I am an active experimenter and will share what I can.

First and foremost, because it is current and because it could skew any other results, is the matter of electrolyte imbalence. In my case, a routine physical two weeks ago showed low potassium. I began with just 200 mg and was startled by the effect. Twenty-four hours before, I had ended my day as I had been doing increasingly frequently for two years by crawling to bed at midnight - primarily due to balance and lower leg weakness plus freezing. Less than 24 hrs after the first potassium I walked to bed more or less normally and have not crawled since. Further, I have reduced my Requip dosage from 32 mg/day (too much) to 24 mg/day and Sinemet CR (200/50) from 4x/day to 3x. Also, my blood pressure is under better control and my early morning dystonia in my left foot was missing this morning. Pretty good for two weeks.

Some points about this- 1) Despite my study I made the basic mistake of assuming that I was getting sufficient electrolytes. My neuro seemingly was too. Only my sturdy GP was sensible enough to check the basics. So, beware of initial assumptions. 2) The symptoms of electrolyte imbalance are uncannily like those of advanced PD as are some of the other nutrients (B12 in particular). 3) Since it only required 200 mg to make a difference and the daily amount suggested is 20x that, one cannot assume a smooth curve as things gradually worsen. Instead, there is a threshold effect. In my case, that means that for the last several months and maybe over a year I was slipping above and below that threshold. I would have a bad day for no obvious reason just because my diet had slipped the day before. A difficult situation.

Finally, while one might test this at home, caution is advised. If you have any kidney problems (and don't assume you would know) potassium levels can build up to fatal levels. This is one where you need a doc monitoring you system since your heart can stop.

I'll return to topic in a following post, but keep in mind that some of my data may be screwed.