I believe he said he found several people with tinnitus and let them try, all with similar results. So I Would guess the sample size is probably more like three or four.