Besides conventional power spectral analysis and computed
estimates of the dimensional complexity (D2) of the EEG and ERP
time series, we use analysis approaches that give equal weight to
the data's temporal and spatial characteristics. In these
approaches, preference is given to data-driven procedures where
top-down con- straints are kept to a minimum. Three approaches
are: (1) Parsing of the series of brain electric momentary maps
into epochs of quasi-stable topography, the so-called
microstates. Since different momentary potential maps must have
been generated by the activity of different neural assemblies, it
is reasonable to assume that they represent different brain
functional states, i.e., different steps or modes of information
processing. Examining the series of momentary field maps showed
that the changes of map landscape are discontinuous, stepwise.
Hence, there is no continual development of brain state; rather,
there are distinct, brief states concatenated by rapid state
transitions in the sub-second time range. This lead to the
concept of brain electric microstates, the manifestation of
hypothetical 'atoms of thought', i.e., to the possibility to
identify a repertoire of building blocks of mentation. Overviews
of concept and results are in Lehmann 1994, 1995. A general
solution to the parsing problem is presented in Pascual- Marqui
et al. (1995), using a boot-strapping approach. (2) Source
modelling of the potential fields in the time domain (a novel
approach which requires minimal assumptions is presented in
Pascual-Marqui et al. 1994), and source modelling in the
frequency domain (the latter method was developed by the KEY
Institute in 1990). (3) Computation of the global dimensional
complexity of the brain electric signals. In addition to the
conventional dimensional complexity single time series (after
state space reconstruction following Takens' procedure), we apply
computations of the 'global dimensional complexity' of the
trajectory of the brain state (as assessed by the momentary maps)
in state space, using either D2-calculations (where the embedding
dimension is the number of simultaneously recorded time series or
the independent measure OMEGA (e.g., see 'chewing gum...' in
section 'Chemical Effects').
Key words: power spectral analysis - microstates - state
transitions - 'atoms of thought' - source modelling of the
potential field - dimensional complexity - D2