Correlation Analysis Between Cortical Structural Features and Acoustic Features in Patients With Parkinson's Disease.

Abstract

Purpose: Parkinson disease (PD) is a progressive neurodegenerative disease. The aim of this study is to investigate the association between acoustic and cortical brain features in Parkinson's disease patients.

Methods: We recruited 19 (eight females, 11 males) Parkinson's disease patients and 19 (eight females, 11 males) healthy subjects to participate in the experiment. Speech samples of three vowels (/i/, /a/, /u/), six plosives (/p/, /pʰ/, /t/, /tʰ/, /k/, /kʰ/), and three voiced consonants (/l/, /m/, /n/) were collected for the experiment, and the acoustic parameters were extracted for fundamental frequency (F0), voice onset time (VOT), voicing onset-vocalic voicing onset (VO-VVO), first formant (F1), second formant (F2), third formant (F3), first bandwidth (B1), second bandwidth (B2), third bandwidth (B3), Jitter, Shimmer, and Harmonics-to-noise ratio (HNR). We also used Ingenia CX 3.0 T to complete the cranial magnetic resonance scanning and did image processing based on the Desikan-Killiany-Tourville Atlas. We assessed the differences in acoustic and neuroimaging parameters between the PD and healthy controls (HCs) groups using the Levene's test (LT), two-sample independent t test (TT), and Mann-Whitney U test (MWUT), and calculated Spearman's bias correlations for acoustic and neuroimaging parameters in the PD and HC groups, respectively.

Results: The results showed that in acoustic features, based on the results of the TT, it can be seen that the F3 of the PD group regarding the vowel /i/ is significantly smaller than that of the HC group. The jitter on the vowel /u/ was significantly higher in the male PD group than in the male HC group. For other acoustic measures, there were no statistically significant differences between the two groups. In the cortical features, the thickness, area, and volume of the cortex were reduced in the vast majority of the brains of the PD patients, however, there is also a small portion of the cortex that appears to be thickened. In the correlation analysis between cortical and acoustic features, F0, F1, F2, F3, B2, B3, VO-VVO, Jitter, HNR, and VOT acoustic parameters showed significant and strong correlation with thickness, area, and volume of cortical sites such as frontal, temporal, entorhinal, fusiform, and precuneus in PD patients, whereas no significant correlation was found in HC group.

Conclusions: This suggests that Parkinson's disease does have an effect on the acoustic and cortical features of the patient's brain, and that there is a correlation between the two features.