From Cognitive Representation to Technical Synthesis of Manual Action

2008-03 till 2012-10
Research Areas: 

We investigated hand kinematics and mental representations of grasping movements directed towards real and virtual spherical objects systematically varying in size. Results suggest that grasping movements are influenced by object size at an early stage of the movement for real and virtual objects. The analyses of mental representations (via SDA) and of motor synergies (via PCA) reveal a separation of the smallest three objects from the larger ones, pointing towards a conceptual influence on the grasping movement.

Methods and Research Questions: 

Cognitive processing is of high relevance for manual action, and hand movements are strongly influenced by object representations. Our aim in this project is to investigate how mental representations of human hands and related actions, such as grasping, and the biomechanical measures of these actions relate to each other. In manual action, the large number of degrees of freedom in the human hand has to be controlled efficiently, and it has been proposed that hand control is organized in a modular way. Studies have supported the notion of motor synergies, degrees of freedom coupled into functional groups, in the control of manual action. Principal component analysis has proved to be a useful tool to extract such motor synergies in human motor action, including manual action. In this project, we investigate motor synergies and mental representations of grasping real and virtual objects that have no distinctive features and vary linearly in selected dimensions, such as size or shape. … In the first study, we investigated the kinematics of grasping real versus virtual spherical objects of different size. The kinematic data was analyzed using principal component analysis (PCA) in order to extract motor synergies and to determine invariant movement characteristics of real and virtual object grasping. Ten subjects grasped spherical objects while wearing a wireless data glove that allowed for recording whole hand kinematics with 22 degrees of freedom. Hand trajectories were tracked using a VICON motion capture system with 14 infrared cameras that monitored three retro-reflective markers on the back of the data glove. In the first experiment, eight white plastic spheres varying linearly in size were presented on a holding device. In the second experiment, corresponding images were displayed on a computer screen behind the empty holding device, and subjects were instructed to imagine the displayed object lying on the holding device. Based on pooled joint angle time courses, we computed motor synergies via PCA for each experiment separately and compared the results of the two experiments. In a third experiment, mental representations of grasping movements directed towards the same virtual objects were analyzed using structure dimensional analysis (SDA) which is based on a hierarchical sorting paradigm.


Results of the PCA indicate that the grasping movements are influenced by object size at an early stage of the movement in both the real and the virtual case. Clusters that mark objects in principal component space can be distinguished, especially the three smallest objects are clearly separated from medium and large objects. For the final grasping posture, more than 70% of the variance of real and virtual grasping movements can be described by the first three principal components, indicating strong linear relationships between the involved joints. The subjects used similar motor synergies during real and virtual grasping which reflect the physical properties of the grasped objects. These findings allow for a compact description of grasping movements in terms of motor synergies. Results of the SDA also show a separation between the three smallest objects and the larger ones, which points towards a conceptual influence on the grasping movement that might relate to a general cognitive distinction between precision and power grasp in motor planning.