Timothy C. Hain, MD. Page last modified: April 16, 2013
This page describes the approach of a lab studying predictive motor control of the head and trunk. It reflects work done by Dr. Wynne Lee (presently retired), and Dr. Timothy Hain, at Chicago Dizziness and Hearing in Chicago Illinois.
Stabilization of the head and neck is a "mission critical" function. The head is a platform for the eyes and ears, and controlling it's motion is critical to interpreting information from these senses. When the head is uncontrolled (such as in a motor vehicle accident), serious injury may result. The head and neck in the upright position are intrinsically unstable and would fall over without stabilization.
Consider for example, control of a nuclear power plant. For the head, it is easy to point out several control interactive systems:
These four types of control differ critically in their timing with respect to a perturbation:
Thus for situations where biomechanical controls are inadaquete, predictive strategies are obviously optimal. The long delay involved in voluntary control means that it will be ineffective for many situations.
There are quite a few possibilities:
Let us consider the situation where the head is being pulled backward by a pulley attached to a weight, and at some time, the weight is dropped (under control of the subject). Clearly one might use any one or a combination of the above mechanisms to stabilize the head. One might even try several out until the most effective one were find. This might result in considerable variability in performance until a methodology for stabilization the head is found by the subject. This example points out two important things related to redundant control:
In other words, in redundant control situations, you don't find "the solution", but you just find one acceptable solution where behavioral goals are met.
Our general approach is to use control system engineering techniques (mathematical modeling) to simulate our data. In general, we set up a redundant control system incorporating what is known about biomechanics, and sensory feedback systems. Generally, we use Matlab/Simulink to implement the system. For an experimental dataset, we find an optimal solution by varying parameters. We then explore the error surface to see if there are families or relationships between parameters that provide equally good solutions.
|Linear sled at Northwestern University Dept. of Physical therapy Human Movement Sciences.|
Our current experimental paradigm is to use a high-performance linear sled to move seated persons on a track. The input in this situation is linear sled motion. As an output, we measure head position with rate sensors and linear acceleration sensors. This provides us the linear and angular position of the head, trunk and neck. By comparing output for predictable and unpredictable motion, we can infer differences in control.
Please direct inquiries to:
Timothy C. Hain, MD
Professor of Neurology, Otolaryngology and Physical therapy,
645 N. Michigan, Suite 1120
Chicago Il 60611
Fax: 773-373-0294, email: email@example.com
References from the Hain/Lee lab: