UC Davis

Motivation: Motor adaptation involves utilizing sensory feedback information to modify motor output in response to movement disturbances. This process can occur along a variety of different behavioral aspects (e.g., timeframe: short or long-term changes) and can involve a single effector (unimanual, e.g., brushing your teeth) or multiple effectors (bimanual, e.g., hammering a nail). Understanding the mechanisms underlying sensorimotor adaptation is crucial in developing solutions to motor disorders (e.g., Huntington’s Disease, pediatric upper limb deformities, assistive robotics). While there is a well-developed scientific literature basis for unimanual motor learning, there is a gap in understanding the interplay of sensory feedback and external factors such as time delay on adaptation. Additionally, little work has centered around bimanual motor adaptation. By investigating these areas, we may contribute to developing therapeutics and procedures to make daily tasks more accessible to people experiencing motor impairments.

Objective: To (1) investigate the spatiotemporal properties of unimanual visuomotor adaptation through examining the effects of different types of visual feedback (either provided throughout the movement or restricted to only the endpoint) on the local and global components of intra-limb generalization and inter-limb transfer; and (2) provide proof of concept for a method toward further investigation of these properties in multi-manual coordination strategies

Methods: Three motor adaptation experiments were conducted using a custom experimental setup that consisted of an LCD monitor mounted horizontally 20 centimeters above a digitized drawing tablet (12 in x 19 in.; Intuos 3, WaCom) and a 3D printed stylus (2.5 cm diameter). Using the stylus, participants made point-to-point reaching movements on the tablet. View of the reaching arm was obstructed by the monitor. All subjects first underwent a baseline reaching block in which they received full visual feedback of a cursor that directly represented the arm position. Subjects in all three experiments were split into two training groups: those receiving full visual feedback (FF) (Decay n = 10; Generalization n = 23; Transfer n = 20) and those receiving visual feedback at the onset of the trial and upon reaching the end point only (EPF) (Decay n = 11; Generalization n = 32; Transfer n = 20). In the training block, subjects experienced a visuomotor rotation such that the cursor path deviated ±30° from the hand path. In the decay experiment (Chapter 2), subjects were probed in the absence of visual feedback after a given time delay that varied from 3 to 120 seconds. In the generalization experiment (Chapter 3), subjects moved into a generalization block where the retention of adaptation was probed at different spatial locations (from 0° to ±135° in 15° increments). In the transfer experiment (Chapter 4), subjects were probed using the untrained (left) limb to targets at the different spatial locations outlined above. These studies culminated in the preliminary bimanual motor learning paradigm outlined in Chapter 5 in which subjects were instructed to place a rectangular cursor (1 cm length, 0.5 cm width) in an oriented rectangular target (1.5 cm length, 0.75 cm width) located 12 cm from the starting position using a KINARM robotic manipulandum.

Results: The temporal decay of the adaptation was largely unaffected by the form of visual feedback; retention after a delay of two minutes for endpoint (73.55 ± 7.60%) and full visual feedback (73.43 ± 7.76%) were not significantly different, as well as the respective time constants (τ_EPF= 25.31, τ_FF= 29.96). Adaptation generalization was best represented by a two-Gaussian model while a single Gaussian model best characterized transfer. Modelling results showed that only the local magnitude of intra-limb generalization was temporally modulated, independent of the feedback condition. In contrast, the spatiotemporal patterns of inter-limb transfer remained near constant across time, again for both types of feedback.

Conclusions: Collectively, these results from Chapters 2-4 suggest that learning mechanisms with different temporal properties underlie the generalization and transfer of visuomotor adaptation. The results of Chapter 5 confirm the preliminary paradigm as a method of future bimanual coordination and control study.