Neural Mechanisms Underlying the Auditory Looming Bias

Abstract

Our auditory system constantly keeps track of our environment, informing us about our surroundings and warning us of potential threats. The auditory looming bias is an early perceptual phenomenon, reflecting higher alertness of listeners to approaching auditory objects, rather than to receding ones. Experimentally, this sensation has been elicited by using both intensity-varying stimuli, as well as spectrally varying stimuli with constant intensity. Following the intensity-based approach, recent research delving into the cortical mechanisms underlying the looming bias argues for top-down signaling from the prefrontal cortex to the auditory cortex in order to prioritize approaching over receding sonic motion. We here test the generalizability of that finding to spectrally induced looms by re-analyzing previously published data. Our results indicate the promoted top-down projection but at time points slightly preceding the motion onset and thus considered to reflect a bias driven by anticipation. At time points following the motion onset, our findings show a bottom-up bias along the dorsal auditory pathway directed toward the prefrontal cortex.

Keywords: connectivity; electroencephalography; looming.

Figure 1

Magnitude responses of listener-specific stimuli as a result of spectral contrast factor C manipulation. Different factor values correspond to different simulated positions in space: C = 1 for an external position (red), C = 0 for a perceived location inside the listener's head (blue), C = 0.5 an intermediate of the two (green). [Adapted from Baumgartner et al. (2017)].

Figure 2

The spectral contrast factor C was manipulated to create looming and receding sounds. Noise filtered with factor C₁ was cross-faded to C₂ with a temporal jitter of ±50 ms. Listeners were asked to report whether the sound was approaching, receding or static during the open response period. [Adapted from Baumgartner et al. (2017)].

Figure 3

Violin and box plots depicting the median, interquartile range, and density of the looming bias of response times in ms (magenta, left x-axis) and accuracies in % (cyan, right x-axis) depending on the different step sizes of spectral contrast manipulation as well as predictability of the second stimulus on the y-axis.

Figure 4

Group-average (N = 15) source strength responses in PAC elicited for trials with looming and receding stimuli. The PAC exhibits stereotypical auditory ERP dynamics (cluster-based permutation test).

Figure 5

Brain maps highlighting the regions that show a significant emergence of the looming bias in the left and right hemisphere, averaged over the time window between 120 and 200 ms (Baumgartner et al., 2017). Differences are mostly localized in temporal, parietal and frontal regions.

Figure 6

Group average connectivity values (dPTE, SEM, N = 15) between (a) PFC and PAC, (b) PFC and IPL and (c) IPL and PAC based on the individual single-trial source activities for looming and receding trials plotted against the center of the sliding time window (duration 250 ms). Significant occurrences of the looming bias are apparent at times around as well as before the stimulus change (vertical colored bars, p < 0.05; colors of the bars correspond to regions in fig. 7). Positive dPTE values represent connectivity in the targeted direction, indicated by the plot title, while negative ones denote the opposite direction.

Figure 7

Average dPTE connectivity values over the significant occurrences of the looming bias, calculated as the difference between dPTE for looming and dPTE for receding stimuli, based on the single-trial source activities and plotted against the center of the sliding time window (duration 250 ms). Additionally to the connections presented above (Figure 6(a,b,c)) significant connections are found from PAC, PostCG and PreCG to SFG. These additional significant occurrences of the looming bias are apparent at times corresponding to the timing of maximum looming bias occurrence (120–200 ms), as indicated by Baumgartner et al. (2017). Positive dPTE values resent connectivity in the direction mention as the legend of each bar (PAC to SFG, PreCG to SFG, PostCG to SFG, SFG to POP, PFC to PAC, PFC to IPL and IPL to PAC), negative values in the opposite direction.