116
which prevented this calculation, then the raw data were repaired using linear interpolation. Processed position sample data were used to find the X axis, Y axis, and rotational sway orientation positional standard deviations, plus the X axis and Y axis average velocities and velocity standard deviations for each trial.
Individual Fourier power spectra of the processed X
axis, Y axis and rotational data for all subjects for each trial were determined using a Fast Fourier Transform (FFT). Trials with less than 100 contiguous samples (due to stimulus clipping) were excluded from EFT analysis due to the lack of low frequency information. Within subject, the power spectra of identical trial conditions were averaged within sway orientations. Fourier spectrum gains were calculated as the ratios of power spectra for different trial conditions (within sway orientations). Spectrum gains were computed from (a) spectral power ratios of each of the feedback conditions to the corresponding jitter simulation conditions, and (b) spectral power ratios of each of the jitter simulation conditions, the eyes closed condi-tion, and the +2 feedback gain condition, to the visible stationary stimulus "ideal baseline" condition. The data were collapsed by averaging these Fourier spectrum gains across subjects, and by grouping the sway spectral gains into
0.5 Hz wide frequency bins from 0 to 5 Hz. The resulting output data of the DPS were the means and standard errors
..