Looming sounds as warning signals: The function of motion cues

Abstract

Sounds with increasing intensity can act as intrinsic warning cues by signalling that the sound source is approaching. However, intensity change is not always the dominant motion cue to a moving sound, and the effects of simple rising intensity sounds versus sounds with full three dimensional motion cues have not yet been directly compared. Here, we examined skin conductance responses, phasic alertness, and perceptual and explicit emotional ratings in response to approaching and receding sounds characterised either by full motion cues or by intensity change only. We found a stronger approach/recede effect in sounds with full motion cues for skin conductance response amplitude, suggesting sustained mobilisation of resources due to their greater saliency. Otherwise, the approach/recede effect was comparable in sounds with and without full motion cues. Overall, approaching sounds elicited greater skin conductance responses and phasic alertness, and loudness change was estimated higher. Also, they were rated as more unpleasant, potent, arousing and intense, and the probability of such sounds to signal a salient event or threat was rated higher. Several of these effects were modulated by sex. In summary, this study supports the suggestion that intensity change is the dominant motion cue mediating the effects of approaching sound sources, thus clarifying the interpretation of previous studies using such stimuli. Explicit emotional appraisal of such sounds shows a strong directional asymmetry and thus may reflect their implicit warning properties.

Introduction

Detecting an approaching object can be critical for survival. From an evolutionary perspective, it has therefore been argued that a perceptual bias for approaching (or looming) sounds in preference to receding sounds would be of selective advantage (Neuhoff, 2001). Evidence for such a bias comes from studies showing that the distance or time-to-arrival of looming sound sources is underestimated by humans, both in absolute terms and compared to receding sounds (Neuhoff, 2001, Neuhoff et al., 2009, Schiff and Oldak, 1990).

Moving sound sources provide a vast array of specific acoustic cues. Among them are binaural cues such as interaural level and time differences, and monaural cues such as overall intensity change, Doppler shift and atmospheric filtering. It has been suggested that an overall increase in intensity is dominant for inferring the position of an approaching sound source (Rosenblum et al., 1987), at least at slow sound source speed (i.e. 10 m/s, Lutfi and Wang, 1999). Since it is easier to quantify this intensity increase than the other complex characteristics of approaching sound sources, a number of studies have used diotic stimuli with rising sound intensity as a model for looming sound sources. Loudness change in such sounds is overestimated (Neuhoff, 1998, Neuhoff, 2001). An intensity increase of 15 dB over 2 s elicits a stronger autonomic orienting reaction than an intensity decrease of the same magnitude, quantified as phasic skin conductance reaction [SCR] 4–5 s after stimulus onset, and heart rate deceleration 2–3 s after stimulus onset (Bach et al., 2008). In this context, enhanced SCR has been interpreted as reflecting mobilisation of energetic resources (see also Barry, 1987). At the same time, rising intensity appears to accelerate reaction times [RTs] to subsequent auditory targets (Bach et al., 2008). This indicates enhanced phasic alertness and resembles responses to experimentally learned warning cues (Roberts et al., 2006).

However, the heavy reliance on simple intensity change to examine putative auditory motion perception can be called into question. For example, at higher sound source speeds (i.e. 50 m/s), motion cues other than overall intensity change can be dominant (Lutfi and Wang, 1999). Thus, it is possible that the perceptual bias for approaching sounds is not only due to the differential effects of simple rising and falling intensity. The approach/recede effects found in previous work could be even more pronounced when sounds contain full motion cues and are therefore easier to characterise in terms of speed and position. However, the opposite prediction might also be true. In neuroimaging studies, stimulus relevance has been found to be represented in the amygdala (Sander et al., 2003), and the amygdala also responds to looming sounds (Bach et al., 2008). There is some evidence that such amygdala responses are increased when potentially relevant stimuli lack precise information about their relevance (Whalen, 1998; see also Whalen et al., 2001). Amygdala responses are supposed to reflect stimulus saliency in this context, and are related to phasic SCRs (Laine et al., 2009, Williams et al., 2001). Sounds containing intensity change as only motion cue lack precise motion information. Hence, one might expect stronger SCRs, and possibly perceptual bias and phasic alertness to approaching sounds that are characterised by simple intensity change only, while for receding sounds there may be no difference as these sounds might be perceived as bearing little relevance anyway.

In the present study, we compared complex full-cue binaural approaching/receding sounds to diotic rising/falling intensity sounds. In this 2 × 2 factorial design, we expected differential effects of motion cues in the interaction sound source direction × motion cues. One aim of this study was to help interpret the results of previous experiments. Thus, stimuli and dependent measures were chosen to be comparable with those of previous studies. In a first experiment, we measured ratings of loudness change to examine the magnitude of the perceptual bias in each condition. In the second experiment, we measured SCR and RTs to subsequent targets as measures of the warning properties of each type of sound. This was conducted on a sample independent from the first experiment in order to prevent effects of repeated exposure to the stimuli. We also sought to examine how participants would explicitly characterise approaching and receding sounds. Recent work has argued that the perceptual bias for approaching sounds could be due to perceptual characteristics alone (Grassi and Darwin, 2006). Thus, we used a novel measure to more explicitly assess the significance of these sounds by collecting emotional ratings. We hypothesised that these ratings would reflect the implicit warning properties of approaching sounds. Finally, because there is evidence for effects of sex on the perceptual bias (Neuhoff et al., 2009), we included this as additional factor into all analyses.