Evaluation of Various Methods to Calculate the EGG Contact Quotient

Work carried out under the ERASMUS exchange programme. The EGG contact quotient (CQegg) is the ratio of the duration of the 'contact phase' to the entire glottal cycle period. Various methods for determining the contacting and de-contacting 'events' and then the CQegg have been advanced by different authors. There are criterion-level methods which look for threshold crossings in the short-term peak-to-peak amplitude of the EGG signal; peak-picking methods applied to the first derivative of the EGG signal (DEGG); and hybrid methods that use both the EGG signal and its first derivative. The results from six published EGG-based methods for calculating the CQegg were compared with the closed quotients derived from simultaneous videokymographic imaging (CQkym). Two trained male singers were asked to phonate at 295 Hz (D4) in falsetto and in chest register, with two degrees of adduction in both registers. The maximum difference in the computed CQegg between methods was 0.3 (out of 1.0). The CQegg was generally lower than the CQkym. Within subjects, the CQegg co-varied with the CQkym, but with changing offsets depending on method. DEGG methods had a low success rate in determining the CQegg, due to the occurrence of multiple peaks in the DEGG signal. No reasonable results can be expected for falsetto phonation with little adduction (untrained, non-counter-tenor-style falsetto). Basic criterion-level methods with thresholds of 0.2 or 0.25 gave the best match to the CQkym data. The temporal and spatial resolution of the kymograms were however only just adequate for this study. The results suggest that contacting and de-contacting events that are derived from the electroglottographic signal might not refer to the same physical events as do the beginning and cessation of airflow. Hence the concept of contacting and de-contacting 'phases' may be more appropriate than contacting and de-contacting 'events'.