Auditory short-term memory persistence for tonal signals in a songbird.

This paper presents an animal model for studying the persistence of auditory memory for tonal signals. Five European starlings (Sturnus vulgaris) were trained in a Go/NoGo delayed nonmatching-to-sample task to discriminate between a series of identical "sample stimuli" and a single "test stimulus." Frequencies of sample and test stimuli should be classified as being either the same or different. The performance of the birds is measured as the percentage of correct classifications. Three parameters were varied: The delay between sample and test stimuli, the number of sample stimuli presented before the test stimulus, and the salience of the difference between sample and test stimuli. Auditory memory persistence time was estimated as a function of the delay between the last sample stimulus and the test stimulus. The performance of the birds deteriorated with increasing delay before the test stimulus. Increasing the number of sample stimuli in the series of tones presented before the test stimulus improved performance. Performance was also better for more salient differences between sample and test stimuli. The individual auditory memory persistence time varied between 4 and 20 s.