For 31.2% of the music plays, people did not recognize highly salient animal voices (regarding the type of acoustic source as well as the frequency spectra) when executing a counting task

The unnoticed zoo: Inattentional deafness to animal sounds in music. Sandra Utz, Friedericke Knauss & Claus-Christian Carbon. Attention, Perception, & Psychophysics, Aug 25 2022. https://rd.springer.com/article/10.3758/s13414-022-02553-9

Abstract: Inattentional unawareness potentially occurs in several different sensory domains but is mainly described in visual paradigms (“inattentional blindness”; e.g., Simons & Chabris, 1999, Perception, 28, 1059–1074). Dalton and Fraenkel (2012, Cognition, 124, 367–372) were introducing “inattentional deafness” by showing that participants missed by 70% a voice repeatedly saying “I’m a Gorilla” when focusing on a primary conversation. The present study expanded this finding from the acoustic domain in a multifaceted way: First, we extended the validity perspective by using 10 acoustic samples—specifically, excerpts of popular musical pieces from different music genres. Second, we used as the secondary acoustic signal animal sounds. Those sounds originate from a completely different acoustic domain and are therefore highly distinctive from the primary sound. Participants’ task was to count different musical features. Results (N = 37 participants) showed that the frequency of missed animal sounds was higher in participants with higher attentional focus and motivation. Additionally, attentional focus, perceptual load, and feature similarity/saliency were analyzed and did not have an influence on detecting or missing animal sounds. We could demonstrate that for 31.2% of the music plays, people did not recognize highly salient animal voices (regarding the type of acoustic source as well as the frequency spectra) when executing the primary (counting) task. This uncovered, significant effect supports the idea that inattentional deafness is even available when the unattended acoustic stimuli are highly salient.

Discussion

Inattentional deafness was firstly shown by Dalton and Fraenkel (2012) were participants missed by 70% a voice repeatedly saying “I’m a Gorilla” when they were focusing on a primary conversation. In the present study, not only the validity perspective was extended by using ten excerpts of popular musical pieces from different music genres, but also acoustic signals highly distinctive from the primary sound (i.e., animal sounds) were presented. Including sounds created by nonhumans into music is not only used as artificial stimuli in experimental studies but also in experimental music—the so-called biomusic. Some famous examples are the symphonic poem “Pini di Roma,” by Respighi first performed in 1924, where a recording of a real nightingale was included in the orchestra performance, or Pink Floyd using howling and barking dogs in their songs “Seamus” (1971) and “Dogs” (1976).

Results of the present study using those biomusic elements revealed a successful transfer of the results by Simons and Chabris (1999) and by Dalton and Fraenkel (2012) to the auditory domain of musical processing. Many of the clearly audible, very salient “auditory Gorillas” with no connection to music (in contrast to Koreimann et al., 2014) passed unnoticed by participants listening to musical pieces while being occupied with an attention-consuming counting task.Footnote3 Importantly, this finding based on a paradigm where the primary and secondary auditory signal was very different as they stemmed from different domains (primary signal was a piece of familiar music, the secondary signal was an animal sound) essentially extends the preliminary finding of Koreimann et al. (2014) where the domain was not different between both signals.

Attentional focus

Regression analyses showed that only the objective measure of deviation from the correct number in the counting task seems to be a significant predictor of inattentional deafness. A higher deviation from the correct answer in the counting task led to a higher likeliness to perceive the animal sounds. Therefore, more counting errors on the task led to a lower susceptibility for inattentional deafness. It could therefore be assumed that inattentional deafness is due to a lack of directing the entire attentional focus on the task. It might be that those participants with more errors generally focus less on the primary counting task and have more cognitive resources available to detect the animal sound in the first place. Another explanation might be that errors are a consequence of detecting the animal sound, diverting the attention from the primary task to the sound, and as a result, losing the count. These results are not only consistent with inattentional blindness/deafness literature (e.g., Wayand et al., 2005), but are also in accordance with findings on change deafness (see, e.g., Neuhoff, & Bochtler, 2018; tendentially in Vitevitch, 2003). However, Koreimann et al. (2014) did not find significant differences in the primary task between detecting and not detecting the unusual event. Also, the performance in the study by Vitevitch (2003) was only tendentially slower for those who detected the voice change. Future research has to look closer to the precise parameters of attentional focus on different cognitive processing levels responsible for inattentional deafness.

