Multiple coordinate frames for reaches revealed through adaptation

Abstract

In what coordinate system are visually guided reaches encoded? We introduce a novel, sensitive method for measuring the adaptive response to reach feedback perturbation, and provide evidence that both polar and Cartesian coordinate systems are used, because adaptive responses can be made to both polar and Cartesian perturbations. Methods: Subjects made center-out reaches on a tabletop with fixed starting point and target distance, with target direction random over trials. Starting point, target, and reach endpoint feedback were shown on a frontoparallel display. During the reach, only the target was shown. After reach completion, fingertip location was shown shifted spatially. The amount of shift was a sinewave over trials. For Cartesian perturbation, on trial t, x_display = x_finger + A sin(2πf_xt) and similarly for y perturbed with frequency f_y. For polar perturbation, gain of displayed relative to actual reach extent was perturbed around 1.0 with a sinewave of frequency f_γ and reach direction was perturbed at frequency f_θ. For Cartesian, f_x=5 and f_y=7 cycle/session; A = 6 mm. For polar, the same frequencies were used, and amplitudes were set to achieve similar maximum perturbations. Bayesian methods were used to detect sinusoidal corrections in response to perturbation and to estimate parameters. Results: Under Cartesian perturbation, subjects corrected for both x and y perturbations; under polar perturbation, independent correction of gain and direction was found. In addition, correction was found for reaches to targets at which feedback was never given. Thus, observers could adapt using either coordinate system, as appropriate. The sinewave-perturbation technique provides a far more sensitive technique for measuring adaptation than the typical step-function technique, allowing the use of perturbations so small that subjects are not consciously aware of them.

Meeting abstract presented at VSS 2012