[Study of non-visually induced smooth pursuit eye movements and their predictability using pseudorandom target motion].

It has been found that the smooth pursuit eye movements (SPEM) are elicited by not only visual stimuli but also non-visual information such as the subject's fingertip movement and a moving sound source. We have already reported the quantitative analysis of SPEM which were induced by somatosensory and acoustic information. In the previous study, we used a sinusoidal waveform that could be highly predictable. Since it is wellknown that predictive control has an important role in the normal SPEM, we expect the predictive control to function in non-visually induced SPEM (NVSPEM). We quantitatively analyzed NVSPEM and normal SPEM evoked by pseudorandom target motion in ten human subjects who had no ocular, oculomotor or vestibular disorders. NVSPEM were induced by the following two non-visual targets: 1, subjects' fingertip motion as a somatosensory target ("Somato"), 2, a small loudspeaker (3-cm diameter.) generating white noise with an intensity of about 60 dB (A) as an acoustic target ("Acoustic"). A servo-controlled swing arm of 50cm was used to drive the subject's fingertip and the acoustic target of the small loudspeaker. The horizontal motion of the swing arm was controlled by a personal computer. The pseudorandom target motion was generated by mixing four sinusoids (0.1, 0.2, 0.4, 0.8 Hz) of which the phases were randomly selected and the peak velocities were equally set at 19 deg/s. The mean peak velocity of the target was 26.2 deg/s and the amplitude was limited within 15 deg. Horizontal eye movements were recorded by DC electro-oculography and on an analogue datatape. The experiment was performed for 30 s in complete darkness so that the subjects' fingertip and loudspeaker as such remain invisible to the subject. Signals from the data recorder were smoothed by a low pass analogue filter of 20Hz, after digitization with a sampling frequency of 200 Hz and precision of 12 bits, and stored on a computer. The slow and quick eye movement components, both of which were present in each class of horizontal eye movement investigated, were identified and separated by a computer. Then we developed a method of automatic quantitative analysis of ocular tracking eye movement. Gain and phase values for the smooth pursuit eye movements were obtained in each condition. In the lower frequency area, the gain elicited by the pseudorandom stimulation was lower than the smooth pursuit gain for sinusoidal (predictable) stimulation in all conditions. In the highest frequency, gain values did not differ significantly among the three. For the sinusoidal stimulation, the phase of the smooth component of "Visual" always had a lag and that of "Somato" and "Acoustic" had a lead in lower frequencies. All conditions had a phase shift, decreasing with increasing frequency. For the pseudorandom stimulation the phase of the SPEM had a lead only in the lowest frequency (0.1 Hz). On the other hand, in the NVSPEM the phases of the three lower frequencies had a lead which had a tendency of a larger phase lead with decreasingly frequency. In the highest frequency (0.8 Hz), we could see a short phase lag. These findings support the idea that SPEM and NVSPEM have a mutual or similar physiologic system and overlap part of the anatomical pathway.