Perceptual load

The frequency of inattentional deafness could not be explained by the variation of perceptual load in this study. Higher perceptual load did not induce a higher susceptibility for inattentional deafness. Considering the participants’ mean ratings of task difficulty (see Table 3), this might partially be due to the fact that tasks were generally perceived as very demanding. The mean task difficulty of 5.53 (on a rating scale ranging from 1 = very easy to 7 = very demanding) can be considered as very high, especially given the fact that the “easiest” task still received an average rating of 4.20. Therefore, the variance of task difficulty might have been too low and the tasks generally too difficult to be able to clearly differentiate high from low perceptual load.

Feature similarity and feature saliency

The absence of an effect of feature similarity on the frequency of inattentional deafness can at least partially be attributed to a slight distinction between the two groups. Since the main aim of the present study was to show inattentional deafness to a highly striking auditory stimulus in music, attention was primarily focused on transferring the findings of the visual domain into the auditory domain by using a similar paradigm as was used by Koreimann et al. (2014). Therefore, particular emphasis was put on fitting animal sounds to the musical pieces regarding low feature similarity in order to maximize the effect. The musical pieces with particular distinct low feature similarity were then sorted into the low feature similarity condition, and the rest was sorted into the high feature similarity condition. However, the high feature similarity condition was less distinct since the fit between musical pieces and animal sounds was far more heterogeneous. The small sample size of five musical pieces in each of the groups needs to be taken into account as well, especially with regard to heterogeneity within these relatively small groups. Therefore, interpretation of the differences between musical pieces concerning feature similarity should also be conducted on an individual basis. As can be seen in Table 3, the lowest rate of inattentional deafness was found in The Moldau, by Bedřich Smetana, and in Symphony No. 5, by Ludwig van Beethoven. Whereas feature similarity was high in the Symphony No. 5 (due to the task requiring to follow different instrumental voices in the piece which varied in pitch and were similar in pitch to the Gorilla) it was low in The Moldau (because the task required to pay attention to the violin which was consistently distant in absolute pitch from the roar of the lion that served as the unexpected animal sound). Yet the roar of the Lion and of the Gorilla was easily detected. A similar finding can be reported for the pieces with the highest percentage of inattentional deafness—Ain’t No Sunshine, by Tom Jones, and In the Hall of the Mountain King, by Edvard Grieg. Whereas the high howling of a Wolf and the low-frequency bass voice of the singer in Ain’t No Sunshine suggest a low feature similarity, the cymbals and the cock-a-doodle-doo in the Hall of the Mountain King were comparatively much closer in tone pitch and therefore had high feature similarity. Yet inattentional deafness had a high frequency in both pieces. Differences in feature similarity might not be sufficient to entirely explain why some musical pieces were comparatively much more prone to inattentional deafness than others. Another factor might be, that both the Lion and the Gorilla were often described as a growling sound. A growling animal can be considered as a warning sound. If a significantly lower rate of inattentional deafness could be demonstrated in those musical pieces containing an aggressive sounding animal, this would support the assumption made by Murphy et al. (2013) of the auditory modality having an early warning function which can be crucial for the detection of alarm sounds in the environment. Lastly, it should be taken into consideration that a musical piece is a very complex construct containing several different streams of musical voices. The overall complexity of the piece and the number of instruments or voices in a musical piece might have an impact and should be considered in future research regarding inattentional deafness in music.

In addition to the subjective descriptions, the feature saliency of animal sounds was further analyzed by looking at replicable psychoacoustical measures such as loudness, specific loudness, roughness, and impulsiveness. All manipulated music pieces with animal sounds were compared with the original pieces without animal sounds. All these measures were not able to predict the percentage of missed animal sounds.

