j_hartling/paper_2025
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Updated/made figure captions up to Fig. 7.

Browse Source
This commit is contained in:
j-hartling
2026-05-03 19:55:26 +02:00
parent 9c5811d97c
commit 69f172ff2c
8 changed files with 1299 additions and 1223 deletions
Download Patch File Download Diff File Expand all files Collapse all files

108
main.aux
View File

@@ -195,9 +195,9 @@
\newlabel{eq:bandpass}{{1}{5}{}{}{}}
\abx@aux@page{61}{5}
\abx@aux@page{62}{5}
\@writefile{lof}{\contentsline {figure}{\numberline {1}{\ignorespaces \textbf {Schematic organisation of the song recognition pathway in grasshoppers compared to the structure of the functional model pathway.} \textbf {a}:~Simplified course of the pathway in the grasshopper, from the tympanal membrane over receptor neurons, local interneurons, and ascending neurons further towards the supraesophageal ganglion. \textbf {b}:~Schematic of synaptic connections between the three neuronal populations within the metathoracic ganglion. \textbf {c}:~Network representation of neuronal connectivity. \textbf {d}:~Flow diagram of the different signal representations and transformations along the model pathway. All representations are time-varying. 1st half: Preprocessing stage (one-dimensional). 2nd half: Feature extraction stage (high-dimensional). }}{6}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {1}{\ignorespaces \textbf {Schematic organisation of the grasshopper song recognition pathway and structure of the functional model pathway.} \textbf {a}:~Simplified course of the pathway in the grasshopper, from the tympanal membrane over receptor neurons, local interneurons, and ascending neurons further towards the supraesophageal ganglion. \textbf {b}:~Schematic of synaptic connections between the three neuronal populations within the metathoracic ganglion. \textbf {c}:~Network representation of neuronal connectivity. \textbf {d}:~Flow diagram of consecutive signal representations~(boxes) and transformations~(arrows) along the model pathway. All representations are time-varying. 1st half: Preprocessing stage~(one-dimensional representation). 2nd half: Feature extraction stage~(high-dimensional representation). }}{6}{}\protected@file@percent }
\providecommand*\caption@xref[2]{\@setref\relax\@undefined{#1}}
\newlabel{fig:pathway}{{1}{6}{}{}{}}
\newlabel{fig:pathway}{{1}{6}{\textbf {Schematic organisation of the grasshopper song recognition pathway and structure of the functional model pathway.} \textbf {a}:~Simplified course of the pathway in the grasshopper, from the tympanal membrane over receptor neurons, local interneurons, and ascending neurons further towards the supraesophageal ganglion. \textbf {b}:~Schematic of synaptic connections between the three neuronal populations within the metathoracic ganglion. \textbf {c}:~Network representation of neuronal connectivity. \textbf {d}:~Flow diagram of consecutive signal representations~(boxes) and transformations~(arrows) along the model pathway. All representations are time-varying. 1st half: Preprocessing stage~(one-dimensional representation). 2nd half: Feature extraction stage~(high-dimensional representation). }{}{}}
\abx@aux@cite{0}{suga1960peripheral}
\abx@aux@segm{0}{0}{suga1960peripheral}
\abx@aux@cite{0}{gollisch2002energy}
@@ -222,8 +222,8 @@
\abx@aux@page{68}{7}
\abx@aux@page{69}{7}
\newlabel{eq:highpass}{{4}{7}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {2}{\ignorespaces \textbf {Representations of a song of \textit {O. rufipes} during the preprocessing stage.} \textbf {a}:~Bandpass-filtered tympanal signal. \textbf {b}:~Signal envelope. \textbf {c}:~Logarithmically scaled envelope. \textbf {d}:~Intensity-adapted envelope. }}{8}{}\protected@file@percent }
\newlabel{fig:pre}{{2}{8}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {2}{\ignorespaces \textbf {Representations of a song of \textit {O. rufipes} during the preprocessing stage.} \textbf {a}:~Bandpass filtered tympanal signal $x_{\text {filt}}(t)$. \textbf {b}:~Signal envelope $x_{\text {env}}(t)$. \textbf {c}:~Logarithmically compressed envelope $x_{\text {log}}(t)$. \textbf {d}:~Intensity-adapted envelope $x_{\text {adapt}}(t)$. }}{8}{}\protected@file@percent }
