Temporal integration and decision-making in crocodiles

Abstract

To make appropriate behavioural decisions, animals continuously process a flow of information provided by different sensory channels. Could temporality, i.e. the order in which independent stimuli are perceived, lead the animal to give greater importance to one stimulus than to another? Here we show that the decision of a crocodile to move preferentially towards the source of water surface waves than towards the source of an airborne sound is irrespective of the relative time of arrival of the sound and water vibrations to the animal, as long as the delay between these two stimuli does not exceed a few seconds. To test whether the late arrival of water waves - which travel more slowly than sound - could explain crocodiles' preference for the source of water waves, we controlled the relative timing of stimulus arrival within a time window of a few seconds. Our results reveal that crocodiles preferentially move towards the source of the water waves, whether they arrive after, at the same time as, or before the sound. This suggests that the temporal integration of information from different sensory channels can occur within a certain time window, where the behavioural decision-making remains independent of the arrival order of stimuli. The maximum delay between simultaneously evaluated stimuli probably depends on animal species and context.

Keywords: Crocodiles; Decision-making; Sensory perception; Sound; Temporality; Water surface waves.

Fig. 1.

An airborne sound and a water surface wave independently attract young crocodiles. (A) Experiment 1a. An airborne sound and a water surface wave are emitted at the same time from two different locations (separated stimuli; 22 trials). The crocodiles (N=9 individuals) were attracted to both stimuli, with a tendency to approach more the vibration source. (B) Experiment 1b. Sound (22 trials) and water surface waves (27 trials) are emitted alone or altogether from the same location (co-located stimuli; 22 trials). Crocodiles approached the stimulus source whatever the stimulus. There is no interactive effect between sound and vibration when they are presented together. Violin plots represent the shortest distance to the stimulus source reached by the crocodile (fitted value of the median of the posterior distribution with 95% CI). The distributions are derived from repeated observations across individuals.

Fig. 2.

Effect of stimuli order and arrival time on the crocodile response. (A) Experiment 2. To control the order of arrival of stimuli at the crocodile, sound (S) and water surface waves (WW) are emitted from two different locations with different delays (40 trials for each condition). In the ‘sound-vibration’ condition, the onset-to-onset delay between the last sound and the first water wave is 6 s. In the ‘synchronized’ condition, the onset-to-onset delay between the first sound and the first water wave is set to zero. In the ‘vibration-sound’ condition, the onset-to-onset delay between the last water wave and the first sound is 6 s. (B) Regardless of the order of arrival of the two stimuli, the crocodiles (N=10 individuals) preferentially approached the water surface wave source. Violin plots represent the minimum distance to the stimulus source reached by the crocodile (fitted value of the median of the posterior distribution with 95% CI). (C) Distributions of measured time delays between the arrival of the vibration stimulus and the arrival of the sound stimulus at the crocodile depending on the playback condition (‘natural’ condition=the two stimuli are emitted at the same time). (D) The later the sound arrives, the more the crocodiles are attracted to the source of the vibration [Bayes Factor on the interaction between the delay and the type of stimulation=19,100; slope for vibration=−2 cm, 95% CI (−4, −1), 99.88% of posterior distribution is negative; slope for sound=1 cm, 95% CI (−1, 4), 90.45% of posterior distribution is positive]. The distributions are derived from repeated observations across individuals.

Fig. 3.

Experimental set-up and sensory stimuli. (A) Experimental basin with the sound and vibration emitting devices (remote-controlled loudspeakers and vibrating devices). The black lines on the bottom of the basin are drawn at distances of 2, 3, and 4 m from each device allowing us to determine the initial position of the crocodile before sending the stimuli. (B) From left to right: spectrogram and oscillogram of a crocodile call; the duo loudspeaker (green)/vibratory device (blue); surface wave created by the vibratory device at the source of stimulation. (C) Airborne sounds propagate more rapidly than water surface waves and are less attenuated (wave drawings are not at scale).

Fig. 4.

Calibration curve. Water height h as a function of the propagation distance r and calibration curve that best fit the whole data set. Measurements were done for various water depth H=22, 10 and 7 cm and two different striking disk radii (r₀=25 mm and r₀=15 mm).

Fig. 5.

Model and experimental time signals. Model signals (on top) are compared to measured signals (at the bottom), for the same distances r from the centre of the striking disk. The water depth considered for the calculation is H=22 cm, the disk radius is r₀=25 mm and the physical properties of water are ρ =998 kg m⁻³, γ =72.8×10⁻³ Nm⁻¹, g=9.81 m s⁻² and ν =1.01×10⁻⁶m² s⁻¹.

Fig. 6.

Surface waves dispersion. Phase and group velocities as a function of wavelength λ, with and without deep water approximation. The phase velocity strongly depends on the wavelength. As a consequence, the different components of a complex waveform propagate at different velocities, leading to constant deformation of this waveform. The water depth in shallow water case is H=7 cm. The physical properties used for these calculations are for water at 20° and atmospheric pressure: ρ =998 kg m⁻³, γ =72.8×10⁻³Nm⁻¹, and g=9.81 m s⁻².