ABSTRACT
Objective: To compare the variability of acoustic measures of speech presencial and remote recording during the assessment of individuals with Parkinson's Disease (PD). Background: The Covid-19 pandemic scenario has restricted inperson speech recordings. Thus, telephone recording has become an option as it is simple to drive, reduces time and investment, and can be done remotely from the owners' home. Methods: This study was approved by the Ethics Committee and all participants signed an informed consent form. The in-person evaluation was carried out in a silent environment, whose maximum noise level did not exceed 40 dB, using the Karsect HT2® microphone earphone placed five centimeters from the participant's mouth and connected to a recorder. The remote data were collected using a phone and a call recording application. Each individual was recorded only once, performing three speech tasks that involved: sustained production of the vowel / a /, diadochokinesis of the syllabic set / PA-TA-KA /, and a 60-s monologue. The acoustic characteristics analyzed using the PRAAT® software were: vocal profile and articulation. Results: 10 patients with PD were included, aged 57.7 years (± 9.4), disease duration of 14.9 years (± 4.45), 60% (6) men and 40% (4) women. The difference between in-person and remote recording was observed in the acoustic analysis of the minimum fundamental frequency (F0) (p = 0.001) and the harmonic-noise level (HNR) (p = 00.005). No important differences were found in the variables of articulation of diadochokinesia and monologue, respectively: the number of syllables (p = 0.661;p = 0.861), number of pauses (p = 1,000;p = 0.702), duration (p = 0.491;p = 0.814), phonation time (p = 0.443;p = 0.904), speech rate (p = 0.929;p = 0.300), articulation rate (p = 0.823;p = 0.611) and average syllable duration (ASD) (p = 0.255;p = 0.750). Conclusion: A comparison between remote speech recording and in-person evidence that the statistically significant differences are related only to the noise variables, which is not enough to interfere in the patient's final diagnosis. Thus, remote recording can be a promising option for acoustic analysis of speech.