The Effects of Age-related Differences in State Estimation on Sensorimotor Control of the Arm in School-age Children

Thumbnail Image
Publication or External Link
King, Bradley Ross
Clark, Jane E
Previous research examining sensorimotor control of arm movements in school-age children has demonstrated age-related improvements in performance. A unifying, mechanistic explanation of these improvements is currently lacking. This dissertation systematically examined the processes involved in sensorimotor control of the arm to investigate the hypothesis that improvements in performance can be attributed, in part, to developmental changes in state estimation, defined as estimates computed by the central nervous system (CNS) that specify current and future hand positions and velocities (i.e., hand `state'). A series of behavioral experiments were employed in which 5- to 12-year-old children and adults executed goal-directed arm movements. Experiment 1 demonstrated that improvements in proprioceptive functioning resulted in an increased contribution of proprioception to the multisensory estimate of hand position, suggesting that the CNS of children flexibly integrates redundant sensorimotor feedback based on the accuracy of the individual inputs. Experiment 2 demonstrated that improvements in proprioceptive functioning for localizing initial hand position reduced the directional variability of goal-directed reaching, suggesting that improvements in static state estimation contribute to the age-related improvements in performance. Relying on sensory feedback to provide estimates of hand state during movement execution can result in erroneous movement trajectories due to delays in sensory processing. Research in adults has suggested that the CNS circumvents these delays by integrating sensory feedback with predictions of future hand states (i.e., dynamic state estimation), a finding that has not been investigated in children. Experiment 3 demonstrated that young children utilized delayed and unreliable state estimates to make on-line trajectory modifications, resulting in poor sensorimotor performance. Last, Experiment 4 hypothesized that if improvements in state estimation drive improvements in sensorimotor performance, then exposure to a perturbation that simulated the delayed and unreliable dynamic state estimation in young children would cause the adults to perform similarly to the young children (i.e., eliminating age-related improvements in performance). Results from this study were equivocal. Collectively, the results from these experiments: 1) characterized a developmental trajectory of state estimation across 5- to 12-year-old children; and, 2) demonstrated that the development of state estimation is one mechanism underlying the age-related improvements in sensorimotor performance.