*Presented at MDS Congress 2026

Introduction: Speech impairment is common in PD and PSP and reflects broader motor system dysfunction. Digital speech measures may sensitively capture disease-related change and provide clinically meaningful markers of functional communication deficits, including intelligibility.

Methods: Speech was recorded in-clinic using the Mili mobile app and speech samples were processed with SIGMA extracting an automatically generated intelligibility score [1] from reading and acoustic features.

Tasks included:

• Maximum phonation /a/

• Sustained vowels /i, a, u/

• DDK of /pataka/ and /kakaka/

• Reading

• Picture description

Results- App Usability: Participants rated app usability on a 1 to 5 scale (higher is better). Usability was high across groups, with most ratings ranging between 4 and 5 (Fig. 1).

Results- Speech Characteristics: Speech intelligibility differed across all groups (HC: 0.94 ± 0.07; PD: 0.88 ± 0.1; PSP: 0.62 ± 0.29), with PSP showing the greatest reduction (Fig. 2). Articulation rate strongly correlated with intelligibility across groups, highlighting its primary role in reduced intelligibility (Fig. 2, right).

Acoustic measures further differentiated groups: PSP showed overall more slowness than PD (reduced DDK and reading rates) and more pauses during connected speech. Vowel Space is reduced in PD compared to HC and PSP (Fig. 3), quantified via Vowel Articulation Index. (HC 1.02 ± 0.13), PD 0.96 ± 0.14, PSP 1.01 ± 0.15). A shorter maximum phonation time is in line with expected reduced breath support in PSP (Fig. 4).

Conclusion: App-based speech assessment is feasible in PD and PSP and detects objective speech impairment, with more severe intelligibility reduction in PSP than PD, particularly driven by articulation rate. These findings support scalable speech monitoring for disease tracking, patient stratification, and differential characterization.