To define specific acoustic markers in the speech of Progressive Supranuclear Palsy and Corticobasal (PSP-CBS) spectrum disorders and associate them with underlying pathology.

We obtained semi-structured speech samples during picture descriptions from PSP-CBS patients (n=100), non-fluent/agrammatic primary progressive aphasia patients (naPPA, n=25), and 42 healthy controls (hc). Nine PSP-CBS patients had concomitant naPPA (PSP-CBS+naPPA). Groups matched for age, sex, education, and disease duration. A speech activity detector (SAD) automatically segmented the acoustic signal into speech and non-speech segments. We calculated fundamental frequency (f0 range, which represents perceived pitch range), total speech duration of sample, speech and pause segment durations, and pause rate (pauses per minute). In a subset of patients with cerebrospinal fluid (CSF) information, we associated specific acoustic markers with pTau levels.

PSP-CBS without naPPA had shorter speech segments (1.42±0.48 sec), shorter total speech duration (31.65±14.44 sec), longer pause durations (1.82±1.00 sec) and higher pause rates (45.05±13.27 pauses per minute) compared to hc (1.97±0.60 sec, p<0.001; 50.26±21.69 sec, p<0.001; 0.93±0.44 sec, p<0.001; 33.07±13.34 ppm, p<0.001, respectively). PSP-CBS with naPPA had the most impaired f0 range (3.56±1.12 ST) compared to hc (5.81±1.95 ST, p=0.015). This parallels patients with purely naPPA who had reduced f0 range compared to hc (4.30±1.59 ST, p=0.026). f0 range was associated with pTau CSF levels in all patients (r=-0.42, p=0.004).

Automatically derived acoustic measures characterize speech in PSP-CBS in an objective and reliable manner and these measures are related to markers of likely pathology.