![[ Visit Client Website ]](images/banner.JPG)
Thursday, May 7, 2009
Northwest Hall (Chicago Hilton)
1:30 PM
Background:
Children with autism exhibit motor dysfunctions including poor coordination and difficulty with performing/imitating skilled gestures. One of the crucial steps in motor learning is for the brain to form internal models: a mapping between motor commands and the expected visual and proprioceptive sensory feedback. These internal models are the basis for which the brain understands actions of others. However, it is not clear yet how the neural mechanism of internal model is disordered in the autistic brain.
Objectives:
In order to understand a mechanism of the motor disorder in the autistic brain, we examined the differences of the neural representation of internal model between high functioning children with autism (HFA) and typically developing children (TD). If the internal model is a mapping between motor commands and visual sensory feedback, the skill generalizes in Cartesian coordinates; whereas, if it were formed on proprioceptive space, the memory would generalize in the intrinsic coordinates of joints and muscles. The objective of the study was to quantify the property of the generalization of learning of internal model by examining how the learned motor memory could transfer to generalize across arm posture.
Methods:
HFA and TD children performed a reaching task that involved learning an internal model of a novel tool (a robotic arm). Subjects were trained to reach to the forward direction in left workspace while holding a robotic arm; the robotic arm produced a curl force field so that subjects had to learn to adapt their movements to hit the target. Learning was then tested in the left, as well as the right, workspace using a channel that clamped the trajectory error so that the force that the subject produced to compensate the applied force was measured. Generalization of learning to the right workspace was assessed using two directions: one required production of the identical movement in Cartesian (visually-based) coordinates and the other required the movement to be produced in joint coordinates.
Results:
Both HFA and TD adapted to the force similarly (F(1,408)=0.892, p=0.3543). We found the learning generalized in joint coordinates for both HFA and TD. This supports the results in our previous study, which suggests that an internal model relies on an association between proprioception and muscle forces. The new finding here is that HFA generalized in joint coordinates to significantly larger extent than TD (F(1,408)=8.91, p=0.0064).
Conclusions:
More generalization in joint coordinates implies that in learning an internal model of self generated action, the HFA brain builds a stronger than normal association between motor commands and proprioceptive feedback. Because the action perception involves information transformation between the visual feedback and the motor command, the larger than normal reliance on proprioception may explain deficits in action perception in HFA. Furthermore, because the brain of autistic children shows an overgrowth of localized white matter connections, it is possible that this abnormally strong association between motor commands and proprioception in HFA is a correlate of this anatomical feature.
Children with autism exhibit motor dysfunctions including poor coordination and difficulty with performing/imitating skilled gestures. One of the crucial steps in motor learning is for the brain to form internal models: a mapping between motor commands and the expected visual and proprioceptive sensory feedback. These internal models are the basis for which the brain understands actions of others. However, it is not clear yet how the neural mechanism of internal model is disordered in the autistic brain.
Objectives:
In order to understand a mechanism of the motor disorder in the autistic brain, we examined the differences of the neural representation of internal model between high functioning children with autism (HFA) and typically developing children (TD). If the internal model is a mapping between motor commands and visual sensory feedback, the skill generalizes in Cartesian coordinates; whereas, if it were formed on proprioceptive space, the memory would generalize in the intrinsic coordinates of joints and muscles. The objective of the study was to quantify the property of the generalization of learning of internal model by examining how the learned motor memory could transfer to generalize across arm posture.
Methods:
HFA and TD children performed a reaching task that involved learning an internal model of a novel tool (a robotic arm). Subjects were trained to reach to the forward direction in left workspace while holding a robotic arm; the robotic arm produced a curl force field so that subjects had to learn to adapt their movements to hit the target. Learning was then tested in the left, as well as the right, workspace using a channel that clamped the trajectory error so that the force that the subject produced to compensate the applied force was measured. Generalization of learning to the right workspace was assessed using two directions: one required production of the identical movement in Cartesian (visually-based) coordinates and the other required the movement to be produced in joint coordinates.
Results:
Both HFA and TD adapted to the force similarly (F(1,408)=0.892, p=0.3543). We found the learning generalized in joint coordinates for both HFA and TD. This supports the results in our previous study, which suggests that an internal model relies on an association between proprioception and muscle forces. The new finding here is that HFA generalized in joint coordinates to significantly larger extent than TD (F(1,408)=8.91, p=0.0064).
Conclusions:
More generalization in joint coordinates implies that in learning an internal model of self generated action, the HFA brain builds a stronger than normal association between motor commands and proprioceptive feedback. Because the action perception involves information transformation between the visual feedback and the motor command, the larger than normal reliance on proprioception may explain deficits in action perception in HFA. Furthermore, because the brain of autistic children shows an overgrowth of localized white matter connections, it is possible that this abnormally strong association between motor commands and proprioception in HFA is a correlate of this anatomical feature.