Limits...
Efficient temporal processing of naturalistic sounds.

Lesica NA, Grothe B - PLoS ONE (2008)

Bottom Line: We find that the onset of ambient noise evokes a change in receptive field dynamics that corresponds to a change from bandpass to lowpass temporal filtering.We show that these changes occur within a few hundred milliseconds of the onset of the noise and are evident across a range of overall stimulus intensities.Using a simple model, we illustrate how these changes in temporal processing exploit differences in the statistical properties of vocalizations and ambient noises to increase the information in the neural response in a manner consistent with the principles of efficient coding.

View Article: PubMed Central - PubMed

Affiliation: Department of Biology II, Ludwig-Maximilians-University Munich, Martinsried, Germany. lesica@zi.biologie.uni-muenchen.de

ABSTRACT
In this study, we investigate the ability of the mammalian auditory pathway to adapt its strategy for temporal processing under natural stimulus conditions. We derive temporal receptive fields from the responses of neurons in the inferior colliculus to vocalization stimuli with and without additional ambient noise. We find that the onset of ambient noise evokes a change in receptive field dynamics that corresponds to a change from bandpass to lowpass temporal filtering. We show that these changes occur within a few hundred milliseconds of the onset of the noise and are evident across a range of overall stimulus intensities. Using a simple model, we illustrate how these changes in temporal processing exploit differences in the statistical properties of vocalizations and ambient noises to increase the information in the neural response in a manner consistent with the principles of efficient coding.

Show MeSH
Temporal processing of vocalization stimuli with and without ambient noise.a) The temporal RFs estimated from responses to the V (black) and VN (red) stimuli for a typical cell. The preferred frequency of the cell was 12.7 KHz. The error bars represent 95% confidence bounds. The RFs were normalized to have the same peak value for plotting. b) The MTFs corresponding to the RFs in a. MTFs were obtained by computing the power spectrum of the RF. Before computing the power spectrum, RFs were normalized such that the variance of the result of the convolution of the RF with the vocalization signal was one. c,d) RFs and MTFs for two additional cells with preferred frequencies of 2.5 and 12.5 KHz, respectively. RFs and MTFs were normalized as described in a and b. e) The area above the MTF corresponding to the attenuation of low frequencies was measured on logarithmic axes, from the lowest non-zero frequency measured (20 Hz) to the peak. Before calculation of the LF area, the MTF was normalized to have a peak value of one. The LF area was zero for those neurons whose MTFs were monotonically decreasing. f) The LF area of the V and VN MTFs for the population of 63 cells. The crosses denote the population mean±one standard deviation. The colored circles correspond the RFs and MTFs shown in a–d as indicated by the colored boxes. g) A histogram showing the percent changes between the LF areas in f.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2249929&req=5

pone-0001655-g003: Temporal processing of vocalization stimuli with and without ambient noise.a) The temporal RFs estimated from responses to the V (black) and VN (red) stimuli for a typical cell. The preferred frequency of the cell was 12.7 KHz. The error bars represent 95% confidence bounds. The RFs were normalized to have the same peak value for plotting. b) The MTFs corresponding to the RFs in a. MTFs were obtained by computing the power spectrum of the RF. Before computing the power spectrum, RFs were normalized such that the variance of the result of the convolution of the RF with the vocalization signal was one. c,d) RFs and MTFs for two additional cells with preferred frequencies of 2.5 and 12.5 KHz, respectively. RFs and MTFs were normalized as described in a and b. e) The area above the MTF corresponding to the attenuation of low frequencies was measured on logarithmic axes, from the lowest non-zero frequency measured (20 Hz) to the peak. Before calculation of the LF area, the MTF was normalized to have a peak value of one. The LF area was zero for those neurons whose MTFs were monotonically decreasing. f) The LF area of the V and VN MTFs for the population of 63 cells. The crosses denote the population mean±one standard deviation. The colored circles correspond the RFs and MTFs shown in a–d as indicated by the colored boxes. g) A histogram showing the percent changes between the LF areas in f.

