The New Jersey Institute of Technology's
Electronic Theses & Dissertations Project

Title:	Characterizing motor control signals in the spinal cord
Author:	Guo, Yi
View Online:	njit-etd2014-091 ([xviii], 97 pages ~ 29.6 MB pdf)
Department:	Department of Biomedical Engineering
Degree:	Doctor of Philosophy
Program:	Biomedical Engineering
Document Type:	Dissertation
Advisory Committee:	Foulds, Richard A. (Committee chair) Sahin, Mesut (Committee member) Adamovich, Sergei (Committee member) Berlin, Joshua R. (Committee member) Alvarez, Tara L. (Committee member) Tunik, Eugene (Committee member)
Date:	2014-05
Keywords:	Central command signals Forearm Motor function control
Availability:	Unrestricted
Abstract:	The main goal of this project is to develop a rodent model to study the central command signals generated in the brain and spinal cord for the control of motor function in the forearms. The nature of the central command signal has been debated for many decades with only limited progress. This thesis presents a project that investigated this problem using novel techniques. Rats are instrumented to record the control signals in their spinal cord while they are performing lever press task they are trained in. A haptic interface and wireless neural data amplifier system simultaneously collects dynamic and neural data. Isometric force is predicted from force signal using a combination of time-frequency analysis, Principle component analysis and linear filters. Neural-force mapping obtained at one location are subsequently applied to isometric data recorded at other locations. Prediction errors exhibited negative relationship with the isometric position at upper half of movement range. This suggests the presence of restorative forces which are consistent with positional feedback at spinal level. The animal also appears to become unstable in the lower half of their movement ranges, likely caused by a transition from bipedal to quadruped posture. The presence of local feedback and ability for animals to plan postures that are unstable in absence of external forces suggest that descending signal is a reference trajectory planned using internal models. This has important consequences in design of neuroprosthetic actuators: Inverse dynamic models of patient limbs and local positional feedbacks can improve their performance.