Audiometria automática por meio da estratégia de detecções mínimas consecutivas baseados em simulação de Monte Carlo e detectores multivariáveis usando respostas auditivas em regime permanente na banda de 80 Hz

• Ano de defesa: 2020
• Autor(a) principal: Tiago Zanotelli
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; ENG - DEPARTAMENTO DE ENGENHARIA ELETRÔNICA; Programa de Pós-Graduação em Engenharia Elétrica; UFMG
• Palavras-chave em Português: Resposta auditiva em regime permanente (ASSR); Eletroencefalografia (EEG); Estratégia para testes sequenciais; Detector multivariado; Audiometria automática; Engenharia elétrica; Eletroencefalografia; Método de Monte Carlo
• Link de acesso: http://hdl.handle.net/1843/34527

Auditory Steady-State Responses (ASSRs) are electrical potentials generated in the auditory system that are usually obtained by electroencephalography (EEG) and whose detection is performed objectively and frequency-specific. In clinical applications, ASSRs have been used to obtain an electrophysiological audiogram, benefiting patients who are unable to respond to conventional audiometry, e.g., neonates and patients with motor or cognitive disorders. However, the technique has limitations, especially regarding its exam time, which requires the development of new detection techniques and electrophysiological threshold search strategy. Thus, the present work proposes a method for the automatic, online estimation of the electrophysiological threshold using ASSR (automatic audiometry). To this end, new multivariate detectors were developed and applied together with two strategies: one for sequential testing where the number of consecutive detections is used as a stop criterion and another for electrophysiological threshold search with independent and automatic adjustment of stimulus intensity. The electrophysiological threshold of 18 adult volunteers with normal hearing was determined by automatic audiometry for multiple stimulation (four frequencies per ear) at a modulation frequency of 80 Hz. The average audiometric exam time was 33 minutes, showing a reduction of 50% time compared to the conventional strategy. In addition, the average difference between electrophysiological and tonal thresholds was around 19 dB, similar to the results reported in other studies. Additionally, results of an offline analysis showed that the strategy can be used to perform even faster exam, thereby obtaining an average exam time of 24 minutes with a variation of around 22 dB between the thresholds. The results showed that the automatic audiometry proposed in this work is able to reduce the exam time, without a significant loss in the electrophysiological threshold accuracy, thus being able to be adapted in the future in hospitals and clinics, helping in the investigation of hearing loss in neonates, or adult patients who do not spontaneously cooperate with behavioral.