\newlabel{fig:stages_pre}{{2}{8}{\textbf {Representations of a song of \textit {O. rufipes} during the preprocessing stage.} \textbf {a}:~Bandpass filtered tympanal signal $x_{\text {filt}}(t)$. \textbf {b}:~Signal envelope $x_{\text {env}}(t)$. \textbf {c}:~Logarithmically compressed envelope $x_{\text {log}}(t)$. \textbf {d}:~Intensity-adapted envelope $x_{\text {adapt}}(t)$. }{}{}}
\@writefile{toc}{\contentsline {subsection}{\numberline {2.2}Feature extraction by individual neurons}{8}{}\protected@file@percent }
\newlabel{eq:conv}{{5}{8}{}{}{}}
\newlabel{eq:gabor}{{6}{8}{}{}{}}
@@ -234,68 +234,68 @@
\abx@aux@cite{0}{bhavsar2017brain}
\abx@aux@segm{0}{0}{bhavsar2017brain}
\@writefile{lot}{\contentsline {table}{\numberline {1}{\ignorespaces Values of phase $\phi $ that are specific for the four major groups of Gabor kernels.}}{9}{}\protected@file@percent }
\newlabel{tab:gabor_phases}{{1}{9}{}{}{}}
\newlabel{tab:gabor_phases}{{1}{9}{Values of phase $\phi $ that are specific for the four major groups of Gabor kernels.}{}{}}
\newlabel{eq:binary}{{9}{9}{}{}{}}
\abx@aux@page{70}{10}
\abx@aux@page{71}{10}
\abx@aux@page{72}{10}
\newlabel{eq:lowpass}{{10}{10}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {3}{\ignorespaces \textbf {Representations of a song of \textit {O. rufipes} during the feature extraction stage.} Different colors indicate Gabor kernels with different lobe number $n$ and sign, with lighter colors for higher $n$~($1\,\leq \,n\,\leq \,4$; both $+$ and $-$ per $n$; two kernel widths $\sigma $ of $4\,$ms and $32\,$ms per sign). \textbf {a}:~Kernel-specific filter responses. \textbf {b}:~Binary responses. \textbf {c}:~Finalized features. }}{10}{}\protected@file@percent }
\newlabel{fig:stages_feat}{{3}{10}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {3}{\ignorespaces \textbf {Representations of a song of \textit {O. rufipes} during the feature extraction stage.} Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; two kernel widths $\sigma $ of $4\,$ms and $32\,$ms per type; 8 types, 16 kernels in total). \textbf {a}:~Kernel-specific filter responses $c_i(t)$. \textbf {b}:~Binary responses $b_i(t)$. \textbf {c}:~Finalized features $f_i(t)$.}}{10}{}\protected@file@percent }
\newlabel{fig:stages_feat}{{3}{10}{\textbf {Representations of a song of \textit {O. rufipes} during the feature extraction stage.} Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; two kernel widths $\sigma $ of $4\,$ms and $32\,$ms per type; 8 types, 16 kernels in total). \textbf {a}:~Kernel-specific filter responses $c_i(t)$. \textbf {b}:~Binary responses $b_i(t)$. \textbf {c}:~Finalized features $f_i(t)$.}{}{}}
\@writefile{toc}{\contentsline {section}{\numberline {3}Two mechanisms driving the emergence of intensity-invariant song representation}{10}{}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.1}Full-wave rectification \& lowpass filtering}{11}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {4}{\ignorespaces \textbf {Intensity invariance by logarithmic compression and adaptation is restricted by the noise floor.} Synthetic input $x_{\text {filt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with (\textbf {c}{} and \textbf {d}) or without (\textbf {a}{} and \textbf {b}) additive noise component $\eta (t)$. Input $x_{\text {filt}}(t)$ is transformed into envelope $x_{\text {env}}(t)$, logarithmically compressed envelope $x_{\text {log}}(t)$, and intensity-adapted envelope $x_{\text {adapt}}(t)$. \textbf {Left}:~$x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ for different scales $\alpha $. \textbf {Right}:~Ratios of the standard deviation of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ relative to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {filt}}(t)=\eta (t)$. \textbf {a}{} and \textbf {b}:~Ideally, if $x_{\text {filt}}(t)=\alpha \cdot s(t)$, then $x_{\text {adapt}}(t)$ is intensity-invariant across all $\alpha $. \textbf {c}{} and \textbf {d}:~In practice, if $x_{\text {filt}}(t)=\alpha \cdot s(t)+\eta (t)$, the intensity invariance of $x_{\text {adapt}}(t)$ is limited to sufficiently large $\alpha $. Shaded area indicates saturation of $x_{\text {adapt}}(t)$ at $95\,\%$ curve span. }}{12}{}\protected@file@percent }
