Intra-individual variability of saccadic velocity measured with the infrared reflection and magnetic scleral search coil methods.
BACKGROUND: The infrared (IR) and the magnetic scleral search coil (MSC) systems for eye tracking were studied with regard to the intra-individual variability in saccadic eye movement recordings. METHOD: Three healthy subjects performed similar saccadic eye movement tasks at five different occasions with both the IR (Orbit XY-1000) and the MSC (Skalar Medical) techniques. The maximum velocity (VMAX) and slope constant (C) of the main sequence plots were analyzed with regard to the coefficient of variation (CV) and the intraclass correlation coefficient (Ricc). In addition, the possible reasons for variability in the IR recordings, especially different causes for noise, were analyzed and discussed. RESULTS: The main sequence data showed intra-individual variation with both recording systems, but the coefficient of variation was higher for VMAX with the IR compared to the MSC method. Ricc analysis showed that 36% of the variance of VMAX and 49% of the variance of C resulted from intra-individual variability in recordings of the IR system. The corresponding results for the MSC recordings regarding VMAX and C were 48% and 88%. CONCLUSIONS: Saccadic eye movement recordings yielded a larger intra-individual variability with the IR system than with the MSC system. The effect that the MSC annulus may have on the ocular motor command signal and the possible low pass filter caused by the coil slipping on the surface of the eye may partly explain the relatively lower velocity in the MSC recordings. Also, noise in the IR recordings induces peaks of eye velocity, which can be reduced considerably by filtering. However, the variability in the recordings, which was larger in the IR than in the MSC recordings, did not seem to be decreased by filtering. The basic level of noise in the recordings was not clearly associated with the amount of reduction of VMAX when the IR recordings were filtered. We suggest that artefacts of the saccadic signal, which can be related to changes in the reflecting surface of the eyes and eyelids, are important factors for explaining the variability and high-velocity peaks in the IR recordings. Lighting conditions was confirmed as a cause for noise, but temperature and air humidity changes in the goggles were not suspected to influence data in the normal experimental setting. Although noise, shortcomings of the recording technique and procedure may offer explanations for the intra-individual variability, the calibration procedure and changes in attention and fatigue of the subject should also be considered.