Investigations into the neural representation of prosodic, lexical, and syntactic properties of spontaneous, natural speech production using electrocorticography (ECoG)

Abstract: The present thesis reports on interdisciplinary attempts to elucidate the hitherto unexplored neural correlates of linguistic processing in conditions of non-experimental, natural overt speech production. To this end, the author of this thesis, supported by colleagues, built up a multimodal neurolinguistic corpus (“The Freiburg/First Neurolinguistic Corpus”), composed of synchronized audio, video, electrocorticographic (ECoG) materials and linguistic annotations on different levels of linguistic abstraction. This corpus allowed us to study the neural effects related to several aspects of natural language, each of whom was treated in a separate study. Study 1 was dedicated to prosody and addressed the neural activity related to the production of the focus accent, Study 2 was dedicated to questions related to the neural representation of word complexity, and Study 3 investigated the syntactic processing accompanying natural clause production. The psycholinguistic methods we used consisted of (i) application of a matching procedure to select controlled word categories out of the natural language data (Study 1), (ii) orthogonalization of the linguistic parameters with the help of a linear regression model to overcome the problem of collinearity between correlated linguistic parameters, and (iii) the usage of a principal component analysis to extract most informative components of the linguistic material. The neuroscientific approach consisted either of group comparisons, in which neural effects underlying linguistically distinctive groups of words were compared (Study 1) or of correlation of neural activity with individual linguistic parameters (Study 2) and with principal components explaining most of the variances in the linguistic data (Study 3). We were interested in neural effects reflecting differences in the effort related to the production of speech units of different linguistic complexity. We were looking for neural effects which would be spatially focalized and which would be manifested in gamma activity (>35 Hz), since it known as a reliable and functionally specific marker of event-related effort. Based on knowledge from previous research, we were expecting neural effects to be spatially focalized, extended in time and extended over a broad range of gamma frequencies. We were, indeed, able to observe anatomically local effects. In Study 1, activity in postcentral areas proved to convey information about prosodic properties of content words. In Study 2, the proportional relation between the number of consonants and vowels, which was the most informative parameter with regard to the neural representation of word complexity, showed effects in two anatomically focal areas: the frontal one was located at the junction of the premotor cortex, the prefrontal cortex, and the posterior portion of Broca’s area (Brodmann area 44). The postcentral one lay directly above the lateral sulcus and comprised effects on the ventral central sulcus, in the parietal operculum and in the adjacent inferior parietal cortex. The preliminary findings of Study 3 indicate that pericentral cortical areas implicated in mouth motor functions show correlations with syntax-relevant information, reflected in the first principal component explaining over 30% of the variances in the linguistic data. While Study 3 yielded temporo-frequentially extended effects in high gamma frequencies, as we had expected, Studies 1 and 2 showed temporally and frequentially narrow effects with little reproducibility in terms of these neural characteristics. This may indicate moderate representation of the phenomena investigated in Studies 1 and 2 in the investigated neural signals. Alternatively, since the spectrum of gamma frequencies is a composite phenomenon relying on multiple cell types, it is also conceivable that these temporo-frequentially narrow effects may point to the functionally specific activation of small local populations of neurons, whose signal properties vary between linguistic parameters and subjects. Since we are not aware of published ECoG works investigating neural effects of linguistic processing during natural speech production, further validation of these observations and speculations is required. Beyond the here summarized findings, the overall contribution of this work to the field of neurolinguistics is that we have developed ways to study the neural effects related to natural language production, to obtain control over potentially confounding parameters in such data, and to surmount the problem of collinearity between multiple linguistic features of the neurolinguistic material

Angaben zum Objekt

• Standort: Deutsche Nationalbibliothek Frankfurt am Main
• Umfang: Online-Ressource
• Sprache: Englisch
• Anmerkungen: Universität Freiburg, Dissertation, 2021

Klassifikation und Themen

• Schlagwort: Speech ; Investigations

Beteiligte, Orts- und Zeitangaben

• Ereignis: Veröffentlichung
• (wo): Freiburg
• (wer): Universität
• (wann): 2021
• Urheber: Glanz, Olga
• Beteiligte Personen und Organisationen: Auer, Peter ; Ball, Tonio ; Albert-Ludwigs-Universität Freiburg. Sprachwissenschaftliches Seminar ; Hermann-Paul-Centrum für Linguistik ; DFG-GRK 1624 "Frequenzeffekte" ; Albert-Ludwigs-Universität Freiburg. Fakultät für Biologie ; BrainLinks - BrainTools Acting.Thoughts ; Neurochirurgische Universitätsklinik. Freiburg im Breisgau ; Albert-Ludwigs-Universität Freiburg. Philologische Fakultät

Weitere Informationen

• DOI: 10.6094/UNIFR/219721
• URN: urn:nbn:de:bsz:25-freidok-2197218
• Rechteinformation: Kein Open Access; Der Zugriff auf das Objekt ist unbeschränkt möglich.
• Letzte Aktualisierung: 25.03.2025, 13:49 MEZ