Limits...
Probing real sensory worlds of receivers with unsupervised clustering.

Pfeiffer M, Hartbauer M, Lang AB, Maass W, Römer H - PLoS ONE (2012)

Bottom Line: Our results show that the reliability of burst coding in the time domain is so high that identical stimuli lead to extremely similar spike pattern responses, even for different preparations on different dates, and even if one of the preparations is recorded outdoors and the other one in the sound proof lab.Our study shows that burst coding can provide a reliable mechanism for acoustic insects to classify and discriminate signals under very noisy real-world conditions.This gives new insights into the neural mechanisms potentially used by bushcrickets to discriminate conspecific songs from sounds of predators in similar carrier frequency bands.

View Article: PubMed Central - PubMed

Affiliation: Institute for Theoretical Computer Science, TU Graz, Graz, Austria. pfeiffer@ini.phys.ethz.ch

ABSTRACT
The task of an organism to extract information about the external environment from sensory signals is based entirely on the analysis of ongoing afferent spike activity provided by the sense organs. We investigate the processing of auditory stimuli by an acoustic interneuron of insects. In contrast to most previous work we do this by using stimuli and neurophysiological recordings directly in the nocturnal tropical rainforest, where the insect communicates. Different from typical recordings in sound proof laboratories, strong environmental noise from multiple sound sources interferes with the perception of acoustic signals in these realistic scenarios. We apply a recently developed unsupervised machine learning algorithm based on probabilistic inference to find frequently occurring firing patterns in the response of the acoustic interneuron. We can thus ask how much information the central nervous system of the receiver can extract from bursts without ever being told which type and which variants of bursts are characteristic for particular stimuli. Our results show that the reliability of burst coding in the time domain is so high that identical stimuli lead to extremely similar spike pattern responses, even for different preparations on different dates, and even if one of the preparations is recorded outdoors and the other one in the sound proof lab. Simultaneous recordings in two preparations exposed to the same acoustic environment reveal that characteristics of burst patterns are largely preserved among individuals of the same species. Our study shows that burst coding can provide a reliable mechanism for acoustic insects to classify and discriminate signals under very noisy real-world conditions. This gives new insights into the neural mechanisms potentially used by bushcrickets to discriminate conspecific songs from sounds of predators in similar carrier frequency bands.

Show MeSH

Related in: MedlinePlus

Analysis of separability of bursts in response to artificial stimuli from other stimulus classes.A) Conditional entropy of class labels given the cluster indices, averaged over six recording sessions, three in which all 5 stimulus classes were used, and three in which only classes 1,2, and 4 were used (errorbars denote standard deviations). Classes of bursts with lower conditional entropy form more homogeneous clusters. Artificial stimuli that consist of temporally more structured and/or longer stimuli (classes 2,4, and 5) are better separable from noise or other stimuli than single pulse stimuli (classes 1 and 3). B) Confusion matrix for assigned cluster labels vs. actual labels of bursts. In every row we plot the average relative frequencies of burst labels occurrences in clusters that were assigned to one of the classes N ( = Noise) or 1–5. One can see that most mistakes are due to noise bursts assigned to one of the artificial stimulus classes. Also bursts in response to classes 2 and 3 are sometimes clustered together.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3368931&req=5

pone-0037354-g008: Analysis of separability of bursts in response to artificial stimuli from other stimulus classes.A) Conditional entropy of class labels given the cluster indices, averaged over six recording sessions, three in which all 5 stimulus classes were used, and three in which only classes 1,2, and 4 were used (errorbars denote standard deviations). Classes of bursts with lower conditional entropy form more homogeneous clusters. Artificial stimuli that consist of temporally more structured and/or longer stimuli (classes 2,4, and 5) are better separable from noise or other stimuli than single pulse stimuli (classes 1 and 3). B) Confusion matrix for assigned cluster labels vs. actual labels of bursts. In every row we plot the average relative frequencies of burst labels occurrences in clusters that were assigned to one of the classes N ( = Noise) or 1–5. One can see that most mistakes are due to noise bursts assigned to one of the artificial stimulus classes. Also bursts in response to classes 2 and 3 are sometimes clustered together.

Mentions: Figure 8A shows the average conditional entropy individually for every class, where the average is over all recordings sessions in which those stimuli were used (three sessions for all 5 stimuli, three sessions only for stimulus classes 1,2, and 4). Although these statistics are based on only six recording sessions, and the standard deviations are large, one can observe the same trend that was qualitatively visible from Figure 7. The average conditional entropy is low for classes of bursts in response to long and/or temporally structured stimuli (classes 2,4, and 5), and higher for the single short pulses of classes 1 and 3. This indicates that classes 2,4, and 5 can be better discriminated from other artificial or background signals than the single pulse stimuli. Due to the limited amount of available data, these results are statistically not significant, and more measurements would be required.


