At times, one may see a first harmonic of the EEG alpha rhythm, showing up as a peak in the spectral energy plot, and particularly in the z-score spectral maps. It has been stated that this indicates a problem with the amplifier. This article explains that when this appears, it is not due to any problem with the amplifier, but is a result of the peculiarities of the FFT analysis.
Whenever a signal is not a perfect sine-wave, it will have harmonic energy in the spectrum. The harmonics will appear at 2x, 3x, and higher harmonics of the fundamental wave. Therefore, if an alpha wave is not a perfect sinewave, it is reasonable to expect harmonic energy at twice the alpha band frequency, due to the nonsinusoidal shape of the wave. Also note that although there is a small broad peak in the raw spectra, the z-scores show a large peak. This is likely due to the fact that the database was collected with amplifiers that do not have as strong a response above 25 Hz, so that the harmonic was not seen in the original database. Modern amplifiers, with a wider bandwidth, reproduce the precise shape of the alpha wave, and the harmonic appears in the raw data, but is even more exaggerated when compared with the normative values from the older model amplifiers.
Below are two examples of this. They both show a broad band of energy that is twice the alpha band. They also show a particularly large energy peak in the z-score spectra. Both spectra are done with NeuroGuide. The first is with a Discovery amplifier, and the second is with a Deymed amplifier. Both plots show a nonsinusoidal alpha wave, and a resulting band of harmonic energy.
The following plot, from a Discovery amplifier, subject has large, fast alpha, that is not sinusoidal, but has a “wicket” shape commonly seen when alpha is large. This was published in Collura, T.F. (2014) Specifying and Developing References for Live Z-Score Neurofeedback . Neuroconnections, Spring 2014. 26-39.
The following graph, from a chapter by Acosta-Urquidi and Gunkelman in Larsen, S. (2006) The Neurofeedback Solution, Rochester, VT: Inner Traditions. This was taken with a Deymed amplifier. Again, the subject has large and nonsinusoidal alpha, producing a similar band of alpha harmonics. Again, this is not due to any problem with the amplifier, it is due to the fact that the fft analysis sees the nonsinusoidal alpha, and the mathematical analysis produces energy at the first harmonic, twice the alpha frequency.
