Issn Print: 0736-0258

Publication Date: 1997/03/01


Print

FUNDAMENTALS OF BRAIN SOURCE IMAGING: S301

Michael Scherg

    loading  Checking for direct PDF access through Ovid

Excerpt

EEG and MEG reflect the overlap of brain activity with predominant contributions from synchronized neuronal populations at the cortical convexity and in fissural cortex, respectively. Whereas the number of channels limits spatial resolution, abundant information is available in the time signals. The fundamental problem of brain source imaging resides in utilizing both the spatial and the temporal information.
Source localization is governed by the question: "Where is the generator of this peak in the brain?" Implicitly, this assumes that spatial information is sufficient for accurate source localization. In practice, this assumption is rarely justified. The scope widens in asking the reverse question: "What signal waveforms are contributed from brain regions A, B, etc., to form a particular segment of EEG/MEG?" In other words: "Can we decompose the measured waveforms to identify brain regions that contribute and those that do not contribute?" The overlap of localized brain activities at the surface can be described by matrix algebra. When combining the topographic field distributions of brain regions to be interrogated with topographies describing background noise signals, optimal spatio-temporal filters can be constructed to decompose and image localized brain activities.
Dr. Scherg is a consultant and major stock shareholder of HEGIS Software AmbH.
    loading  Loading Related Articles