In search of the cause in motor control
We search for a cause in everything, it is hard for us to accept that nature doesn't have a plan. Lagrangians/Hamiltonians seem to work very well in mechanics. I have used a similar approach to model arm movements for 'normals' and 'not-normals'.
A good read would be the principle of least action by Feynman (link) to understand the motivation.
The methodology in nutshell is as simple as this, when we move our arm to do a task I assume we are minimizing a cost function. But! as I don't know this cost function I use data and Inverse-optimal control (fancy term for searching the cost function that governs motion). This lets me produce a set of cost functions that govern the motion for different groups of people.
What I find is that there is a significant difference between two groups and there is also some intragroup variability. This says something deep? :) that probably there is nothing deep in motor control. No central principle by which the motion happens. Evolution is a 'nonoptimal' process, therefore, we should accept for these nonprincipled theories in biology.
For those who are pedantic: Read cost functions as cost functionals.
subscribe via RSS