Acoustic measures in natural speech of progressive supranuclear palsy and corticobasal spectrum disorders

Abstract

AbstractBackgroundProgressive supranuclear palsy and corticobasal syndrome (PSP‐CBS) are neurodegenerative disorders involving motor and cognitive deficits. These patients may have slowed and dysfluent speech. However, few studies have investigated specific speech and language deficits in these disorders, although some patients may have non‐fluent/agrammatic primary progressive aphasia (naPPA). In our study, we used novel, automated speech analysis methods to quantify acoustic properties of natural speech in PSP‐CBS. We hypothesized patients would exhibit slowed speech and longer pauses, and predicted that prosodic changes would be related to underlying cognitive and motor impairments. Because PSP and CBS are tauopathies, we related their acoustic performance to cerebrospinal fluid (CSF) analytes associated with Tau pathology.MethodWe obtained digitized samples of natural speech with the Cookie Theft picture description task from PSP‐CBS patients (n = 95), naPPA patients (n = 25) and healthy controls (HC, n = 41). Eight PSP‐CBS patients had concomitant naPPA (PSP‐CBS+naPPA). The groups were matched for age, sex, education, and disease duration. A speech activity detector program automatically segmented the acoustic signal into speech and silent pause segments. We automatically calculated the pitch range by measuring fundamental frequency (f0, semitones), total speech duration, mean speech and pause segment durations, and pause rate (pauses per minute). In a subset of patients with available CSF analytes, we related acoustic markers with levels of phosphorylated tau (pTau).ResultPSP‐CBS without naPPA had shorter mean speech segments (1.42 ± 0.48 sec), shorter total speech durations (31.77 ± 14.66 sec), longer mean pause segments (1.82 ± 1.01 sec) and higher pause rates (45.10 ± 13.43 pauses per minute) compared to HC (2.00 ± 0.57 sec, p < .001; 51.08 ± 21.30 sec, p < .001; 0.93 ± 0.44 sec, p < .001; 31.61 ± 9.51 ppm, p < .001, respectively). PSP‐CBS only differed from naPPA for pause rate (PSP‐CBS, 45.10 ± 13.43 ppm; naPPA, 57.81 ± 19.95 ppm; p = .026). PSP‐CBS+naPPA patients were most impaired for f0 range (3.56 ± 1.21 ST) compared to HC (5.87 ± 1.94 ST, p = .021), paralleling naPPA patients, who had reduced f0 range compared to HC (4.30 ± 1.59 ST, p = .02). f0 range was associated with pTau levels in all patients (r = ‐0.43, p = .004).ConclusionPatients with PSP‐CBS have impaired speech as quantified by automated analysis of acoustic measures of speech. These measures are related to pTau, a marker of likely pathology.