Speech masking effects in speech communication across the lifespan

Our ability to communicate successfully with others can be strongly affected by the presence of noise and other voices in the environment, and children and older adults are more greatly affected than young adults in these situations. Even greater disruption is experienced for populations with hearing or language impairments, or even healthy adults who are non-native speakers. Some of the disruption is due to physical masking by interfering sounds (energetic masking EM) but if the disrupting sound can be understood, this also causes further difficulty (informational masking IM). Previous work suggests that informational masking causes relatively more disruption for children and older adults than for young adults but these findings are based on laboratory tests that are far from realistic communication.

Although the impact of adverse conditions on speech communication has been the object of studies in different age groups, no study to date has taken a full lifespan view, looking at the relative impact of IM and EM on participants aged from 8 to 85 using a common experimental design. Also, many studies have focused on the impact of EM and IM on speech perception using recorded sentences or words; this ignores the fact that speakers make dynamic adaptations in speech communication to counter the effects of masking. We evaluate the impact of adverse conditions in an interactive task using measures reflecting speech communication efficiency (e.g., task transaction time, rate of dysfluencies). Finally, there is little evidence to date as to whether laboratory-based evaluations reflect the level of difficulty experienced in everyday life. The proposed project will, for the first time, relate evaluations of speech communication difficulties in adverse listening conditions as measured in the laboratory with real-life ratings of difficulty collected in real time over a two week period. It will also test whether primarily informational masking causes greater interference for some age groups (e.g. children, older adults), and if the underlying reasons for the interference differ between children and adults.

In Study 1, 120 individuals aged 8 to 85 yrs will be tested in pairs while they carry out a problem-solving task in conditions varying in the degree of informational and energetic masking present. A secondary task will be added to make the task more cognitively demanding, thus reflecting real-life multitasking situations. Baseline measures of hearing, speech perception and cognitive function and a standardised questionnaire of auditory disability (SSQ) will also be collected. In Study 2, the same participants will be asked to report perceived communication difficulty and their listening environment on 6-7 occasions per day during a 2-week period using a smartphone-based app. Data from this study will be related to measures of communication effectiveness and SSQ data collected in Study 1. Finally, in Study 3, spontaneous speech will be collected for a group of 4-7 year olds using the same interactive task to complement measures from our previous corpora (and data for Study 1) and gain a full lifespan set of acoustic descriptors for conversational speech in good listening conditions.

This project will lead to a better understanding of how the impact of adverse listening conditions changes across the lifespan, of the relative effects of IM and EM in different age groups, evaluated in realistic communicative conditions, and of the true ecological validity of laboratory-based evaluations. It will also provide normative data for a set of acoustic-phonetic measures across a 4-85 year age range. This benchmark will be of use for practitioners such as speech and language therapists and audiologists who work on aspects of communication with individuals of all ages. Importantly, this research will also contribute to our basic understanding of speech perception and production development across the lifespan.

Although this project is primarily a basic science project concerned with advancing our knowledge of the impact of masking on speech communication across the lifespan, we consider that the work has the potential to have impact in a number of fields. The main non-academic beneficiaries of this research will be audiologists, speech and language therapists (SLT), developers of speech technology applications. Finally a project leading to a better understanding of the causes of speech communication difficulties in noisy environments is of wider relevance to the general public.

Audiologists: It is well documented that speech communication in noisy environments is a main cause of difficulty for older listeners who are beginning to suffer age-related hearing loss as well as for any child or adult with hearing loss or auditory processing disorder. Although standardised audiometric tests of speech intelligibility in noise exist, they are limited to recorded words or sentences and have poor ecological validity as they do not include the adaptations that interlocutors would naturally make to counter masking effects in real-life interactions. Similarly, questionnaires asking a client to rate their auditory disability in different situations may under- or over-estimate actual difficulty. The findings of Study 1 will provide audiologists with a better understanding of speaker-related factors influencing the effect of different adverse conditions on speech communication in clients across the lifespan. The findings of Study 2, using the experiential sampling method (ESM), could be especially relevant, in terms of informing audiologists about the ecological validity of standardised tests of speech perception in noise. It will be a test also of the potential of ESM for clinical evaluations as alternatives to standardised tests and questionnaires.

Speech and Language Therapists: Although a wide range of instruments are available to SLTs for evaluating speech production and perception abilities in their clients in all age groups, these instruments do not directly test the ability to communicate in an ecologically-valid way, as they do not actually involve communication between two people or a situation in which clients are doing more than one task simultaneously. Our diapix method which is being developed in this project as a tool to gather measures of communication efficiency, rather than primarily as a means to elicit clear speech as previously used, has potential as such a tool especially given the recent development of the diapix app which gives automatic measures of transaction time.

Developers of speech technology: There is much current interest in trying to reduce the listening effort that results from communication in difficult conditions by automatically enhancing or 'enriching' the speech signal. Gaining a better understanding of how speaker-related factors affect such communication and of the relative impact of IM/EM masking in different age groups will help tailor such enrichment techniques for different populations of users.

Wider public: A great majority of the population experiences difficulty communicating in noisy environments, because of their age, their hearing, the fact that they are speaking in a foreign language, etc. The fact that speaking in the presence of other voices causes additional difficulty is not typically known even though it reflects the everyday listening situation of many people who may actually experience particular difficulties with informational masking (e.g, TV or radio in the background in many homes, as well as in care homes, background chatter in school environment). A better understanding of the impact of informational masking on particular age groups or populations could lead to simple adjustments being made to reduce such masking effects.