A neural model of human prehension
Althea R. Iberall · 1987
BEHAVIORAL STUDIES HAVE SHOWN THAT AS THE HUMAN HAND REACHES OUT TO GRASP AN OBJECT, IT PRESHAPES FOR THE ANTICIPATED INTERACTION. AS IT GETS CLOSE TO THE OBJECT, IT ENCLOSES IT. IN ORDER TO EVALUATE SUCH STUDIES, A MODEL IS NEEDED WHICH CAN BE USED TO ANALYZE GRASPING BEHAVIOR CHARACTERIS- TICS AS THEY RELATE TO OBJECT AND TASK PROPERTIES. WE PRESENT HERE A FRAME WORK FOR STUDYING HUMAN PREHENSILE POSTURES AND MOVEMENTS, WHICH IS BASED ON ARCHITECTURAL FEATURES OF THE HAND. WHILE A GRASPING TASK WILL INVOLVE MANY FINGERS, WE NOTE THAT FINGERS GROUP TOGETHER TO APPLY AN OPPOSING FORCE AGAINST OTHER FINGERS OR AGAINST TASK TORQUES. WE CALL THESE GROUP- INGS VIRTUAL FINGERS, AND WE CALL THE COORDINATE FRAME IN WHICH THEY MOVE AN OPPOSITION SPACE. THIS SPACE PROVIDES A DESCRIPTION OF THE WAYS VIRTUAL FINGERS CAN APPLY OPPOSITIONS, AS WELL AS A METHOD FOR DEFINING THE RELE- VANT OBJECT/TASK PROPERTIES. IN ORDER TO TEST THIS STYLE OF ANALYSIS, WE PERFORMED BEHAVIORAL EXPERIMENTS, AND SHOW HERE HOW OUR FRAMEWORK CAN BE USED TO MOTOR AND SENSORY PSYCHOLOGISTS AS AN EFFECTIVE METHOD FOR ANALYZ- ING GRASPING BEHAVIORS. TO ACCOUNT FOR THE COORDINATION AND CONTROL OF THESE COMPLEX MOVEMENTS, WE USE SCHEMA THEORY. A PERCEPTUAL SCHEMA MUST CHOOSE WHERE TO PLACE THE HAND RELATIVE TO THE OBJECT. WE SHOW HOW SUCH A CHOICE COULD BE PERFORMED BY A NEURAL NETWORK, SIMULATED ON A COMPUTER HERE USING AN AMARI-ARBIB COOP