Acoustic Evaluation of Nasality in Cerebellar Syndromes

M. Novotný, Jan Rusz, K. Spálenka, Jiří Klempíř, D. Horáková, Evžen RůžičkaInterspeech 2017

TL;DR

This study investigates nasality in individuals with cerebellar disorders, finding that while overt hypernasality isn't common, increased nasality fluctuation is a significant indicator of impaired velopharyngeal control. This fluctuation, measured using 1/3-octave spectra analysis, correlates with cerebellar dysfunction, suggesting it could be a novel marker for these conditions.

What Problem Does This Paper Address?

Cerebellar damage significantly impacts motor control, leading to ataxic dysarthria, a speech disorder characterized by articulation inaccuracies, prosodic issues, and phonatory-prosodic insufficiency. One manifestation of this can be inconsistent velopharyngeal closure, resulting in variable nasality. However, previous research on nasality in cerebellar disorders is limited, partly due to the subjective and unreliable nature of perceptual assessment and the subtle presentation of nasality disturbances. This paper aims to provide an objective acoustic measure of nasality in cerebellar syndromes to better understand its presence and relationship with neurological deficits.

Methods

The researchers employed a 1/3-octave spectra analysis, a previously validated objective acoustic method, to measure nasality. This technique analyzes the energy distribution across different frequency bands in speech signals. Specifically, they focused on the energy in the 1000 Hz frequency band as a marker for nasality and calculated both the average nasality level (EFn mean) and nasality fluctuation (EFn SD) across sustained vowel /i/ recordings. The EFn SD was particularly emphasized as a measure of inconsistent velopharyngeal control. Statistical analyses included Kruskal-Wallis tests for group comparisons and Spearman correlations to assess relationships between acoustic measures and clinical scores. A Wald task was used to establish a threshold for classifying utterances as healthy, indecisive, or pathological for nasality fluctuation.

Datasets

The study analyzed speech recordings from 74 Czech native speakers divided into four groups:

  • Multiple System Atrophy (MSA): 20 participants (9 men, 11 women)
  • Cerebellar Ataxia (CA): 13 participants (7 men, 6 women)
  • Multiple Sclerosis (MS): 20 participants (7 men, 13 women)
  • Healthy Controls (HC): 20 participants (9 men, 11 women)

Participants were recruited from previous studies investigating speech in neurological conditions. Clinical data, including disease duration and specific neurological scales (NNIPPS, SARA, EDSS), were also collected.

Key Insights and Ideas

  • Nasality Fluctuation, Not Just Level: The study highlights that the key acoustic indicator of velopharyngeal dysfunction in cerebellar syndromes isn't necessarily a consistently high level of nasality (hypernasality), but rather fluctuations in nasality. This suggests a problem with the dynamic control of the soft palate.
  • Cerebellar Impact on Velopharyngeal Control: The findings strongly suggest that damage to the cerebellum, even if not directly controlling the velopharyngeal mechanism, can indirectly impair its precise and coordinated function, leading to speech disturbances like nasality fluctuation.
  • Potential Biomarker: Nasality fluctuation, as measured by EFn SD, shows promise as a sensitive and objective marker for cerebellar dysfunction, particularly in conditions like MSA.

Key Results

  • Increased Nasality Fluctuation: The study found significantly increased nasality fluctuation (EFn SD) in patients with cerebellar disorders compared to healthy controls. Specifically, 65% of MSA, 43% of CA, and 30% of MS participants showed increased fluctuation, compared to 15% of HC speakers.
  • MSA vs. HC Difference: A statistically significant difference was found between the MSA and HC groups for EFn SD (p < 0.001), with a large effect size (d = 1.39).
  • Correlation with Clinical Markers: A significant positive correlation was observed between nasality fluctuation (EFn SD) and the NNIPPS cerebellar subscore in MSA patients (r = 0.51, p < 0.05). This indicates that greater cerebellar impairment is associated with more pronounced nasality fluctuations.
  • No Overt Hypernasality: The study did not detect a clear cut-off threshold for increased average nasality (EFn mean), suggesting that overt hypernasality is not a consistent feature of these conditions.

Limitations and Future Work

  • No Aerodynamic Measurements: The authors acknowledge the lack of aerodynamic measurements, which would provide a more direct assessment of velopharyngeal function.
  • Age and Disease Onset Mismatch: The study groups could not be perfectly age-matched due to differing disease onset ages.
  • MSA-P Inclusion: The MSA group included participants with the parkinsonian subtype (MSA-P), which can have its own speech characteristics (hypokinetic-rigid syndrome). While efforts were made to account for this, it remains a potential confound.
  • Sustained Vowels Only: The current acoustic analysis is limited to sustained vowels and cannot be easily applied to continuous speech, which is more ecologically valid.

Future work could involve extending the analysis to continuous speech and incorporating aerodynamic measures to further validate these findings. Investigating the specific neural pathways linking cerebellar damage to velopharyngeal control would also be valuable.

Summary updated

BibTeX
@inproceedings{novotny2017acoustic,
  author = {M. Novotný and Jan Rusz and K. Spálenka and Jiří Klempíř and D. Horáková and Evžen Růžička},
  title = {{Acoustic Evaluation of Nasality in Cerebellar Syndromes}},
  booktitle = {Interspeech},
  year = {2017},
}
Figure