A generative learning model for saccade adaptation.

Consistent with the qualitative observations mentioned above, while type of adaptation is highly significant for the learning rates in every model, frequency shows significance for K and D only in the models that feature double error sampling ( KmDG , for both datasets, KAmDG only for FREQ) but not in those learning just from the last feedback ( KAm , KAmG ).

also in 132,142 other papers