Abstract
Goal. Saccadic adaptation restores the correct targeting of visual objects when saccades miss their goals systematically. Here, we assessed global changes in the gain of saccade amplitudes (Rolfs et al., Vision Res., 2010) for continuously varying post-saccadic visual feedback using a Bayesian approach for parameter estimation. Methods and Results. Observers made saccades following a sequence of 304 target steps with amplitudes drawn randomly from a uniform distribution between 4 and 12 degrees with unconstrained saccade directions. During each saccade, an intra-saccadic shift (ISS)—following a continuous sinusoidal variation as a function of the trial number—shifted the target along its vector by -25% (inward) to 25% (outward) of the presaccadic target eccentricity. We modified a method developed for the study of adaptation in manual reach movements (Hudson & Landy, 2012) to model changes in saccadic gain as proportional to the target ISS with four parameters: an overall frequency, a saccade amplitude gain, a time lag, and a global shift to account for intrinsic hypometria of the saccades. We constructed a probability distribution over model parameters after integrating over an internal variable that characterizes the width of the landing error distribution. Subsequently, we obtained a posterior marginal for each parameter by integrating over the remaining three. Using this method, we were able to extract the slow time-scale of the overall ISS variation with a delay of 15 to 30 trials, in spite of the high variability in the saccade landing error and a low overall degree of adaptation. Conclusion. Global saccadic gain adaptation tracks a continuously varying ISS with a similar frequency of variation and a temporal lag of a few trials. Evidence of this fast plasticity can be detected and extracted from a short experimental session in spite of the high variability in the saccade landing error.
Meeting abstract presented at VSS 2014