Time-varying properties of respiratory fluctuations in pupil diameter of human eyes.

The time varying properties have been investigated of respiratory components in fluctuations of the pupil diameter (PDF) of the human eye. PDF and instantaneous lung volume were simultaneously measured for 45 minutes under natural binocular viewing conditions. The short-term Fourier transform and the autoregressive signal modeling together with the recursive least squares algorithm have been applied for the estimation of nonstationary power spectra. Data analysis showed a bursting type non-stationarity in the appearance of the respiratory component in PDF.