Latent State-Trait and Latent Growth Curve Modeling of Smooth Pursuit Eye Movements

Smooth pursuit eye movement (SPEM) performance has previously been shown to have good reliability. While quantifying the relative amounts of reliable trait and state influences on SPEM is relevant to different lines of research including individual differences, clinical and experimental research, this has not yet been done. Here, we apply latent state-trait (LST) theory to SPEM for the first time to examine reliability and to explicitly decompose trait and situational variance. SPEM tasks with sinusoidal and triangular movement patterns were performed by N = 163 healthy participants at three measurement occasions. LST and latent growth curve (LGC) modeling was used to calculate model-based reliability and to distinguish reliable trait variance (consistency) and variance due to influences of the situation and of the person × situation interaction (occasion specificity). We found mostly excellent reliabilities (0.86-0.98), except for the intra-individual standard deviation of root mean square error (RMSE) in both SPEM tasks where reliability was good (0.70-0.74). Consistencies and occasion specificities indicated that a higher proportion of variance was due to trait influences (62% on average) than due to situational influences (26% on average). There were mostly no changes on trait level over time. We conclude that SPEM performance is highly reliable and mainly reflects relatively stable trait components, but is also characterized by substantial state influences. Overall, these findings further support the use of SPEM in individual differences studies. However, potential state influences should be considered more explicitly in future studies examining SPEM.