\newlabel{fig:inv_rect-lp}{{4}{12}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {4}{\ignorespaces \textbf {Rectification and lowpass filtering improves SNR but does not contribute to intensity invariance.} Input $x_{\text {raw}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is successively transformed into tympanal signal $x_{\text {filt}}(t)$ and envelope $x_{\text {env}}(t)$. Different line styles indicate different cutoff frequencies $f_{\text {cut}}$ of the lowpass filter extracting $x_{\text {env}}(t)$. \textbf {Top}:~Example representations of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$ for different $\alpha $. \textbf {a}:~Noiseless case. \textbf {b}:~Noisy case. \textbf {Bottom}:~Intensity metrics over a range of $\alpha $. \textbf {c}:~Noiseless case: Standard deviations $\sigma _x$ of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$. \textbf {d}:~Noisy case: Ratios of $\sigma _x$ of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$ to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {raw}}(t)=\eta (t)$. \textbf {e}:~Ratios of $\sigma _x$ to $\sigma _{\eta }$ of $x_{\text {env}}(t)$ as in \textbf {d} for different species (averaged over songs and recordings, see appendix Fig.\,\ref {fig:app_rect-lp}). }}{12}{}\protected@file@percent }
\newlabel{fig:rect-lp}{{4}{12}{\textbf {Rectification and lowpass filtering improves SNR but does not contribute to intensity invariance.} Input $x_{\text {raw}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is successively transformed into tympanal signal $x_{\text {filt}}(t)$ and envelope $x_{\text {env}}(t)$. Different line styles indicate different cutoff frequencies $f_{\text {cut}}$ of the lowpass filter extracting $x_{\text {env}}(t)$. \textbf {Top}:~Example representations of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$ for different $\alpha $. \textbf {a}:~Noiseless case. \textbf {b}:~Noisy case. \textbf {Bottom}:~Intensity metrics over a range of $\alpha $. \textbf {c}:~Noiseless case: Standard deviations $\sigma _x$ of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$. \textbf {d}:~Noisy case: Ratios of $\sigma _x$ of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$ to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {raw}}(t)=\eta (t)$. \textbf {e}:~Ratios of $\sigma _x$ to $\sigma _{\eta }$ of $x_{\text {env}}(t)$ as in \textbf {d} for different species (averaged over songs and recordings, see appendix Fig.\,\ref {fig:app_rect-lp}). }{}{}}
\@writefile{toc}{\contentsline {subsection}{\numberline {3.2}Logarithmic compression \& spike-frequency adaptation}{12}{}\protected@file@percent }
\newlabel{eq:toy_env}{{11}{13}{}{}{}}
\newlabel{eq:toy_log}{{12}{13}{}{}{}}
\newlabel{eq:toy_highpass}{{13}{13}{}{}{}}
\newlabel{eq:toy_snr}{{14}{13}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {5}{\ignorespaces \textbf {Intensity invariance by logarithmic compression and adaptation is restricted by the noise floor.} Synthetic input $x_{\text {filt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with (\textbf {c}{} and \textbf {d}) or without (\textbf {a}{} and \textbf {b}) additive noise component $\eta (t)$. Input $x_{\text {filt}}(t)$ is transformed into envelope $x_{\text {env}}(t)$, logarithmically compressed envelope $x_{\text {log}}(t)$, and intensity-adapted envelope $x_{\text {adapt}}(t)$. \textbf {Left}:~$x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ for different scales $\alpha $. \textbf {Right}:~Ratios of the standard deviation of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ relative to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {filt}}(t)=\eta (t)$. \textbf {a}{} and \textbf {b}:~Ideally, if $x_{\text {filt}}(t)=\alpha \cdot s(t)$, then $x_{\text {adapt}}(t)$ is intensity-invariant across all $\alpha $. \textbf {c}{} and \textbf {d}:~In practice, if $x_{\text {filt}}(t)=\alpha \cdot s(t)+\eta (t)$, the intensity invariance of $x_{\text {adapt}}(t)$ is limited to sufficiently large $\alpha $. Shaded area indicates saturation of $x_{\text {adapt}}(t)$ at $95\,\%$ curve span. }}{14}{}\protected@file@percent }