General discussion

The results of this study demonstrate that inattentional deafness in the musical realm exists, even when a highly bizarre, noticeable auditory “Gorilla” appears twice during a known musical piece. Contrary to what had previously been shown by Simons and Chabris (1999) feature similarity or our additionally used concept of feature saliency could not predict the susceptibility to inattentional deafness. One explanation might be that our measures were not the most adequate ones to capture those effects of feature similarity or feature saliency. However, as we took great care of addressing subjective as well as established objective measures, this line of argument does not seem to be very probable. Another straightforward reason could be that as soon as a certain level of perceptual load (in our case: the counting tasks) occupies our resources, we are susceptible to inattentional deafness because our attentional resources are becoming too limited. So, we will only detect the auditory “Gorillas” if our attention strays from the primary task leaving enough cognitive resources available. Indeed, we were able to document an impact of attentional focus: Whereas the subjective rating of the ability to focus attention on the specific task by the participants themselves showed no effect, the objective measure of deviance from the counting task could be found to go along with a higher frequency of inattentional deafness. A conclusion of a cause–effect relationship could not be drawn from this result, however.

Although the overall mean frequency of inattentional deafness of 31.2% may seem relatively low in comparison to findings of other authors (e.g., 57% in Koreimann et al., 2014), the results are highly remarkable. It has to be taken into account that not only did inattentional deafness appear in all but two participants and had a frequency rate of 50% or more in ten participants, but it also appeared in musical pieces that were presented after an animal sound in a previous musical piece had already been detected! More than a third (36 out of 101) of the missed animal sounds was missed after the previous detection of a different animal sound. The effect of inattentional deafness was so strong that it even appeared after participants had the chance to develop first ideas regarding the aim of the conducted experiment and to build up an expectation of more animal sounds appearing in the following musical pieces.

Whereas most of the participants expressed genuine surprise when hearing the animal sounds, because they perceived the animal sounds as very salient, other participants reported that they blended in very well with the music and could possibly be mistaken for belonging to the musical piece. This finding confirmed the decision to exclude all unknown musical pieces from further analysis to ensure that the appearance of inattentional deafness was not only due to interpreting the animal sounds as belonging to the musical piece.

Overall, very conservative and strict criteria for actual misses were applied. Animal sounds were also not counted as missed if participants reported hearing the animal sound only once after first presentation of the musical piece for the possibility that this resulted from a lack of remembering how often the animal sound was noticed rather than from inattentional deafness in one of the two appearances of animal sounds within the musical pieces. Considering these limitations in classifying inattentional deafness, the rate of approximately 31.2% is all the more impressive. Not only did inattentional deafness occur in almost all participants and all musical pieces (with very diverse rates of courses), in some participants, the phenomenon also extended over several musical pieces. To the best of the authors’ knowledge, no other study has revealed repeated susceptibility to inattentional deafness before. The likeliness of the appearance of inattentional deafness was analyzed regarding differences between people and differences between characteristics of the musical pieces. Not all findings investigated in previous studies could be transferred to the domain of inattentional deafness in music. Yet the focus of the present study lay more on investigating a possible influence of previously unattended facets and generating new hypotheses for future studies. Due to the considerably high heterogeneity of the musical pieces as well as the respective animal sounds used in this study and the high overlapping of different variables, it cannot be ruled out that parameters other than the attended variables played an important role in the frequency of inattentional deafness. Future research should attempt to identify the underlying mechanisms and musical properties that are involved in inducing or constraining inattentional deafness going beyond concepts like conceptual load or feature similarity. Special focus should be directed on the question why some individuals are especially and even repeatedly prone to the phenomenon of inattentional deafness. Individuals’ ability to concentrate and narrow attention to one task as well as differences in motivation and the eagerness to succeed, should hereby be investigated further in order to precisely identify the underlying mechanisms.