Selected sounds from the Aquamark100 pinger were used in this study. These stimuli as well as the sound field were measured regularly using a pistonphone calibrated (+/- 1 dB) Reson TC4032 hydrophone connected to an Olympus LS-10 digital recorder. The tests were run via a LabView program on a laptop connected with USB to DAQPAD (National Instruments USB-9162). The output of the DAQPAD was connected to a Basetech PA-1001 amplifier and to an UW 30 (University Sound) underwater loudspeaker, which emitted the bridge sound, and to a hydrophone (Sonar products HS-70), which emitted the pinger sounds.

The underwater apparatus consisted of a rig made of polyvinyl chloride (PVC) tubes, designed to be placed at the edge of the pool and fixed with weights. This rig contained a response target and a chin station. The chin station consisted of a PVC plastic plate on which the seal could rest its head, indicating that it was prepared for a trial. This controlled the seals’ head position in the water and ensured consistency across trials and sessions. The chin station was placed 75 cm below the water surface. The response target was a red cylindrical object located at the water surface 20 cm to the left of the chin station (Fig. 3). An underwater speaker was mounted at the same level as the chin station (Fig. 3). A hydrophone was attached to the speaker to pick up the transmitter sounds. The sounds were run through a click detector (ECD1, NewLeap Ltd., UK), to make the pinger sounds audible to the operator, and thus, ensure that the system worked as intended. This click detector enabled the trainer to hear the pinger sounds above the water through a set of headphones. Using headphones prevented the seal to hear air-borne sounds coming from the click detector. 

For each session, the subject was moved to a 5-m diameter test pool (Fig 4). Each trial started after the seal entered the water and placed its head on the chin station (Fig. 4).

The experimental procedure was carried out as an auditory go/no-go procedure. This means that, if the seal detected a sound, it touched the response target with its snout, and if it did not detect a sound, it remained on its position at the chin station. These tests were performed using the staircase method. This is a psychophysical technique (i.e. measured through behavior) that aims to reveal absolute thresholds. Here, a stimulus is initially presented at a level known to be above the animal’s threshold. The stimulus level is lowered (in 6 dB steps in the present study) until the subject fails to detect it, and it is increased again (also in 6 dB steps) until the subject detects the stimulus again.

Thereupon, the stimulus level is lowered in steps again until the subject fails to detect it. A light placed in front of this station at eye level was turned on to mark the initiation of each trial, and it remained on throughout the whole trial. A trial ended either after the subject touched the response target or when the response time limit (4 s) was over and the light turned off.

Two types of trials were tested, signal-present and signal-absent. Therefore, there were four possible outcomes: correct detection, correct rejection, incorrect detection (false alarm), and incorrect rejection (miss). 

Correct detection: When the subject left the chin station upon presentation of the pinger sound and touched a response target with its snout (Fig. 5A). The subject was given a 4 s response time limit, within which a response could be considered correct. Reinforced with a 1-s shrill from a dog whistle (secondary reinforcer) followed by a fish reward (primary reinforcer) given by the trainer.

Incorrect detection (false alarm): When the subject touched the response target without the presentation of a signal. No reinforcement was provided.

Correct rejection: When the subject remained at the chin station until the end of the trial (i.e. after 4 s) (Fig. 5B), marked by the light being turned off. Reinforced by an underwater 0.5 s broadband buzz (secondary reinforcer) along with a fish reward.

Incorrect rejection (miss): When the subject remained at the station when a signal was presented.

In total, 200 Aquamark trials were carried out in 6 sessions for one seal. Each session included from 25 to 35 trials, including 4 warm-up and 4 cool-down trials. Sessions had a duration ranging from 5 to 15 min depending on the motivation of the subject and equipment functioning. During the warm-up and cool-down trials, a test signal known to be audible for the seals was presented to maintain the subject’s motivation by starting and ending each session with easily detectable signals, for which the seal was reinforced. The sequence of trials was randomized for every session, but did not include more than 3 signal-present or signal-absent trials in a row to avoid predictability by the subject.