Speech intelligibility of speakers with masks - How facial masks affect communication in children with cochlear implant

The thesis work focuses on the assessment of speech intelligibility of children with cochlear implants in real complex acoustic scenarios, which include both competitive sound environments (i.e., with reverberation and noise) and the use of face masks that modify speech intelligibility towards a listener. The research is inspired by the recent events related to the global pandemic, as well as by the need of ensuring optimal acoustic scenarios for the support of auditory abilities especially for hearing impaired listeners. On the one side, the dramatic spread of the Coronavirus disease made it necessary face masks in public and private spaces. Although the use of face masks is needed to prevent the spreading of Coronavirus, several studies have already shown their negative effects on speech intelligibility (SI) and comprehension due to the coverings’ acoustic attenuation. On the other side, research is now focusing on the understanding of the extent to which realistic acoustic scenarios, that account for the combined presence of noise with informative content and reverberation time, affect speech intelligibility. Besides behavioural (i.e., the use of face masks) and environmental (i.e., complex acoustic scenarios) issues, there are only few studies that consider them in relation to children with hearing impairments, and particularly wearing cochlear implants (CI). These subjects have special acoustic needs that are difficult to be met in everyday life and would need the built environment support their hearing abilities. This research thus followed the open questions that arouse from the abovementioned baseline. Children between 7 and 15 years of age have been involved, who were divided into an experimental group of 14 children with CI, and a control group of 6 normal hearing children and were administrated via web-platform a SI test. Different acoustic conditions were evaluated in terms of noise (i.e., no noise, signal-to noise ratio of 0 dB, +5 dB, +10 dB) and considering three types of masks that differed for intrinsic characteristics (i.e., acoustic attenuation, filtration efficiency and breathability). SI was evaluated in terms of percentage of items correctly understood in relation to the speech material used in the test (which is the one of the Simplified Matrix Sentence Test for the Italian language), and the listening difficulty was also evaluated based on the subjective rating given by the subjects on a 5-points scale. Results showed that SI was higher in the condition without a mask and varied depending on the type of mask used. More challenging conditions, e.g., with a lower thus more competitive signal-to-noise ratio corresponded to lower SI as expected. Results on SI related to the case of no-mask were similar to those obtained with the mask that had the lowest acoustic attenuation, although characterized by a low filtration efficiency. The optimal match between intrinsic features of masks thus needs to be investigated with respect to the consequences on SI and listening difficulty. The preliminary results discussed in this work suggest the need of guaranteeing optimal acoustic conditions, especially in terms of high signal-to-noise ratios, and of selecting the most appropriate face masks for teachers to be understood by children wearing CIs. These outcomes can be transferred to improve speech communication (that is, speech production and speech understanding) in classrooms for everyday design and usage practices to be implemented to support optimal listening conditions.