How to measure multiple, concurrent tremors?

Context

My mantra, as always, is: if you can measure something, you are a step on the road to being able to control it.

I've posted in the past on using a laptop's pressure pad to measure the tremor in a finger. You can try this online at:
http://www.parkinsonsmeasurement.org/toolBox/tremor.htm

Songfellow has also developed a tool to measure tremor, see thread:
http://neurotalk.psychcentral.com/thread182476.html

Others have used smart phone apps that measure tremor.

One of the problems with my existing approach is that it can only measure the tremor of one finger at a time. This is probably not too much of a problem in the early stages of the disease, when symptoms are minor, but as the disease progresses it misses much of the complexity of the process.

For instance, currently, 8 years post diagnosis, in addition to finger tremor in a number of fingers on both sides of my body, I have on one side whole hand tremor.

You could put an accelerometer on each of the parts and measure them seperately. But the vector of results that that would provide would be difficult to work with.

In an ideal world there would be a single number describing all a person's tremors. This would encourage a "hill climbing" approach to theraputic intervention: try something if it gives an improvement, intensify the intervention and so on, until a maxima is reached.

Some physics

Let me talk you through a simple experiment.

I stand on a bathroom scale. I note the reading on the dial. Let's say it's 70kg. I move my hand upward. What happens to my weight reading? Does it go up? go down? stay the same?

Let's define the "apparent weight" as the reading given on the scale's dial. I'll mean this when I use the term weight below. Well it's possibly counter-intuitive, but the initial change, as the hand begins to accelerate upward is for the weight to increase, once the acceleration is constant the apparent weight is constant, but higher than intially. This is reversed as the hand comes to rest, and it ends up the same weight.

Approximating total tremor

The situation in the above experiment is just a simplification of a tremor. A tremor shows itself as a sequence of pulses. The muliple tremors give a set of superimposed pulses.

The problem becomes simply one of measuring the apparent weight.

There are two cases that may cause a problem:
- movements that have a horizontal component. At first sight these appear to be ignored, but they cause the body to flex, altering the centre of gravity and hence the apparent weight.
- equal and opposite movements will cancel. I don't know how common these are.

A consequence of this approach is that a suitable unit of measurement of tremor is the joule.

I'll be grateful for comments.

John