Spatiotemporal patterning of acoustic gaze in echolocating bats navigating gaps in clutter

Abstract

We challenged four big brown bats to maneuver through abrupt turns in narrow corridors surrounded by dense acoustic clutter. We quantified bats' performance, sonar beam focus, and sensory acquisition rate. Performance was excellent in straight corridors, with sonar beam aim deviating less than 5° from the corridor midline. Bats anticipated an upcoming abrupt turn to the right or left by slowing flight speed and shifting beam aim to "look" proactively into one side of the corridor to identify the new flightpath. All bats mastered the right turn, but two bats consistently failed the left turn. Bats increased their sensory acquisition rate when confronting abrupt turns in both successful and failed flights. Limitations on biosonar performance reflected failures to switch beam aim and to modify a learned spatial map, rather than failures to update acquisition rate.

Keywords: Animals; Biological Sciences; Ethology; Zoology.

Graphical abstract

Figure 1

Diagram of the flight room and chain array Dimensions of the flight room (m; left y and x axes) and the location of the hanging chains (filled circles) are shown. Bats were released through an opening in hanging felt (small filled rectangle) at the release point (black X) and flew down a corridor (40 cm width) through the chain array. Chains were hung in rows and columns 20 cm apart. In the Straight condition, chains were removed (open circles) from the section outlined in green, and both the red and blue sections were blocked with chains. Bats were rewarded for landing on Wall B. In the Right Turn condition, chains were removed from the blue section of the chain array, and the green and red sections were blocked with chains. Bats were rewarded for landing on Wall C. In the Left Turn condition, chains were removed from the red section, and the blue and green sections were blocked with chains. Bats were rewarded for landing on Wall A. In the Reversed Right Turn condition, bats were released from the blue X on Wall B and flew through the green and red sections, landing on Wall A. The blue section was blocked by chains. The colored lines on the right side of the array marked beam 1 (brown) and beam 2 (dark purple) show the two segments of the flightpath used for calculating beam aim. These two segments are referred to distance from the turn, as shown by the gray numbers; the turn is at 0 cm and the beginning of measured flightpath is at 300 cm. The colored lines marked speed 1 (orange) and speed 2 (purple) show the two segments of the flightpath used to calculate flight speed. Small numbered boxes around the perimeter of the room show the locations of recording microphones, and boxes labeled “v” show the position of recording cameras. Black dotted lines indicate a net that prevented the bat from flying into the rest of the room, and recording equipment was located behind the curved gray dotted line.

Figure 2

Performance of individual bats on each flight day in the three different task conditions Bats performed 15–18 flights per day. All bats showed improved performance (increased percentage of successful flights) with experience in Right Turn flights; only two bats showed improved performance with experience in Left Turn flights.

Figure 3

Mean and standard deviation (SD) of beam aim of all calls emitted in successful flights (A–C) (A) Straight flights; (B) Right Turn flights; (C) Left Turn flights. The y axis shows the portion of the flightpath (see beam 2 line, Figure 1) immediately before the turn (or equivalent point in Straight flights). The position of the turn (or equivalent location in straight flights) is shown by the green, blue, and red dashed rectangles in (A, B, and C), respectively. Filled circles show the positions of hanging chains in each condition. Colored vectors plotted along the midline of the chain array show mean beam aim calculated in 10-cm bins. Colored envelopes around these vectors show the SD within each window. Blue vectors denote positive beam aims >0°, red vectors denote negative beam aims <0°. N = Numbers of calls used to calculate beam aim. In (B and C), red horizontal arrows denote the area in which beam aim diverges more than 1 SD from the midline.

Figure 4

Beam angle of calls emitted from 150–0 cm before the turn (or equivalent location in Straight flights) on first and last day of each flight condition Each panel shows beam angles for the first day of flights in that condition (open black circles), beam angles for the last day of flights in that condition (red crosses), and regression lines (first day, black; last day, red) through the data points. Distance to the turn (150–0 cm) is shown on the x axis. y Axis shows beam angle (positive values show shifts toward the right, negative values show shifts toward the left). Asterisks next to “Last day” indicate that beam aim regression in the final day was significantly different from that in the first day. Plots show beam aim values from both successful and failed flights, with the exception of Left Turn flights for Bats 3 and 4, which show values from failed flights only.

Figure 5

Results of linear mixed effects models (A and B) Linear mixed effects model predictions of changes in IPI in successful flights as the bat approaches the turn (A) and as the bat completes multiple days of flights in the same condition (B). Asterisks indicate that slopes for the turn flights are significantly different from the slope for Straight flights.

Figure 6

Spectrogram of calls emitted by Bat 1 during a single successful flight through the Straight corridor SSGs were categorized using the algorithm devised by Kothari et al. (2014) and modified by Warnecke et al. (2016, . White labels and numbers underneath calls identify these SSG categorizations and the number of pulses within each group. White triangles show the bat's position relative to the point at which the position of the turn (at 0 cm) in the Turn conditions (beam 2 in Figure 1). After flying past the 0 cm mark, the bat lowers the frequency of its pulses in preparation for landing on the wall.

Figure 7

Proportions of calls emitted as Sonar Sound Groups (SSGs) Bats emitted the majority of their calls in SSGs (as opposed to single calls) in all flight conditions, although the relative proportion varies. SSG of 1 (blue) = single call; 2 (pink) = doublet; 3 (yellow) = triplet; 4 (green) = quadruplet. Solid outlined bars show data from successful flights. The dashed outlined bars (Bat 3 and Bat 4, Left Turn) show failed flights. SSG/single call proportions in Turn flights are significantly different from SSG/single call proportions in Straight flights for all bats (McNemar chi-square tests, p < 0.001).

Figure 8

The distribution of pre-IPI and post-IPI of all calls emitted in all flights Calls along the gray solid diagonal line have equal pre- and post-IPI values. Plots with blue symbols show IPIs in failed Left Turn flights for Bat 3 and Bat 4. Bat 2 in the Left Turn condition emitted fewer calls, so the plot seems less dense. These plots reveal identifiable and consistent qualitative differences in the timing patterns that individual bats employ when navigating through clutter. Note persistence of distinctive plot patterns for each bat across conditions.