Voice onset time as a specific classifier for the PIGD subtype in Parkinson's Disease, independent of non-motor symptom severity

Background Parkinson’s disease (PD) is a heterogeneous disorder where the postural instability/gait difficulty (PIGD) subtype carries a poorer prognosis. This study investigates acoustic biomarkers to examine the complex relationship between motor deficits and non-motor symptoms (NMS) in PD subtypes. We hypothesized that distinct speech features—articulatory timing (Voice Onset Time, VOT) and phonatory stability—could serve different roles in characterizing the PIGD phenotype. Methods This study enrolled 73 patients with early Parkinson's disease (PD) and 30 healthy controls (HC). PD patients were subtyped as tremor-dominant (TD) or postural instability/gait difficulty (PIGD) based on their MDS-UPDRS scores. Acoustic parameters, including voice onset time (VOT) and formant centralization ratio (FCR), were collected and analyzed. Statistical analyses involved the Kruskal-Wallis test, Spearman correlation, as well as linear and binary logistic regression models. Results The PIGD group exhibited significantly prolonged /ta/-VOT and greater vocal jitter. A notable functional dissociation was observed within the PIGD subtype: phonatory instability (jitter) correlated significantly with higher anxiety (HAMA, R = 0.509, P = 0.022) and depression (HAMD, R = 0.449, P = 0.047), and speech dysrhythmia (pause duration) with lower cognitive scores (MoCA, R=-0.522, P = 0.022). In contrast, prolonged VOT showed no such correlations. A logistic regression model identified /ta/-VOT as a key significant predictor (P = 0.044) for discriminating PIGD from TD subtypes, achieving an AUC of 0.821 (sensitivity = 91.7%, specificity = 70.6%). Conclusion We demonstrate a functional dissociation of acoustic biomarkers in PD. Prolonged VOT is a strong and specific biomarker for classifying the PIGD motor subtype, likely reflecting a core articulatory timing deficit. Concurrently, phonatory instability and dysrhythmia serve as independent indicators of the NMS burden (anxiety, depression, cognitive decline) within this subtype. This non-invasive, multi-faceted speech analysis offers a potential approach to improve subtype classification and objectively track the distinct motor and non-motor components of PIGD.