ASU Learning Sparks

Parkinson's Speech Analysis: Early Detection Insights

Parkinson's disease is a common neurodegenerative condition with motor symptoms that significantly impact quality of life. Detection of early signs of Parkinson's is crucial for potential treatments. Speech analysis can play a role by identifying subtle changes in speech patterns through acoustic analysis. Measures such as voice weakness, imprecise articulation, altered ...

Parkinson's disease is a common neurodegenerative condition with motor symptoms that significantly impact quality of life. Detection of early signs of Parkinson's is crucial for potential treatments. Speech analysis can play a role by identifying subtle changes in speech patterns through acoustic analysis. Measures such as voice weakness, imprecise articulation, altered speaking rate, and rhythm can be among the first signs of Parkinson's disease.

Parkinson’s disease is one of the most common neurodegenerative diseases in the world, second only to Alzheimer’s disease. While people diagnosed with Parkinson’s can live a long life, the motor symptoms can substantially decrease quality of life, robbing those afflicted of their livelihood, independence, and social relationships. While early motor changes may be subtle and seemingly within normal limits, progression of the disease eventually results in a hypokinetic movement disorder. Hypokinesia, which is not enough movement, includes slow movements with restricted range, muscle rigidity, tremor, shuffling gait, masked facial expression, stooped posture, and hypokinetic dysarthria. Hypokinetic dysarthria - when there is not enough movement of the speech articulators -  is the condition patients develop that makes it difficult for people to understand their weak, breathy, mumbled, monotone speech. This significantly impacts quality of life.

Parkinson’s disease pathology in the brain usually occurs slowly, over many years where no obvious symptoms are present. Like for all neurodegenerative diseases, a search for a cure requires that potential treatments be tested in individuals who are experiencing very early disease, when minimal damage to the brain has occurred. Most people will not seek medical attention until their symptoms become problematic and the damage has already commenced to a significant degree. The holy grail in developing a cure relies on very early detection of disease. In Parkinson’s disease, the long pre-clinical phase offers a wide window of opportunity for early detection, if only there were sensitive and reliable clues to set off the proverbial check-engine light. 

And that is where speech analysis may play an important role. Many of the early speech changes are subtle but measurable using acoustic analysis of the speech signal.  And additional detection power is gained when speech can be sampled at regular intervals, over time, to document patterns of speech change through acoustic analysis.

Acoustic analysis refers to digitally decomposing the speech signal into its constituent frequencies and amplitudes, not unlike a prism decomposes white light into its spectrum of colors. Measuring various acoustic components of the speech signal allows for tracking and measuring changes in one’s speech over time. For someone who may be experiencing very early Parkinson’s disease, the acoustic analyses might show changes in voice that reflect an increasingly weak, breathy voice that may have a vocal tremor. Other measures may show reduced precision in producing speech sounds (imprecise articulation).  And other measures can capture speaking rate and rhythm changes and reductions in the melody of speech.   

Taken together with other clinical results, speech changes can be a harbinger of early disease like Parkinson’s.