[Continuous frequency analysis (spectral trend analysis) in the evaluation of long-term EEG recordings]

This study demonstrates the possibilities of continuous frequency analysis supporting the evaluation of long lasting EEG recordings which present problems in analysing large amount of data. Avoiding some disadvantages of the usual presentation of power spectra by compressed spectral array compression of the spectral data will be obtained in a more suitable way by sequential plots of the data on the abscissa as time axis. The spectra will be separated in only a few number of frequency bands. A special conversion procedure (linear multiplication of the data within each frequency band by logarithmic calculated factors) provides for the essential enhancement of low spectral power in the upper EEG frequency domain preserving the full dynamic of spectral variations. The efficiency of this spectral trend analysis is shown by two examples: (1) In the study of the variations of cerebral excitability in a patient with epileptic seizures depending on the waking-sleeping-cycle (Fig. 3), and (2) in the documentation of the EEG dynamics in the course of intensive care monitoring. Fig. 5 shows the development of a cerebral autorhythmicity representing reintegration of cerebral function in a patient with severe head injury.