Assessment of the quality of the fits of the parametric phenomenological model to the group data.

(a,c) Each bar is split into the log of the odds ratio of the full model to a drift only model that lacks the sinusoidal component (darker tone of the bars) added to the log of the odds ratio of the drift only model to the noise only model described above (lighter tone of the bars). For all but one condition (Global adaptation, 24 cpd), the full model provides the best account of the data. (b,d) Estimates of the frequency of the periodic component of the oculomotor response, for dataset ORIG (b) and dataset FREQ (d). Error bars are SEM.