Sound localization acuity of the common marmoset ( Callithrix jacchus )

Evan D. Remington, Chenggang Chen, Xiaoqin WangbioRxiv 2022 · DOI

Abstract

The common marmoset (Callithrix jacchus) is a small arboreal New World primate which has emerged as a promising model in auditory neuroscience. One potentially useful application of this model system is in the study of the neural mechanism underlying spatial hearing in primate species, as the marmoset’s visually occluded natural habitat in the forest would make sound localization an essential behavior for survival. However, interpretation of neurophysiological data on sound localization requires an understanding of perceptual abilities, and the sound localization behavior of marmosets has not been well studied. The present experiment measured sound localization acuity using an operant conditioning procedure in which marmosets were trained to discriminate changes in sound location in the horizontal (azimuth) or vertical (elevation) dimension. Our results showed that the minimum audible angle (MAA) for horizontal discrimination was on average 15° for band-passed Gaussian noise and 13° for Random Spectral Shape (RSS) stimuli, whereas the MAA for vertical locations was at 17° and 22° for band-passed Gaussian noises containing more and less high frequency energy, respectively.

Introduction

In the last two decades, the common marmoset has become a premier model for auditory neuroscience due to its complex vocalizations and primate-like brain structure. For an arboreal animal living in dense, visually occluded tropical forests, sound localization is a critical survival skill. However, while we have extensive data on how their neurons respond to sound, we lacked a clear understanding of their actual perceptual limits—specifically, how well they can locate a sound source in 3D space. This study establishes the behavioral "Minimum Audible Angle" (MAA)—the smallest change in sound source position a subject can detect—providing the ground truth necessary to interpret neurophysiological data regarding spatial coding.

Methods

The researchers used an operant conditioning procedure involving five marmosets in a "Go/No-Go" spatial discrimination task. Marmosets sat in a custom-designed wire mesh chair to minimize acoustic reflections and listened to a series of "reference" sounds from a fixed speaker. When the sound location began alternating between the reference and a "target" speaker, the marmoset had to respond by licking a tube to receive a food reward.

Experimental setup and trial structure. Panel A shows the wire mesh chair and reward tube; Panel B illustrates the stimulus timing and response window; Panel C shows the speaker array.

Stimuli included band-passed Gaussian noise (2–32 kHz) and Random Spectral Shape (RSS) stimuli. RSS stimuli vary in frequency profile on every presentation, forcing the animal to rely on spatial cues (like binaural differences) rather than memorizing a specific spectral "fingerprint" of a speaker. To ensure accuracy, an infrared camera and custom algorithm tracked head position in real-time; trials only proceeded if the animal was facing forward.

Results

The study found that marmosets have a horizontal spatial acuity of approximately 15° and a vertical acuity of 17–22°. These findings align with mammalian scaling laws based on head size; because marmosets have small heads, the distance between their ears is short, providing smaller binaural cues (ITD/ILD) than larger primates.

Behavioral performance (Corrected Hit Rate) as a function of angular separation for horizontal (A), rear (B), and vertical (C) localization tasks across subjects.

Key findings include:

  • Horizontal MAA: Average of 15° for Gaussian noise and 13° for RSS stimuli. The slightly better performance with RSS stimuli suggests a heavy reliance on binaural cues.
  • Vertical MAA: Average of 17° for broadband noise. This increased to 22° when high-frequency information (above 12 kHz) was removed, confirming that marmosets rely on spectral notches for elevation cues.
  • Frontal Bias: Marmosets are significantly worse at localizing sounds behind them (22° MAA) than in front of them (15° MAA), suggesting optimization for the frontal visual field.

Discussion and Limitations

These results provide the first comprehensive behavioral benchmarks for marmoset sound localization. While their 15° MAA is higher than the human threshold (~1°), it falls exactly on the predicted regression line for mammals of their head size.

Limitations include the small sample size (n=5) and potential training effects; while 10° seemed unachievable in this study, different reinforcement schedules might yield lower thresholds. Future work will involve recording from the auditory cortex during these tasks to correlate neural population activity with these perceptual limits.

Summary updated

Abstract

Abstract The common marmoset ( Callithrix jacchus ) is a small arboreal New World primate which has emerged as a promising model in auditory neuroscience. One potentially useful application of this model system is in the study of the neural mechanism underlying spatial hearing in primate species, as the marmoset’s visually occluded natural habitat in the forest would make sound localization an essential behavior for survival. However, interpretation of neurophysiological data on sound localization requires an understanding of perceptual abilities, and the sound localization behavior of marmosets has not been well studied. The present experiment measured sound localization acuity using an operant conditioning procedure in which marmosets were trained to discriminate changes in sound location in the horizontal (azimuth) or vertical (elevation) dimension. Our results showed that the minimum audible angle (MAA) for horizontal discrimination was on average 15° for band-passed Gaussian noise and 13° for Random Spectral Shape (RSS) stimuli, whereas the MAA for vertical locations was at 17° and 22° for band-passed Gaussian noises containing more and less high frequency energy, respectively.
BibTeX
@misc{remington2022sound,
  author = {Evan D. Remington and Chenggang Chen and Xiaoqin Wang},
  title = {{Sound localization acuity of the common marmoset ( <i>Callithrix jacchus</i> )}},
  year = {2022},
  doi = {10.1101/2022.08.13.503849},
  howpublished = {bioRxiv},
}
Figure