Cortical hemodynamic response patterns to normal and whispered speech

Whispered speech is often used in direct person-to-person communication as a means to confidentiality. Compared with normally-vocalized speech, whispered speech is predominantly unvoiced, i.e., produced without vocal fold vibration, and has no clear fundamental frequency. By using near-infrared spectroscopy (NIRS), we assessed cortical hemodynamic response patterns to normally-vocalized and whispered speech in adult listeners (n=13). Stimuli consisted of 20-s strings of Japanese word associations spoken by a female voice. Average oxygenated hemoglobin values (oxy-Hb) were obtained over two regions of interest (ROIs). Oxy-Hb values during the perception of normally-vocalized speech were highest over the left temporal ROI, but not significantly different from values measured over other ROIs. Oxy-Hb values during whispered speech were highest over the right temporal ROI and significantly higher (p<0.05) than those obtained over the left temporal ROI. No significant differences, however, were found in oxy-Hb comparisons between normally-vocalized and whispered speech, although the right temporal ROI comparison bordered on significance, with whisper inducing the higher value. Together, the results seem to suggest that whispered speech is a potent catalyst of cortical hemodynamic activity, especially over the right temporal cortex, in spite of its relatively modest sound level as compared to normal speech.