Why movements drift in the dark: Passive and active mechanisms of error accumulation


When people make multiple movements without visual feedback, error accumulates rapidly: within five to ten open-loop movements, a person’s hand may have drifted several centimetres away from the target. Are people aligning their (misestimated) felt hand with the seen target (i.e. active accumulation) (Smeets et al. 2006), or is the hand estimate poor enough that substantial error can accumulate before the error is detected and counteracted (i.e. passive accumulation)? To determine which of these mechanisms causes drift, we asked participants to reach to a visual target and we provided them with no error feedback, veridical error feedback, or false error feedback. The false feedback consisted of a visual error signal randomly drawn from a normal distribution (mean = zero error; SD = derived from participant’s performance in the veridical feedback condition). In other words, the false feedback indicated to the participant that their average performance was accurate, so no systematic movement correction was required. If drift accumulates under false feedback conditions, it would suggest that drift can accumulate without active correction by the participant. We compared movement drift with false feedback to movement drift with no feedback, and we observed comparable amounts of drift in the two conditions. In a second experiment, we manipulated the accuracy constraints of the task instead of the veridicality of the feedback. In one condition, participants reached to horizontal line targets (only amplitude accuracy required), while in another condition they reached to circle targets (amplitude and direction accuracy required). In both conditions visual feedback of amplitude error but not direction error was provided. We observed comparable amounts of drift in the two conditions. The results from both experiments lead us to conclude that passive error accumulation is a viable mechanism for movement drift when people reach without vision of their hand.

Acknowledgments: The research group is supported by Grant 2009SGR00308 from the Catalan government. Brendan D. Cameron was supported by a Juan de la Cierva fellowship from the Spanish Government and the Marie Curie fellowship PCIG13-GA-2013-618407. Joan López-Moliner was supported by an ICREA Academic Distinguished Professorship award.