A paper by Pal and Goetz [1] describes ways to measure the slow movements of PwP.

"... numbers generated must have a direct bearing on the clinical state under question, because an improvement in a movement variable, even if statistically significant, has little importance if it does not correlate strongly with clinically appreciated improvements.

"... such tools have the potential to register almost limitless information and over-sampling can lead to data sets difficult to interpret.

"... the learning effects need to be considered, as well as confounding influences of cognitive function, motivation, tremor, and the topographic distribution of the motor deficit being measured.

"... rating scales and technology-based tools s are likely to be used in research and clinical care as complementary strategies for tracking disease progression and response to treatment.

"... it is likely that rapid advances in bradykinesia monitoring will be achieved."

Apparently, for years I've been a quantitative digitographer!

For my attempts at online measurement tools see:
http://www.parkinsonsmeasurement.org...eToSideTap.htm
http://www.parkinsonsmeasurement.org/toolBox/tremor.htm
http://www.parkinsonsmeasurement.org/PDMeasure/

The learning effect mentioned above is important. It refers to the process by which the act of repeatedly taking a test leads to better scores without improving the underlying condition. Today, a user of PDMeasure, to whom my thanks, reported that he/she (users are anonymous) had increased their side to side tap test scores by moving their torso as well as their arms, hands and fingers. I wonder whether all clinical trials take the learning effect into account?

Reference

[1] "Assessing Bradykinesia in Parkinsonian Disorders"
Gian Pal and Christopher G. Goetz
Front Neurol. 2013; 4:54
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3669893/

John