Energetic and informational masking of complex sounds by a bottlenose dolphin (Tursiops truncatus)

Abstract

With few exceptions, laboratory studies of auditory masking in marine mammals have been limited to examining detection thresholds for simple tonal signals embedded in broadband noise. However, detection of a sound has little adaptive advantage without the knowledge of what produced the sound (recognition) and where the sound originated (localization). In the current study, a bottlenose dolphin's masked detection thresholds (energetic masking) and masked recognition thresholds (informational masking) were estimated for a variety of complex signals including dolphin vocalizations, frequency modulated signals, and a 10 kHz pure tone. Broadband noise types included recordings of natural sounds and computer generated sounds. Detection thresholds were estimated using a standard go, no-go adaptive staircase procedure. The same dolphin learned to associate whistle-like FM sounds with specific arbitrary objects using a three alternative, matching-to-sample (MTS) procedure. The dolphin's performance in the MTS task was then tested in the presence of the same masking noise types used in the detection task. Recognition thresholds were, on average, about 4 dB higher than detection thresholds for similar signal-noise conditions. The 4 dB difference is likely due to additional cognitive demands of recognition, including attention and pattern recognition.

Human auditory discrimination of bottlenose dolphin signature whistles masked by noise: Investigating perceptual strategies for anthropogenic noise pollution

“…Unsurprisingly, masking noise can impair dolphins' abilities to echolocate accurately, which could impact their ability to forage. One study (Lemonds, Au, Vlachos, & Nachtigal, 2012) which examined a bottlenose dolphin's masked hearing thresholds of a pure tone, found similar masking patterns to the aforementioned body of research (Au et al, 1988;Au & Penner, 1981;Branstetter et al, 2016a;Branstetter et al, 2013a) on masking and echolocation. Branstetter and colleagues (2016a) studied the effects of several types of natural and artificial noise sources on dolphin detection of various natural and artificial dolphin vocalizations: one non-signature whistle, one burst pulse, echolocation clicks, a 10 kHz tone, and four artificial, whistle-like sounds.…”

“…Frequency contour was reportedly the most common cause of mistaken whistle identity. It had the highest frequency of use, and it was the most frequently cited cue for identifying each individual dolphin's whistle as well, indicating that this either does not vary among dolphins or varies only somewhat (e.g., Branstetter et al, 2016a;Branstetter et al, 2016b).…”

“…Current regulations apply to behavior that may disrupt or alter the natural behavior of all marine mammals (Marine Mammal Protection Act, 1972), but in order to properly understand what constitutes such a disruption, it is imperative to understand the full scope of human impacts on the cognition, behavior, and perception of marine mammals. Previous research has focused on the basic acoustic properties of noise and dolphin vocalizations (Albuquerque & Souto, 2013;David, 2006), on the perception of simple tones, broadband clicks, and biosonar echoes in noise (e.g., Au, Moore, & Pawloski, 1988; Au & Penner, 1981;Branstetter et al, 2013b;Weilgart, 2007), or general behavior responses of dolphins to the presence of boats (Janik & Thompson, 1996;Lusseau, 2003;Pirotta et al, 2015). No studies have yet examined how boat noise directly impacts dolphins' ability to discriminate among their whistle vocalizations.…”

“…The extent to which masking noise might impair a dolphin's ability to echolocate has been studied in great detail with animals in human care (e.g., Au et al, 1988;Au & Penner, 1981;Branstetter et al, 2016a;Branstetter, Trickey, Aihara, Finneran, & Liberman, 2013a), revealing thresholds at which sounds affect dolphins' ability to detect or discriminate among objects using echolocation. In general, this research suggests that, like in humans, the effects of masking noise vary with both the frequency and amplitude of both the noise and the signal.…”

“…In addition, the study used two types of artificially-generated noise. It is crucial to note that some signals, especially echolocation clicks, were masked more or less severely than others by certain noises (Branstetter et al, 2016a). This disparity in masking patterns across different noise types emphasizes the need for further study of specific signal types and noise types, especially commonly-used signals such as signature whistles.…”