Control of multi-finger pressing: Studied with mechanical and hypothetical control variables

One of the central issues in the field of motor control is how movements are controlled, or more specifically, what is controlled by the central nervous system (CNS). Typically, movements involve the interaction of multiple effectors to produce a desired output. An attractive approach to addressing this problem is studying the interaction of fingers in multi-finger tasks. One issue that arises when studying finger interaction is that single finger force production cannot be performed independently of the other fingers. A behavior known as enslaving occurs when one tries to press with a single finger -- the other fingers also produce force. The dissertation was divided into two parts. The purpose of the first part of the dissertation was to use single- and multi-finger pressing paradigms to investigate: 1) anatomical changes induced in fingers due to changes in configuration and force production of a single finger; 2) force changes of fingers due to an involuntary perturbation of a single finger during a four-finger pressing task; and 3) force changes of fingers due to voluntary pressing of a single finger which followed a task that required establishing a total force stabilizing synergy among the four fingers. The main findings from the first part of the dissertation were: 1) the changes in configuration and force production of a single finger resulted in no significant anatomical changes in the other fingers; 2) an involuntary perturbation of a single finger resulted in the non-perturbed fingers displaying changes in force that acted to stabilize task performance; and 3) following the establishment of a force stabilizing synergy the CNS was able to quickly abolish that synergy such that typical finger enslaving was observed during voluntary pressing with a single finger. The second part of the dissertation investigated whether it is plausible the CNS uses hypothetical control variables (called neural commands) instead of the mechanical variables (force in our case) to control and regulate movements. The neural commands are essentially free of the enslaving effects. Estimating neural commands is non-trivial and a methodological study was performed to determine the best procedure for computing neural commands. It was determined that a neural network model provided the most accurate estimation of neural commands. Analytical inverse optimization (ANIO) and uncontrolled manifold analysis (UCM) were used to compare performance of neural commands to finger forces in a multi-finger pressing task. Based on the experimental results it appears plausible that the CNS may operate in the space of neural commands. The neural commands performed equally as well to the forces in terms of optimality and outperformed forces in terms of the structure of variance. These findings agree with the notion that the CNS controls patterns of muscle activity, not individual muscles and the findings also support a synergistic hierarchical organization of the elemental variables.