Probing real sensory worlds of receivers with unsupervised clustering.

Pfeiffer M, Hartbauer M, Lang AB, Maass W, Römer H - PLoS ONE (2012)

Analysis of separability of bursts in response to artificial stimuli from other stimulus classes.A) Conditional entropy of class labels given the cluster indices, averaged over six recording sessions, three in which all 5 stimulus classes were used, and three in which only classes 1,2, and 4 were used (errorbars denote standard deviations). Classes of bursts with lower conditional entropy form more homogeneous clusters. Artificial stimuli that consist of temporally more structured and/or longer stimuli (classes 2,4, and 5) are better separable from noise or other stimuli than single pulse stimuli (classes 1 and 3). B) Confusion matrix for assigned cluster labels vs. actual labels of bursts. In every row we plot the average relative frequencies of burst labels occurrences in clusters that were assigned to one of the classes N ( = Noise) or 1–5. One can see that most mistakes are due to noise bursts assigned to one of the artificial stimulus classes. Also bursts in response to classes 2 and 3 are sometimes clustered together.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3368931&req=5

pone-0037354-g008: Analysis of separability of bursts in response to artificial stimuli from other stimulus classes.A) Conditional entropy of class labels given the cluster indices, averaged over six recording sessions, three in which all 5 stimulus classes were used, and three in which only classes 1,2, and 4 were used (errorbars denote standard deviations). Classes of bursts with lower conditional entropy form more homogeneous clusters. Artificial stimuli that consist of temporally more structured and/or longer stimuli (classes 2,4, and 5) are better separable from noise or other stimuli than single pulse stimuli (classes 1 and 3). B) Confusion matrix for assigned cluster labels vs. actual labels of bursts. In every row we plot the average relative frequencies of burst labels occurrences in clusters that were assigned to one of the classes N ( = Noise) or 1–5. One can see that most mistakes are due to noise bursts assigned to one of the artificial stimulus classes. Also bursts in response to classes 2 and 3 are sometimes clustered together.
Mentions: Figure 8A shows the average conditional entropy individually for every class, where the average is over all recordings sessions in which those stimuli were used (three sessions for all 5 stimuli, three sessions only for stimulus classes 1,2, and 4). Although these statistics are based on only six recording sessions, and the standard deviations are large, one can observe the same trend that was qualitatively visible from Figure 7. The average conditional entropy is low for classes of bursts in response to long and/or temporally structured stimuli (classes 2,4, and 5), and higher for the single short pulses of classes 1 and 3. This indicates that classes 2,4, and 5 can be better discriminated from other artificial or background signals than the single pulse stimuli. Due to the limited amount of available data, these results are statistically not significant, and more measurements would be required.

Bottom Line: Our results show that the reliability of burst coding in the time domain is so high that identical stimuli lead to extremely similar spike pattern responses, even for different preparations on different dates, and even if one of the preparations is recorded outdoors and the other one in the sound proof lab.Our study shows that burst coding can provide a reliable mechanism for acoustic insects to classify and discriminate signals under very noisy real-world conditions.This gives new insights into the neural mechanisms potentially used by bushcrickets to discriminate conspecific songs from sounds of predators in similar carrier frequency bands.

View Article: PubMed Central - PubMed

Affiliation: Institute for Theoretical Computer Science, TU Graz, Graz, Austria. pfeiffer@ini.phys.ethz.ch

ABSTRACT
The task of an organism to extract information about the external environment from sensory signals is based entirely on the analysis of ongoing afferent spike activity provided by the sense organs. We investigate the processing of auditory stimuli by an acoustic interneuron of insects. In contrast to most previous work we do this by using stimuli and neurophysiological recordings directly in the nocturnal tropical rainforest, where the insect communicates. Different from typical recordings in sound proof laboratories, strong environmental noise from multiple sound sources interferes with the perception of acoustic signals in these realistic scenarios. We apply a recently developed unsupervised machine learning algorithm based on probabilistic inference to find frequently occurring firing patterns in the response of the acoustic interneuron. We can thus ask how much information the central nervous system of the receiver can extract from bursts without ever being told which type and which variants of bursts are characteristic for particular stimuli. Our results show that the reliability of burst coding in the time domain is so high that identical stimuli lead to extremely similar spike pattern responses, even for different preparations on different dates, and even if one of the preparations is recorded outdoors and the other one in the sound proof lab. Simultaneous recordings in two preparations exposed to the same acoustic environment reveal that characteristics of burst patterns are largely preserved among individuals of the same species. Our study shows that burst coding can provide a reliable mechanism for acoustic insects to classify and discriminate signals under very noisy real-world conditions. This gives new insights into the neural mechanisms potentially used by bushcrickets to discriminate conspecific songs from sounds of predators in similar carrier frequency bands.

Show MeSH
Related in: MedlinePlus