\newlabel{fig:inv_log-hp}{{5}{14}{}{}{}}
\@writefile{toc}{\contentsline {subsection}{\numberline {3.3}Thresholding nonlinearity \& temporal averaging}{14}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {6}{\ignorespaces \textbf {Intensity invariance by thresholding and temporal averaging depends on both the threshold value and the noise floor.} Synthetic input $x_{\text {adapt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with additive noise component $\eta (t)$. Input $x_{\text {adapt}}(t)$ is transformed into kernel response $c_i(t)$, binary response $b_i(t)$, and feature $f_i(t)$. Threshold value $\Theta _i$ is set to multiples of the reference standard deviation $\sigma _{\eta }$ of $c_i(t)$ for input $x_{\text {adapt}}(t)=\eta (t)$. Darker colors correspond to higher $\Theta _i$. \textbf {Left}:~$x_{\text {adapt}}(t)$, $c_i(t)$, $b_i(t)$, and $f_i(t)$ for different scales $\alpha $ and threshold values $\Theta _i$. Left-most column is is the pure-noise reference. \textbf {Right}:~Average value of $f_i(t)$ during the song for the different $\Theta _i$. \textbf {a}:~Input $x_{\text {adapt}}(t)$. \textbf {b}-\textbf {d}:~$c_i(t)$, $b_i(t)$, and $f_i(t)$ for the different $\Theta _i$ based on the same $x_{\text {adapt}}(t)$ from \textbf {a}{}. \textbf {e}:~Average value of $f_i(t)$ during the song for the different $\Theta _i$ in \textbf {b}{}-\textbf {d}. }}{15}{}\protected@file@percent }
\newlabel{fig:inv_thresh-lp_single}{{6}{15}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {7}{\ignorespaces \textbf {Feature representation of different species-specific songs saturates at different points in feature space.} }}{16}{}\protected@file@percent }
\newlabel{fig:inv_thresh-lp_species}{{7}{16}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {8}{\ignorespaces \textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }}{17}{}\protected@file@percent }
\newlabel{fig:inv_full}{{8}{17}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {9}{\ignorespaces \textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }}{18}{}\protected@file@percent }
\newlabel{fig:inv_short}{{9}{18}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {10}{\ignorespaces \textbf {Inter- and intraspecific feature variability.} }}{19}{}\protected@file@percent }
\newlabel{fig:cross_species}{{10}{19}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {11}{\ignorespaces \textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }}{20}{}\protected@file@percent }
\newlabel{fig:inv_field}{{11}{20}{}{}{}}
\newlabel{eq:pdf_split}{{15}{21}{}{}{}}
\newlabel{eq:feat_avg}{{16}{21}{}{}{}}
\newlabel{eq:feat_prop}{{17}{21}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {5}{\ignorespaces \textbf {Intensity invariance through logarithmic compression and adaptation is restricted by the noise floor and decreases SNR.} Input $x_{\text {filt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is successively transformed into envelope $x_{\text {env}}(t)$, logarithmically compressed envelope $x_{\text {log}}(t)$, and intensity-adapted envelope $x_{\text {adapt}}(t)$. \textbf {Top}:~Example representations of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ for different $\alpha $. \textbf {a}:~Noiseless case. \textbf {b}:~Noisy case. \textbf {Bottom}:~Intensity metrics over a range of $\alpha $. \textbf {c}:~Noiseless case: Standard deviations $\sigma _x$ of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$. \textbf {d}:~Noisy case: Ratios of $\sigma _x$ of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {filt}}(t)=\eta (t)$. Shaded areas indicate $5\,\%$ (dark grey) and $95\,\%$ (light grey) curve span for $x_{\text {adapt}}(t)$. \textbf {e}:~Ratios of $\sigma _x$ to $\sigma _{\eta }$ of $x_{\text {adapt}}(t)$ as in \textbf {d} for different species (averaged over songs and recordings, see appendix Fig\,\ref {fig:app_log-hp_curves}). Dots indicate $95\,\%$ curve span per species. }}{14}{}\protected@file@percent }
