QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online May 30, 2008
Journal of Experimental Biology 211, 1850-1858 (2008)
Published by The Company of Biologists 2008
doi: 10.1242/jeb.017715

This Article Summary Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Goerlitz, H. R. Articles by Wiegrebe, L. Search for Related Content PubMed PubMed Citation Articles by Goerlitz, H. R. Articles by Wiegrebe, L. Social Bookmarking                         What's this?

Comparing passive and active hearing: spectral analysis of transient sounds in bats

Holger R. Goerlitz^*, Mathias Hübner and Lutz Wiegrebe

Department Biologie II, Neurobiologie, Ludwig-Maximilians-Universität München, Großhadernerstrasse 2, 82152 Martinsried, Germany

View larger version (28K): [in this window] [in a new window]   Fig. 1. Filter transfer functions and amplitude spectra of example stimuli. (A) Experiment 1 (passive hearing). (B) Experiment 2 (echolocation). The graphs on the left show the transfer functions of the filters used to generate the foreground stimuli and the noise background. The graphs on the right show examples of the amplitude spectra of the finally presented stimuli, i.e. after convolution of the filters with a foreground signal (an impulse or a recorded echolocation call) and with white noise. FS, full scale. Thick lines, noise background (filter slope of 0 dB/octave or –3 dB/octave); thin lines, foreground stimuli (filter slopes from –6 dB/octave to +6 dB/octave). The amplitudes of the spectra do not illustrate the amplitude of the finally presented stimuli, which cannot be easily compared and additionally depend on the applied attenuation and on the call amplitude.

View larger version (28K): [in this window] [in a new window]   Fig. 2. Experiment 1 (passive hearing): classification of filtered impulses by three bats (A–C) and their means ± s.e.m. (D). The performance is plotted as percentage of highpass classification versus the filter slope. Trials with presentation of white-noise background are shown with black symbols, and trials with pink-noise background are shown with red symbols. The two training/control conditions (±6 dB/octave and white noise) are indicated with open symbols. The grey bars show the difference in highpass classification between pink- and white-noise background. Their mean difference across all six filter slopes (indicated by M) is shown as a black bar. The perceptual classification boundaries per background are indicated by vertical lines in the lower part of each panel (black, white-noise background; red, pink-noise background). The shift in the perceptual classification boundary from white- to pink-noise background is given above the vertical lines in units of dB/octave. Only the mean perceptual classification boundary could be tested for significance. *P<0.05, **P<0.01.

View larger version (21K): [in this window] [in a new window]   Fig. 3. Experiment 2 (echolocation): classification of filtered echolocation calls by two bats (A,B) and their means ± s.e.m. (C) during presentation of white-noise (black symbols) and pink-noise background (red symbols). Notation as in Fig. 2. *P<0.05, ***P<0.001.

View larger version (32K): [in this window] [in a new window]   Fig. 4. Experiment 2 (echolocation): influence of the roving level on signal classification. Classification of the same filtered echoes by the same bats as in Fig. 3, but here plotted as a function of echo level. (A,B) Control signals (±6 dB/octave, white-noise background). (C,D) Test signals (white noise: ±1.2 and 3.6 dB/octave, pink noise: ±1.2, 3.6 and 6.0 dB/octave). Numbers above the data points give the number of trials.

View larger version (66K): [in this window] [in a new window]   Fig. 5. Call parameters of Phyllostomus discolor in Experiment 2 (echolocation). (A) Histogram of the durations of the analysed calls. Median and quartiles are indicated by the vertical lines. (B) One example call of Bat 4 plotted as spectrogram (top), as oscillogram (bottom) and as amplitude spectrum (right). The amplitude spectrum was calculated from the section between the vertical lines in the time signal. (C–H) Spectral parameters. For analysis, we used the five calls with the highest signal-to-noise ratio per sequence and calculated the mean per sequence of every parameter. The means per sequence were then grouped, either per stimulus (i.e. the combinations of filter slope and noise backgrounds; the larger part of the panels, between –6.0 and +6.0 dB/octave; N=33–59 sequences per test stimulus, 766–1252 sequences per control stimulus) or per noise background (small right part of the panels, marked B; N=214–2694 sequences per background), and their second order mean ± s.e.m. was plotted. Note the different scales of the y-axes. Grey symbols, white-noise background; coloured symbols, pink-noise background. *P<0.05, **P<0.01, ***P<0.001.

View larger version (22K): [in this window] [in a new window]   Fig. 6. Calculated influence of the stimulus filters and the roving level on the frequency centroid and the level of the recorded calls (mean ± s.d. of 9500 calls for Bat 4 and 15055 calls for Bat 5). (A) The stimulus filters, going from lowpass (–6 dB/octave) to highpass (+6 dB/octave), increased the frequency centroid of the recorded calls by about 5 kHz. (B) The stimulus filters, going from lowpass (–6 dB/octave) to highpass (+6 dB/octave), increased the r.m.s. level of the recorded calls by 2–3 dB SPL. (C) Increasing roving levels, i.e. higher overall levels, reduced the frequency centroid of the recorded calls by about 2 kHz.

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

Twitter