Sensorimotor adaptation affects detection of and pupil response to motor perturbations

Much recent research on sensorimotor adaptation has focussed on its cognitive side and how adaptation differs when it it cognitively controlled. Here, we investigate more closely what makes different types of perturbations detectable. We used different perturbation schedules to contrast how the magnitude of a motor perturbation and the associated error signal contribute to its detectability. We manipulated perturbation either following an unpredictable de Bruijn sequence, maximising the error signal, or a ramp perturbation, incrementally increasing perturbations and thus minimising error by allowing continuous adaptation. We did this in two different experiments using two perturbed everyday actions, walking with a split-belt speed perturbation (experiment 1) and grasping with a visuo-haptic size mismatch (experiment 2). In each experiment, N=48 participants were asked to judge the relevant dimension, either the relative speed of the treadmill belts, or the relative size of the objects to be grasped, while we measured participants' pupil dilation and investigated if this could serve as a no-report marker of perturbation detection. We found that participants reliably adapted to the incrementally introduced ramp perturbation and concurrently showed reduced perturbation detection relative to de Bruijn sequence perturbations. Tonic pupil-dilation parameters were sensitive to both perturbation magnitude and the type of perturbation, consistent with an effect of uncertainty. Pupil responses also differed depending on the detection response, but not sufficiently to allow trialwise classification. These patterns occurred both in walking and in grasping, with only minor differences, suggesting that the underlying mechanisms may be similar and potentially generalisable across motor actions.