Tracing the ventral stream for auditory speech processing in the temporal lobe by using a combined time series and independent component analysis.

Recent models of speech processing claim a ventral and dorsal stream for processing language as for the visual domain. Patients with aphasia are demonstrating selective disturbances within parts of these streams. Here, we are presenting an fMRI study, which mainly focuses on the ventral stream, i.e. the phonological, lexical, and semantic processing of auditory speech. We used three linguistic conditions: (1) listening to reversed words, (2) listening to real words and, (3) covert naming after short description. We analysed the temporal dynamics of the BOLD signal as well as the temporal dynamics of lateralisation. In addition, an independent component analysis (ICA) was applied. By combining the results from these methods, we were able to separate the areas for pure auditory-phonetic perception, located bilaterally in the middle and posterior aspects of the superior temporal lobe, from predominantly left-lateralised areas for semantic and lexical processing, located in posterior-inferior as well as anterior regions of the temporal lobe, at the temporal-parietal-junction, and in the inferior frontal gyrus. These two networks differed with respect to their time course and temporal dynamics of lateralisation, and they were separated into two components of the ICA analysis. By being able to separate these two networks, we further conclude that this complementary analysis method might also be relevant for exploring recovery processes in patients with aphasia.