UC San Diego
The structure and perception of human vocalizations
- Author(s): Tierney, Adam Taylor
- et al.
Music takes on a variety of forms in different cultures, but several cross-cultural regularities exist. Large jumps in pitch, for example, tend to be followed by reversals in pitch direction, phrase-final notes tend to be lengthened, and pitch contours of song phrases tend to be arch-shaped. I show that these patterns exist in a corpus of European and Chinese folksongs and in a phylogenetically diverse selection of bird songs. As birds and humans diverged from a common ancestor more than 200 million years ago, but share similar song-production mechanisms, the simplest explanation for the existence of similar patterns in the songs of both animals is that they are a consequence of motor constraints. I also show that these patterns exist in instrumental composed music and speech. If these patterns are a consequence of motor constraints, the auditory system has had ample time to adapt to their presence. Simple stimuli following these patterns could, then, be useful tools for studying the auditory system. I demonstrate the viability of this approach by using emotional non-verbal vocalizations, simple stimuli with arch-like pitch contours, to tonotopically map human auditory cortex using fMRI. I demonstrate, for the first time, tonotopic organization in the superior temporal sulcus, an area commonly activated by voice stimuli. Speech and song are acoustically highly similar, a fact emphasized by my finding that they share several patterns by virtue of being produced by the same vocal production system. In fact, as Diana Deutsch has shown, it is possible to find fragments of recorded speech that, when repeated out of context, suddenly begin to sound like song. I collected twenty-four additional examples of this phenomenon from audio books, along with twenty-four control stimuli spoken by the same speakers. In an imaging study using these two sets of stimuli I show that perceiving stimuli as song rather than as speech is linked to an increase in activation in a number of areas involved with pitch processing. It appears, therefore, that one of the main differences between the auditory processing of speech and song may lie in the degree of attention directed to pitch information