A comparison of capturing speech breathing using respiratory inductive plethysmography and electromagnetic articulography.

This study introduces a novel approach to speech breathing analysis by evaluating Electromagnetic Articulography (EMA) for tracking speech breathing patterns and thus providing an alternative to the commonly used Respiratory Inductive Plethysmography (RIP). We recorded speech production data from 18 speakers (9 F, 9 M), aged between 23 and 54 years, producing sentences that differ in length and syntactic complexity. Multiple EMA sensor configurations were evaluated to determine which measures best reflect the RIP signal. For the purposes of this study, the analysis was restricted exclusively to thoracic (chest) movements for both EMA and RIP. The results demonstrate that EMA reliably captures key respiratory events in speech, including inhalation and exhalation duration as well as full breathing cycle trajectories. Horizontal EMA distance measures, particularly those derived from a sternum or middle rib cage sensor combined with a sensor on the lowest cervical vertebra, showed the highest correspondence with RIP across temporal and spatial dimensions. Distance measures of these EMA sensors exhibited strong correlations with RIP signals and reproduced comparable speech breathing patterns in the temporal and spatial domain. Vertical EMA measures were more variable, but did not compromise the overall detection of speech breathing intervals, such as inhalation and exhalation phases. In sum, these findings highlight that EMA can be an effective method for the kinematic analysis of articulatory and speech breathing data. The proposed two-sensor setup is easy to apply and provides robust speech breathing data without additional instrumentation. This approach offers a practical path for incorporating speech breathing assessment into articulatory research and has the potential to advance our understanding of speech production in linguistic and clinical contexts.