BACKGROUND: Several measures have been developed to quantify the change in EEG from wakefulness to deep anaesthesia. Measures of signal complexity or entropy have been popular and even applied in commercial monitors. These measures quantify different features of the signal, however, and may therefore behave in an incomparable way when calculated for standardized EEG patterns. METHODS: Two measures widely studied for anaesthesia EEG analysis were considered: spectral entropy and approximate entropy. First, we generated surrogate signals which had the same spectral entropy as a prototype signal, the sawtooth wave. Secondly, EEG samples where rhythmic pattern caused a peak in the power spectrum in the α-frequency band were modified by enhancing or suppressing the corresponding rhythm. RESULTS: We found that the value of spectral entropy does not, in general, correlate with the visual impression of signal regularity. Also, the two entropy measures interpret a standardized artificially modified EEG signal in opposite directions: spectral peak of increasing amplitude in the α-frequency band causes spectral entropy to increase but decreases approximate entropy when low frequencies are present in the signal. CONCLUSIONS: Spectral entropy and approximate entropy of EEG are two totally different measures. They change similarly in deepening anaesthesia due to an increase in slow activity. In some cases, however, they may change in opposite directions when the EEG signal properties change during anaesthesia. Failure to understand the behaviour of these measures can lead to misinterpretation of the monitor readings or study results if no reference to the raw EEG signal is taken.
BACKGROUND: Several measures have been developed to quantify the change in EEG from wakefulness to deep anaesthesia. Measures of signal complexity or entropy have been popular and even applied in commercial monitors. These measures quantify different features of the signal, however, and may therefore behave in an incomparable way when calculated for standardized EEG patterns. METHODS: Two measures widely studied for anaesthesia EEG analysis were considered: spectral entropy and approximate entropy. First, we generated surrogate signals which had the same spectral entropy as a prototype signal, the sawtooth wave. Secondly, EEG samples where rhythmic pattern caused a peak in the power spectrum in the α-frequency band were modified by enhancing or suppressing the corresponding rhythm. RESULTS: We found that the value of spectral entropy does not, in general, correlate with the visual impression of signal regularity. Also, the two entropy measures interpret a standardized artificially modified EEG signal in opposite directions: spectral peak of increasing amplitude in the α-frequency band causes spectral entropy to increase but decreases approximate entropy when low frequencies are present in the signal. CONCLUSIONS: Spectral entropy and approximate entropy of EEG are two totally different measures. They change similarly in deepening anaesthesia due to an increase in slow activity. In some cases, however, they may change in opposite directions when the EEG signal properties change during anaesthesia. Failure to understand the behaviour of these measures can lead to misinterpretation of the monitor readings or study results if no reference to the raw EEG signal is taken.
Authors: Juan P Ugarte; Andrés Orozco-Duque; Catalina Tobón; Vaclav Kremen; Daniel Novak; Javier Saiz; Tobias Oesterlein; Clauss Schmitt; Armin Luik; John Bustamante Journal: PLoS One Date: 2014-12-09 Impact factor: 3.240