If you've ever had a problem with your voice, you might have had an electroglottograph taken to find out how your vocal folds are working. Christian Herbst, from the Palacký University Olomouc, Czech Republic, explains that the vocal folds – which are situated in the larynx at the top of the trachea – open and close as we talk, to alter the flow of air exhaled from the lungs to produce a noise that is then modified by our throats and mouths to produce the bewildering array of sounds that we know as human speech. And when the voice is damaged, one of the first things that the doctors investigate is the vocal fold's vibration. However, Herbst explains that assessment of the vocal folds is costly and invasive, ‘You have to stick an endoscope down a patient's throat and that is not pleasant’, he says. Fortunately, doctors can resort to a less-intrusive approach. They collect electroglottographic data by monitoring variations in high-frequency weak electric current passed through the vocal folds by electrodes that are placed either side of the larynx. ‘It is a low-cost, non-invasive alternative’, says Herbst. However, the precise details of how well the visible vocal fold vibrations correlated at high speeds with the fluctuations in the electric current were not clear. So, while based at the University of Vienna, Austria, Herbst collaborated with colleagues from several European institutions in a high-speed analysis of the relationship between vocal fold vibrations and the electroglottograph data (p.955).
Blowing a carefully controlled flow of air through the larynx of a golden retriever that had recently died from natural causes, Herbst filmed high-speed movies (at 27,000 framess−1) of the vocal fold's vibrations while measuring fluctuations in the weak electric current passing through the vocal folds at a frequency of 44,000 Hz to look for correlations between the vibration and variations in the electric current. Using several graphical methods to analyse the vocal fold's complex rippling motions, Herbst and his colleagues could see that the fluctuations in the electric current were not as tightly correlated with the vocal fold vibrations as had been suggested by earlier video studies at lower speeds.
The team found that the changes in the electroglottograph electric current did not coincide perfectly with the times when the vocal folds opened. Also, the electric current fluctuations – which were thought to indicate the precise moment when the vocal folds opened or closed – were delayed by between 0.02 and 0.61ms, which doesn't sound like much, but is as much as 10.9% of one of the vocal folds' rippling cycles. Herbst and his colleagues also noticed that there is not a single precise moment when the vocal folds separate or come back into contact. This is because of the complex rippling motions that occur not only from the front to the back of the vocal fold slit, but also down through the depth of the slit – which is about 5mm deep – so both edges do not separate completely at one defined moment, but open more gradually depending on the way that the rippling motion propagates through the tissue.
So, Herbst and colleagues conclude that variations in the high-frequency electric current measured by an electroglottograph across vibrating vocal folds do not reflect the motion of the tissue precisely during sound production. However, the team are optimistic that high-speed videos of the vocal fold's movements will give us a better understanding of the relationship between peaks in electroglottograph data and the vibrating motions that produce them, to provide a cheap method for assessing vocal fold vibrations in humans and other animals.
