Answers are grounded in this paper and cite their sources.
The self-organizing map (a neural network algorithm of Kohonen) was applied to the spectral pattern recognition of dysphonia. The speech samples, Finnish words with long [a:] uttered by 17 men and 18 women, were perceptually assessed by a group of speech pathologists and the judgments were compared with the locations of the [a:] samples on a self-organized spectral feature map. The map distinguished between normal and dysphonic speech spectra in a statistically significant way. The most apparent spectral feature underlying the differentiation was the relative amount of energy at 1-2 kHz and at 7-9 kHz.
No takes yet. Share an insight, caveat, or question.
Leinonen et al. (1993) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: