15th Speech in Noise Workshop, 11-12 January 2024, Potsdam, Germany 15th Speech in Noise Workshop, 11-12 January 2024, Potsdam, Germany

P22Session 2 (Friday 12 January 2024, 09:00-11:30)
From 0 (dB) to 15 (dB) in 13 (years): Extended developmental trends in the spatial release of masking for highly informational maskers

Stuart Rosen
UCL Speech, Hearing & Phonetic Sciences, London, UK

Shiran Koifman
UCL Speech, Hearing & Phonetic Sciences, London, UK
Carl von Ossietzky University of Oldenburg, Germany

There is growing evidence that spatial release of masking (SRM) develops over a much longer period through childhood and adolescence than had previously been suspected. Even so, children as young as 5 years old have shown at least some SRM. We have been investigating SRM using a modified version of the Coordinate Response Measure (CRM), the so-called Children’s CRM (CCRM). Here all sentences, both targets and maskers, are in the form of commands ‘Show the [animal] where the [colour] [number] is’, with 6 animals, 6 colours and 8 numbers, all monosyllabic. Sentences are uttered by three male British English talkers. In each trial, a target sentence (‘Show the dog …’) from a randomly selected talker is presented simultaneously with two masker sentences, one from each of the other talkers, with no overlap in animal, colour or number. In the collocated condition, all talkers are placed to the front of the listener in virtual space over headphones. In the spatially separated condition, the two masker sentences are placed symmetrically at ±45° to the participant’s right and left. Listeners are required to click on an interface with 48 buttons to indicate the colour and number of the target sentence. Speech Reception Thresholds (SRTs, the SNR which results in ~50% correct identification of the colour and the number) are determined adaptively. SRM is calculated as the difference between the two conditions. Because all talkers are of the same sex (and with a similar F0), and masker sentences contain valid target words, this situation is one of high Informational Masking (IM), because there are few cues for a listener to use to attend to the correct target, especially in the collocated condition.

We have applied these tests to children aged 5-18, with two striking outcomes. First, in the collocated condition, there is very little change in SRT over age, ~2.6 dB on average over the 13-year age span of the children, with the mean SRT always at positive SNRs. Presumably, listeners are only able to use a loudness cue to focus on the target. Secondly, 5-year-olds displayed no SRM at all, with a clear non-zero value only occurring from age 7.5 years. The oldest children obtained ~15.5 dB release from masking with no indication of a plateau in performance even at age 18. It seems likely that this inability to use spatial separation to improve performance relates to immature executive functions, in particular, attention.

Acknowledgments: Thanks so much to Bethany Davison and Katie Wilkins who ran these studies as part of their MSc project in Speech & Language Sciences at UCL.

Last modified 2024-01-16 10:49:05