Neural Modeling of Imitation Deficits

This abstract addresses the question of human imitation through convergent evidence from neuroscience. We look at deficits in imitation following brain lesion, such as apraxia. We believe that looking at how imitation is impaired can unveil its underlying principles. We also take inspiration from numerous brain imaging studies to ground the functional architecture and information flow of our model. In the end we will use findings from monkey brain neurophysiological studies to implement the details of our processing modules. We aim at developing a model of visuo-motor imitation using tools from neural networks and dynamical systems. The model should account for some of the behaviors observed in faulty imitation. At this stage we have implemented a somatotopically organized neural network with probabilistically impaired transfer of information that simulates lesions at the level of the parietal cortex (a brain center for sensorimotor integration). To validate the model against human motion experimental data, we conduct, in collaboration with the Geneva University Hospital (HUG), kinematic studies with brain damaged adults specifically disabled in gesture imitation. The model will motivate the realization of computer-based rehabilitation tools.