Human communication investigated with magnetoencephalography: Speech, music, and gestures
In this chapter, we report a number of studies applying magnetoencephalography (MEG) to the investigation of the modes of human communication. We seek to elucidate the spatio-temporal organization of the processing of diVerent levels of information during the perception of speech, music, and gestures. In particular, we investigated the preattentive processing of speech sounds, the processing of syntactic information in speech and music, the mechanisms of lexical selection during language production, the perception of musical phrase structure, the coupling between sensory and motor representations in musicians, and the understanding of emblematic hand signs. We applied source localization methods, enabling us not only to localize the active brain areas underlying the respective cognitive processes, but also to attribute precise timing to these processes. The actual choice of methods for both the physical modeling of the head (forward problem) and the reconstruction of the brain activity (inverse problem) has to be made anew in each case, carefully considering the nature and quality of the data, available supporting information (e.g., magnetic resonance images of the head), and the hypotheses on the possible outcome. Consequently, we applied a whole range of techniques throughout the studies reported in this chapter. The solutions to the forward model included simple spherical as well as realistically shaped head models. Inverse methods used in our studies encompassed the reconstruction of focal sources (spatio-temporal dipole fit), of two-dimensional current distributions (brain surface current density mapping), and of threedimensional voxel-based activity distributions (magnetic field tomography, multiple signal classification). Principal component analysis was used to disentangle diVerent neuronal systems or to detect subtle eVects obscured by irrelevant brain activity. Additional information from functional magnetic resonance imaging was also employed. The reported results highlight the ability of MEG to map the brain activity underlying high-level cognitive processes in space and time.