreover, findings on language comprehension can be used for implementation in artificial communicators by members of the SFB 360 "Situierte künstliche Kommunikatoren" (University of Bielefeld).