Spectral analysis of nonsynchronous EEG activity and problem solving
Right and left side electroencephalograms (EEG) and electromyograms (EMG) were recorded in 20 right-handed, male adults during a verbal and a spatial task. EEG and EMG activity was quantified with spectral analysis, and bands of energy centered at 14 Hz (cycles per second), 20 Hz, 30 Hz, 40 Hz, 50 Hz, and 70 Hz were formed. From preliminary analyses of EEG and EMG data, three criteria were determined which demonstrated consistent and reliable dissociations between distributions of energy in EEG and EMG leads. Thorough analyses of the corrected EEG data were then conducted. It was found that for right and left EEG leads, during both tasks, there were slight, yet in most cases highly significant decreases in 14 Hz and 20 Hz bands, with concomitant increases in 40 Hz, 50 Hz, and 70 Hz bands. During spatial items, 30 Hz also increased on both sides. It was also found that 40 Hz was significantly correlated with 50 Hz and 70 Hz for corrected and uncorrected data. No significant indications of lateralization were found. No significant changes in 14 Hz, 20 Hz, or 30 Hz occurred as a result of applying the correction criteria. The major finding of this project was the demonstration of clear differences in the distribution of energy between EEG and EMG leads. The importance of this for future 40 Hz research is that EMG corrections can be made with activity in the high frequency range of the EEG leads alone. Simultaneous recording of EMG activity is not required for elimination of EMG contamination. It was also concluded that there is a shift in energy from the lower frequency bands (14 Hz and 20 Hz) to the higher frequency bands (40 Hz, 50 Hz, and 70 Hz) during problem solving. The changes in the lower bands are consistent with similar research using spectral analysis. The changes in the 50 Hz and 70 Hz bands clearly indicate increased EMG activity. That increases in 40 Hz activity correlated with increases in 50 Hz and 70 Hz activity indicates insufficient EMG artifact elimination. Refinements in the criteria are suggested, and limitations in the application of spectral analysis in fast-frequency EEG research are discussed.