Researchers combined data from 10 different speech collections to study how voice changes relate to disease severity. The group included 890 speakers across five languages who had conditions like Parkinsons disease, cerebral palsy, or amyotrophic lateral sclerosis. They compared these voices to those of healthy controls using a computer method that measures sound features without needing special training on sick voices.
The analysis showed a strong link between specific speech sounds and how serious the disease is. Consonant features correlated with severity scores in a consistent way across all groups. Additionally, changes in nasal sound quality decreased steadily as disease severity increased. The computer model could also clearly separate healthy voices from those of people with severe speech difficulties.
No safety issues were reported because the study used existing voice recordings rather than testing a new drug or procedure. The main reason to be careful is that this research is a meta-analysis of past data, so it shows a connection but does not prove that fixing these sounds will cure the disease. Readers should take from this that speech analysis might help monitor disease progress in places where specialist doctors are not available.