I was able to access the actual paper, rather than merely the abstract in Science Daily.

The graph of their results does not visually appear to show much correlation. The connection is "statistically significant", but not terribly compelling.

The measure of musical aptitude was a test called "Advanced Measure of Music Audiation" (AMMA, in one place it's spelled "Audition") "Short melodies performed on a piano were played in pairs and participants had to state whether the two melodies are the same, tonally different or rhythmically different."

The measure of hallucination proness was a test called the "Revised Launay-Slade Hallucination Scale" (LSHS). It "is a widely used self-report measure of hallucination proneness. It is sensitive in detecting auditory verbal hallucinations (AVH) in a healthy population. The 12 items of the LSHS scale describe clinical and subclinical forms of auditory and visual hallucination-like experiences." Note that term: "self-report". The people being tested judge themselves.

I was amused to see that the participants "were non-musicians (had less than 10 years of professional musical training)". I thought of myself as a musician a long time before I had "10 years of professional musical training". I won't comment whether that was right or wrong; only that it seems like a fairly high bar. I would be interested to see the actual levels of "professional musical training)" the participants actually had.

BB,
NightWing