\newlabel{fig:log-hp}{{5}{14}{\textbf {Intensity invariance through logarithmic compression and adaptation is restricted by the noise floor and decreases SNR.} Input $x_{\text {filt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is successively transformed into envelope $x_{\text {env}}(t)$, logarithmically compressed envelope $x_{\text {log}}(t)$, and intensity-adapted envelope $x_{\text {adapt}}(t)$. \textbf {Top}:~Example representations of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ for different $\alpha $. \textbf {a}:~Noiseless case. \textbf {b}:~Noisy case. \textbf {Bottom}:~Intensity metrics over a range of $\alpha $. \textbf {c}:~Noiseless case: Standard deviations $\sigma _x$ of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$. \textbf {d}:~Noisy case: Ratios of $\sigma _x$ of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {filt}}(t)=\eta (t)$. Shaded areas indicate $5\,\%$ (dark grey) and $95\,\%$ (light grey) curve span for $x_{\text {adapt}}(t)$. \textbf {e}:~Ratios of $\sigma _x$ to $\sigma _{\eta }$ of $x_{\text {adapt}}(t)$ as in \textbf {d} for different species (averaged over songs and recordings, see appendix Fig\,\ref {fig:app_log-hp_curves}). Dots indicate $95\,\%$ curve span per species. }{}{}}
\@writefile{toc}{\contentsline {subsection}{\numberline {3.3}Thresholding nonlinearity \& temporal averaging}{15}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {6}{\ignorespaces \textbf {Intensity invariance through thresholding and temporal averaging is mediated by the interaction of threshold value and noise floor.} Input $x_{\text {adapt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is transformed into single kernel response $c(t)$, binary response $b(t)$, and feature $f(t)$. Different color shades indicate different threshold values $\Theta $ (multiples of reference standard deviation $\sigma _{\eta }$ of $c(t)$ for input $x_{\text {adapt}}(t)=\eta (t)$, with darker colors for higher $\Theta $). \textbf {Left}:~Noisy case: Example representations of $x_{\text {adapt}}(t)$ as well as $c(t)$, $b(t)$, and $f(t)$ for different $\alpha $. \textbf {a}:~$x_{\text {adapt}}(t)$ with kernel $k(t)$ in black. \textbf {b\,-\,d}: $c(t)$, $b(t)$, and $f(t)$ based on the same $x_{\text {adapt}}(t)$ from \textbf {a} but with different $\Theta $. \textbf {Right}:~Average value $\mu _f$ of $f(t)$ for each $\Theta $ from \textbf {b\,-\,d}, once for the noisy case (solid lines) and once for the noiseless case (dotted lines). Dots indicate $95\,\%$ curve span (noisy case). \textbf {e}:~$\mu _f$ over a range of $\alpha $. \textbf {f}:~$\mu _f$ over the standard deviation of noisy input $x_{\text {adapt}}$ corresponding to the values of $\alpha $ shown in \textbf {e}. }}{16}{}\protected@file@percent }
\newlabel{fig:thresh-lp_single}{{6}{16}{\textbf {Intensity invariance through thresholding and temporal averaging is mediated by the interaction of threshold value and noise floor.} Input $x_{\text {adapt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is transformed into single kernel response $c(t)$, binary response $b(t)$, and feature $f(t)$. Different color shades indicate different threshold values $\Theta $ (multiples of reference standard deviation $\sigma _{\eta }$ of $c(t)$ for input $x_{\text {adapt}}(t)=\eta (t)$, with darker colors for higher $\Theta $). \textbf {Left}:~Noisy case: Example representations of $x_{\text {adapt}}(t)$ as well as $c(t)$, $b(t)$, and $f(t)$ for different $\alpha $. \textbf {a}:~$x_{\text {adapt}}(t)$ with kernel $k(t)$ in black. \textbf {b\,-\,d}: $c(t)$, $b(t)$, and $f(t)$ based on the same $x_{\text {adapt}}(t)$ from \textbf {a} but with different $\Theta $. \textbf {Right}:~Average value $\mu _f$ of $f(t)$ for each $\Theta $ from \textbf {b\,-\,d}, once for the noisy case (solid lines) and once for the noiseless case (dotted lines). Dots indicate $95\,\%$ curve span (noisy case). \textbf {e}:~$\mu _f$ over a range of $\alpha $. \textbf {f}:~$\mu _f$ over the standard deviation of noisy input $x_{\text {adapt}}$ corresponding to the values of $\alpha $ shown in \textbf {e}. }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {7}{\ignorespaces \textbf {Feature representation of different species-specific songs saturates at different points in feature space.