A quantitative acoustic analysis of the vocal repertoire of the common marmoset (Callithrix jacchus)

Abstract

The common marmoset (Callithrix jacchus), a highly vocal New World primate species, has emerged in recent years as a promising animal model for studying brain mechanisms underlying perception, vocal production, and cognition. The present study provides a quantitative acoustic analysis of a large number of vocalizations produced by marmosets in a social environment within a captive colony. Previous classifications of the marmoset vocal repertoire were mostly based on qualitative observations. In the present study a variety of vocalizations from individually identified marmosets were sampled and multiple acoustic features of each type of vocalization were measured. Results show that marmosets have a complex vocal repertoire in captivity that consists of multiple vocalization types, including both simple calls and compound calls composed of sequences of simple calls. A detailed quantification of the vocal repertoire of the marmoset can serve as a solid basis for studying the behavioral significance of their vocalizations and is essential for carrying out studies that investigate such properties as perceptual boundaries between call types and among individual callers as well as neural coding mechanisms for vocalizations. It can also serve as the basis for evaluating abnormal vocal behaviors resulting from diseases or genetic manipulations.

FIG. 1.

(Color online) Schematic of the setup used to record vocalizations from individually housed adult marmosets in both populations. Directional microphones were pointed at individual monkeys so that the calls from that monkey could be traced during audio replay.

FIG. 2.

(Color online) Signal representations used to measure the acoustic features described in Table I, with representative feature measurements for each signal representation shown. (A) Time waveform (gray) and envelope (black) of a twitter call, with detected envelope peaks marked with “+” symbols. (B) Smoothed magnitude of the frequency spectrum for the beginning phrase of a twitter call. The “*” symbol marks the detected spectral peak. (C) Spectrogram and time-frequency trace for the beginning phrase of a twitter call. The minimum and maximum detected frequencies are shown along with the sweep time. (D) Spectrogram and time-frequency trace for a trillphee call. The + markers indicate detected peaks in the FM sinusoid segment of the call, the “O” markers indicate detected troughs in the FM sinusoid segment of the call, and the O marker indicates where the transition point from the FM sinusoidal to tonal segment of the call was detected. The markers in all signal representations were generated using automated feature detection software.

FIG. 3.

(Color online) Time waveforms and spectrograms of twitter calls observed from six different monkeys. Although the phrased nature and upward tendency of the FM sweeps comprising the phrases makes the twitter call easily recognizable, twitters from different monkeys show a large degree of variation in the interval between phrases, the number of phrases typically uttered, and the specific time-frequency structure of the FM sweeps.

FIG. 4.

(Color online) Examples of observed twitter call features are based on measurements made from both populations.

FIG. 5.

(Color online) Time waveforms and spectrograms of phee calls observed from six different monkeys. Contrasting the time-waveforms of (A), (B), and (C) clearly shows the broad dynamic range used in phee calls. The time-frequency characteristics of phee calls are highly stereotyped and variation between utterances is largely limited to the end of the call.

FIG. 6.

(Color online) Phee call spectro-temporal characteristics are observable in the distributions of measurements made from calls across the two populations.

FIG. 7.

(Color online) Time waveforms and spectrograms of trill calls observed from five different monkeys. The trill is distinguished based on the characteristic sinusoidal FM that comprises the call.

FIG. 8.

(Color online) Examples of observed trill call features are based on measurements made from both populations.

FIG. 9.

(Color online) Time waveforms and spectrograms of trillphee calls observed from four different monkeys. Trillphees are a hybrid form of the trill and phee, containing both a sinusoidal FM segment and a flat tonal segment.

FIG. 10.

(Color online) Examples of observed trillphee call features based on measurements made from both populations.

FIG. 11.

(Color online) The marmoset utters a variety of short duration calls classified as “peeps” which were divided into five types. Time waveforms and spectrograms of the five observed simple peep types are shown. In general, the p-peep resembles a very short phee, the t-peep resembles a very short trill, the sa-peep is characterized by a steeply rising FM sweep, the sd-peep is characterized by a steeply falling FM sweep, and the dh-peep is a declining FM sweep that blends into an FM arch. Unlike other calls such as the phee and twitter, there is a high degree of variability in time-frequency characteristics within each of the peep types.

FIG. 12.

(Color online) Tsiks (A), eggs (B), and ocks (C) are simple calls uttered in a mobbing response to a predator (Epple, 1968). Within our colony, these calls were primarily observed only when a human observer was in close proximity to the marmoset's cage.

FIG. 13.

(Color online) Marmosets frequently concatenate several phee calls in a compound call type as shown. Occasionally, the first phrase in a concatenation of phee calls is actually either a trill or a trillphee, as shown in (D). Phee strings and peep-phees are distinguishable based on whether or not the series of phee calls is preceded by one or more peep types.

FIG. 14.

(Color online) The marmoset utters several compound calls consisting of a single peep and a single trill, phee, or trillphee. In the phee-peep [(A) and (B)], the peep follows the longer call, whereas in the peep-trill and peep-trillphee [(C) and (D)] the peep precedes the longer call.

FIG. 15.

(Color online) Sequences of various simple peep calls are often observed in compound call types. These sequences may compose the compound call entirely, as in the peep string [(A) and (B)], or they may be uttered in conjunction with another simple call type, as shown with the trill-peep (C).

FIG. 16.

(Color online) In a mobbing response to a predator, marmosets frequently combine tsik, egg, and ock simple call types into compound call types. The tsik strings are typically highly variable in the order of tsiks, eggs, and ocks that comprise them. However, the tsik-egg combination surfaced frequently enough that it was considered a separate call type.

FIG. 17.

(Color online) Although twitter calls are phrased, they are often observed as a distinct syllable in a compound call type. In compound call types involving twitter calls, the twitter may be preceded by a trill which often blends into the beginning twitter phrase [(A) and (D)], and/or followed by one or more peeps [(B), (C), and (D)], or followed by one or more phees [(E) and (F)].