Description:
<jats:p>Acoustic recognition of individuals has been demonstrated experimentally in most, if not all, birds and mammals studied from this point of view. The investigator trying to identify the “acoustic signatures” on which recognition is based invariably finds what he seeks, sooner or later, depending on the degree of intra-individual variation in vocalizations compared to inter-individual variation. However, acoustic signatures brought out using analytical methods are not necessarily those used by the animals themselves. Such is the case for peach-faced lovebirds (Agapornis roseicollis). In an attempt to overcome this problem, we designed a set of pattern discrimination techniques which are based on previous psychoacoustic and neurophysiological data and whose validity could be tested experimentally. Our techniques exploit temporal evolution of energy spectrum density and comprise Fourier analysis, various standardization procedures, and three multivariate statistical analyses. Use of these techniques in further psychoacoustic and neurophysiological investigations, particularly those concerning the functioning of “complex” neurons, is discussed.</jats:p>