Predicting vocal frequency from selected physiologic measures.

Simultaneous physiologic measures were obtained on four young adult male subjects as they sustained phonation at seven frequencies within their modal-to-falsetto voice range. Data were analyzed using a multiple regression program to determine the contribution of each measure singly and in combination with other measures to the prediction of the resultant voice frequency. Results showed that by far the best predictor was cricothyroid muscle activity for both the pooled data model, and for each individual subject. The contribution of subglottal air pressure and thyroarytenoid muscle activity increased the variance explained by only 4% while the measure of vertical laryngeal position was a significant factor in only one subject's predictive model. Partial models from the pooled data explained from 67% to 73% of the variance; whereas the obtained measures for individual subjects explained from 90% to 94% of the variance.