Time domain indices and discrete power spectrum in electroencephalographic processing.

The classical fast Fourier transform (FFT) methods to process experimental signals have the problem of providing multivariant measures with dimension and hence depend on the devices and measure units. These features make the comparisons between different teams and equipments rather difficult. The aim of the present study is to prove the exact mathematical relation between the Hjorth parameters and the spectral powers obtained by means of the quoted algorithm. The proof confirms that these quantifiers can also be used for the exploration of spectral properties of univariant recordings. The formulae proposed provide an easy and low-cost procedure to compute the descriptors if a standard FFT algorithm is performed on the signal. A numerical study of electroencephalographic recordings involving spectral and Hjorth parameters is described in the paper in order to quantify the electroencephalogram of children with problems of attention. [ABSTRACT FROM AUTHOR]