The effect of target jumps on implicit adaptation of reaching

Abstract

Sensorimotor adaptation is an implicit motor learning process that corrects and compensates for motor errors, contributing to the successful execution of everyday movements. Previous research into the control of reaching movements suggests that the implicit motor adaptation system is sensitive to two types of error signals - sensory prediction error (SPE), defined as the difference between where visual feedback of hand position is expected to appear and where it actually appears, and target error (TE), defined as the difference between the cursor position and final target position (Leow et al., 2018). Traditional adaptation experiments confound these two error signals, with the SPE and TE typically being the same. This prompted the introduction of shifting targets during reaching movements, a potential means of isolating TE and its effects. We have previously assumed that this method is able to induce a TE by having the target jump to a new location approximately one centimeter into the reach. However, it is unclear whether the motor system processes the target jump as a target error or if the jump introduces uncertainty regarding the SPE, as it is no longer anchored to the original target location. We refer to this latter perspective as a “lowlevel processing error”. Here, we observed the effects of target jumps during three different conditions designed to assess adaptive responses to three combinations of SPE and TE conditions, “amplified”, “consistent”, and “opposite” in reference to whether the target jump amplified the TE, remained consistent throughout the trial, or elicited a TE in opposition to the SPE. This experiment aimed to determine whether target jumps elicit a TE, independent of SPE, or if they result in non-specific attenuation of adaptation due to low-level processing errors. Support for the former hypothesis would result in a graded amount of adaptation, with the most observed during amplified trials and the least during opposite trials. In contrast, if target jumps result in low-level processing errors, then adaptation should be attenuated a similar amount during amplified and opposite trials, as the absolute target jump magnitudes were equivalent for these two trial types. Our results revealed a trend towards greater amounts of adaptation during the consistent condition, which had no target jump, compared with less adaptation in the amplified and opposite conditions. These results argue against the interpretation that target jumps elicit a “pure” TE, and suggest that they may have non-specific effects on adaptation from SPE, potentially due to a lowlevel processing error.