} Same input and processing as in Fig.\,\ref {fig:thresh-lp_single} but with three different kernels $k_i$, each with a single kernel-specific threshold value $\Theta _i=0.5\cdot \sigma _{\eta _i}$. \textbf {a}:~Examples of species-specific grasshopper songs. \textbf {Middle}:~Average value $\mu _{f_i}$ of each feature $f_i(t)$ over $\alpha $ per species (averaged over songs and recordings, see appendix Figs.\,\ref {fig:app_thresh-lp_pure} and \ref {fig:app_thresh-lp_noise}). Different color shades indicate different kernels $k_i$. Dots indicate $95\,\%$ curve span per $k_i$. \textbf {b}:~Noiseless case. \textbf {c}:~Noisy case. \textbf {Bottom}:~2D feature spaces spanned by each pair of $f_i(t)$. Each trajectory corresponds to a species-specific combination of $\mu _{f_i}$ that develops with $\alpha $ (colorbars). Horizontal dashes in the colorbar indicate $5\,\%$ (dark grey) and $95\,\%$ (light grey) curve span of the norm across all three $\mu _{f_i}$ per species. \textbf {d}:~Noiseless case. \textbf {e}:~Noisy case. Shaded areas }}{17}{}\protected@file@percent }
\newlabel{fig:thresh-lp_species}{{7}{17}{\textbf {Feature representation of different species-specific songs saturates at different points in feature space.} Same input and processing as in Fig.\,\ref {fig:thresh-lp_single} but with three different kernels $k_i$, each with a single kernel-specific threshold value $\Theta _i=0.5\cdot \sigma _{\eta _i}$. \textbf {a}:~Examples of species-specific grasshopper songs. \textbf {Middle}:~Average value $\mu _{f_i}$ of each feature $f_i(t)$ over $\alpha $ per species (averaged over songs and recordings, see appendix Figs.\,\ref {fig:app_thresh-lp_pure} and \ref {fig:app_thresh-lp_noise}). Different color shades indicate different kernels $k_i$. Dots indicate $95\,\%$ curve span per $k_i$. \textbf {b}:~Noiseless case. \textbf {c}:~Noisy case. \textbf {Bottom}:~2D feature spaces spanned by each pair of $f_i(t)$. Each trajectory corresponds to a species-specific combination of $\mu _{f_i}$ that develops with $\alpha $ (colorbars). Horizontal dashes in the colorbar indicate $5\,\%$ (dark grey) and $95\,\%$ (light grey) curve span of the norm across all three $\mu _{f_i}$ per species. \textbf {d}:~Noiseless case. \textbf {e}:~Noisy case. Shaded areas }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {8}{\ignorespaces \textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }}{18}{}\protected@file@percent }
\newlabel{fig:pipeline_full}{{8}{18}{\textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {9}{\ignorespaces \textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }}{19}{}\protected@file@percent }
\newlabel{fig:pipeline_short}{{9}{19}{\textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {10}{\ignorespaces \textbf {Inter- and intraspecific feature variability.} }}{20}{}\protected@file@percent }
\newlabel{fig:feat_cross_species}{{10}{20}{\textbf {Inter- and intraspecific feature variability.} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {11}{\ignorespaces \textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }}{21}{}\protected@file@percent }
\newlabel{fig:pipeline_field}{{11}{21}{\textbf {Step-wise emergence of intensity invariant song representation along the model pathway.} }{}{}}
\newlabel{eq:pdf_split}{{15}{22}{}{}{}}
\newlabel{eq:feat_avg}{{16}{22}{}{}{}}
\newlabel{eq:feat_prop}{{17}{22}{}{}{}}
\abx@aux@cite{0}{stumpner1991auditory}
\abx@aux@segm{0}{0}{stumpner1991auditory}
