The article was from 1974 and was a nerve stimulation test vs SFEMG if I understand the difference correctly. As I understand it though I think the results should still apply, at least generally.

At 31C the reduction was 9% which is near the bottom of normal. Warming the muscle 5 degrees to 36C resulted in a 44% reduction which is large. In another test they cooled a muscle from 34.2C to 30C and reduction in response improved from 25% to 4%. From that they conclude that false-negative diagnostic tests can happen due to unduly low temperatures.

Here's the full citation

S. Borenstein, JohnE. Desmedt,
TEMPERATURE AND WEATHER CORRELATES OF MYASTHENIC FATIGUE,
The Lancet,
Volume 304, Issue 7872,
1974,
Pages 63-66,
ISSN 0140-6736,