Body

a 2023 study looking at how machine learning applied to general sound response profiles can be used to categorize Misophonics from other conditions and controls. The authors examined whether specific acoustic features, such as energy in the 2500 to 5500 Hz range and 1 to 16 Hz temporal modulations, could explain why some sounds consistently receive higher aversive ratings. While a few significant correlations were found, none remained after correction for multiple comparisons. Stronger correlations were observed for scrambled, unrecognizable sounds, suggesting that these features may partly account for aversive responses when the sound lacks clear meaning. In contrast, correlations were weak or absent for natural, recognizable sounds. A reanalysis of a larger dataset confirmed this pattern. The authors conclude that although these acoustic properties may contribute to aversiveness in abstract sounds, their influence is likely overshadowed by factors such as meaning and context when the sounds are easily identifiable. (Andermane et., 2023)