| Abstract #138, Date 2/14/99, Session E6, Poster (B106) |
| Parametric modeling of species-specific vocalizations of the common marmoset |
| *R.K. Snider, X. Wang (Johns Hopkins University, Baltimore, MD) |
Species-specific vocalizations are behaviorally important sounds. When studying neural responses to such spectrotemporally complex stimuli, a central question is to understand which aspect of the stimuli a neuron is responding to. This requires the ability to change the stimulus along many dimensions. Traditional methods modify the stimulus by relying on various filtering strategies or global operations such as signal reversal, compression, or expansion which are typically performed on "examples" of representative calls. These methods are limited in two ways. First, when individual calls are used, they may not be statistically accurate in representing a particular class of vocalization. Analysis in our lab has shown considerable intraclass variability in marmoset vocalizations (Agamaite & Wang, ARO Abs 20:573, 1997). Second, there are many aspects of a complex call that cannot be changed by filtering or global operations. In order to address these limitations we have developed a new approach for studying neural responses to species-specific vocalizations. We have constructed functionally descriptive synthetic versions of the major types of marmoset vocalizations. These parametrically controlled stimuli are based on the statistical properties of a large number of vocalization samples ( > 1000 per call type) from eight individual common marmosets (callithrix jacchus jacchus). They are generated by a time-varying frequency synthesis procedure modeled after various marmoset call types. The significance of this approach is that it accurately captures the statistical properties of marmoset vocalizations and it allows the generation of both representative calls and their deviations. Moreover, virtually any conceivable call parameter can be controlled. This provides a powerful tool to probe cortical responses to complex vocalizations along many stimulus dimensions and their combinations. The resulting data are important for us to understand the processing of species-specific vocalizations by neurons in the auditory cortex with complex receptive fields. Neural sensitivity to various changes in stimulus dimension using our method will be described in the poster. |