\@writefile{toc}{\contentsline {section}{\numberline {4}Discriminating species-specific song\\patterns in feature space}{22}{}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {5}Conclusions \& outlook}{22}{}\protected@file@percent }
\abx@aux@page{73}{22}
\@writefile{lof}{\contentsline {figure}{\numberline {12}{\ignorespaces \textbf {} }}{24}{}\protected@file@percent }
\newlabel{}{{12}{24}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {13}{\ignorespaces \textbf {} }}{24}{}\protected@file@percent }
\newlabel{}{{13}{24}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {14}{\ignorespaces \textbf {} }}{25}{}\protected@file@percent }
\newlabel{}{{14}{25}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {15}{\ignorespaces \textbf {} }}{25}{}\protected@file@percent }
\newlabel{}{{15}{25}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {16}{\ignorespaces \textbf {} }}{26}{}\protected@file@percent }
\newlabel{}{{16}{26}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {17}{\ignorespaces \textbf {} }}{26}{}\protected@file@percent }
\newlabel{}{{17}{26}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {18}{\ignorespaces \textbf {} }}{27}{}\protected@file@percent }
\newlabel{}{{18}{27}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {19}{\ignorespaces \textbf {} }}{27}{}\protected@file@percent }
\newlabel{}{{19}{27}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {20}{\ignorespaces \textbf {} }}{28}{}\protected@file@percent }
\newlabel{}{{20}{28}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {21}{\ignorespaces \textbf {} }}{28}{}\protected@file@percent }
\newlabel{}{{21}{28}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {22}{\ignorespaces \textbf {} }}{29}{}\protected@file@percent }
\newlabel{}{{22}{29}{}{}{}}
\@writefile{toc}{\contentsline {section}{\numberline {4}Discriminating species-specific song\\patterns in feature space}{23}{}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {5}Conclusions \& outlook}{23}{}\protected@file@percent }
\abx@aux@page{73}{23}
\@writefile{lof}{\contentsline {figure}{\numberline {12}{\ignorespaces \textbf {} }}{25}{}\protected@file@percent }
\newlabel{fig:app_env-sd}{{12}{25}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {13}{\ignorespaces \textbf {} }}{25}{}\protected@file@percent }
\newlabel{fig:app_rect-lp}{{13}{25}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {14}{\ignorespaces \textbf {} }}{26}{}\protected@file@percent }
\newlabel{fig:app_log-hp_curves}{{14}{26}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {15}{\ignorespaces \textbf {} }}{26}{}\protected@file@percent }
\newlabel{fig:app_log-hp_saturation}{{15}{26}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {16}{\ignorespaces \textbf {} }}{27}{}\protected@file@percent }
\newlabel{fig:app_thresh-lp_pure}{{16}{27}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {17}{\ignorespaces \textbf {} }}{27}{}\protected@file@percent }
\newlabel{fig:app_thresh-lp_noise}{{17}{27}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {18}{\ignorespaces \textbf {} }}{28}{}\protected@file@percent }
\newlabel{fig:app_thresh-lp_kern-sd}{{18}{28}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {19}{\ignorespaces \textbf {} }}{28}{}\protected@file@percent }
\newlabel{fig:app_full_kern-sd}{{19}{28}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {20}{\ignorespaces \textbf {} }}{29}{}\protected@file@percent }
\newlabel{fig:app_short_kern-sd}{{20}{29}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {21}{\ignorespaces \textbf {} }}{29}{}\protected@file@percent }
\newlabel{fig:app_field_kern-sd}{{21}{29}{\textbf {} }{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {22}{\ignorespaces \textbf {} }}{30}{}\protected@file@percent }
\newlabel{fig:app_cross_species_thresh}{{22}{30}{\textbf {} }{}{}}
\gdef\svg@ink@ver@settings{{\m@ne }{inkscape}{\m@ne }}
\abx@aux@read@bbl@mdfivesum{1380DC8C93D2855FDB132CC5A40AD52F}
\gdef \@abspage@last{29}
\abx@aux@read@bbl@mdfivesum{9F6C5EB9B092C01D1CB433A0E1BC4A06}
\gdef \@abspage@last{30}