Mentions: The temporal RFs of a typical cell during stimulation with the V and VN stimuli in figure 3a. The V RF (black) exhibits fast, biphasic dynamics when coding the vocalization stimulus alone, while the addition of noise evokes a change to the slower, more monophasic dynamics of the VN RF (red). As shown in figure 3b, these changes are also evident in the frequency domain representation of the temporal RF, the modulation tuning function (MTF), which reflects the response of the neuron to different modulation frequencies: the V RF has bandpass modulation tuning (black), while the tuning of the VN RF is lowpass (red). The effect of the addition of noise to the vocalization stimulus on temporal processing varied widely, with the MTFs of many neurons changing from bandpass to pure lowpass (figures 3a, b, and c), while the MTFs of other neurons did not change at all (figure 3d). We quantified these effects by measuring the area above the MTF (on logarithmic axes) that corresponds to the attenuation of low frequencies, as shown in figure 3e. Across the population, the LF areas of the VN MTFs were significantly smaller than those of the V MTFs, as shown in figure 3f (paired t-test, p<0.001, n = 63). The average decrease in MTF LF area was 60% (see histogram in figure 3g), indicating that, as a population, IC neurons exhibit a strong shift from bandpass modulation tuning under noise-free conditions to lowpass modulation tuning with the addition of noise.


Efficient temporal processing of naturalistic sounds.

Lesica NA, Grothe B - PLoS ONE (2008)

Temporal processing of vocalization stimuli with and without ambient noise.a) The temporal RFs estimated from responses to the V (black) and VN (red) stimuli for a typical cell. The preferred frequency of the cell was 12.7 KHz. The error bars represent 95% confidence bounds. The RFs were normalized to have the same peak value for plotting. b) The MTFs corresponding to the RFs in a. MTFs were obtained by computing the power spectrum of the RF. Before computing the power spectrum, RFs were normalized such that the variance of the result of the convolution of the RF with the vocalization signal was one. c,d) RFs and MTFs for two additional cells with preferred frequencies of 2.5 and 12.5 KHz, respectively. RFs and MTFs were normalized as described in a and b. e) The area above the MTF corresponding to the attenuation of low frequencies was measured on logarithmic axes, from the lowest non-zero frequency measured (20 Hz) to the peak. Before calculation of the LF area, the MTF was normalized to have a peak value of one. The LF area was zero for those neurons whose MTFs were monotonically decreasing. f) The LF area of the V and VN MTFs for the population of 63 cells. The crosses denote the population mean±one standard deviation. The colored circles correspond the RFs and MTFs shown in a–d as indicated by the colored boxes. g) A histogram showing the percent changes between the LF areas in f.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2249929&req=5

pone-0001655-g003: Temporal processing of vocalization stimuli with and without ambient noise.a) The temporal RFs estimated from responses to the V (black) and VN (red) stimuli for a typical cell. The preferred frequency of the cell was 12.7 KHz. The error bars represent 95% confidence bounds. The RFs were normalized to have the same peak value for plotting. b) The MTFs corresponding to the RFs in a. MTFs were obtained by computing the power spectrum of the RF. Before computing the power spectrum, RFs were normalized such that the variance of the result of the convolution of the RF with the vocalization signal was one. c,d) RFs and MTFs for two additional cells with preferred frequencies of 2.5 and 12.5 KHz, respectively. RFs and MTFs were normalized as described in a and b. e) The area above the MTF corresponding to the attenuation of low frequencies was measured on logarithmic axes, from the lowest non-zero frequency measured (20 Hz) to the peak. Before calculation of the LF area, the MTF was normalized to have a peak value of one. The LF area was zero for those neurons whose MTFs were monotonically decreasing. f) The LF area of the V and VN MTFs for the population of 63 cells. The crosses denote the population mean±one standard deviation. The colored circles correspond the RFs and MTFs shown in a–d as indicated by the colored boxes. g) A histogram showing the percent changes between the LF areas in f.
Mentions: The temporal RFs of a typical cell during stimulation with the V and VN stimuli in figure 3a. The V RF (black) exhibits fast, biphasic dynamics when coding the vocalization stimulus alone, while the addition of noise evokes a change to the slower, more monophasic dynamics of the VN RF (red). As shown in figure 3b, these changes are also evident in the frequency domain representation of the temporal RF, the modulation tuning function (MTF), which reflects the response of the neuron to different modulation frequencies: the V RF has bandpass modulation tuning (black), while the tuning of the VN RF is lowpass (red). The effect of the addition of noise to the vocalization stimulus on temporal processing varied widely, with the MTFs of many neurons changing from bandpass to pure lowpass (figures 3a, b, and c), while the MTFs of other neurons did not change at all (figure 3d). We quantified these effects by measuring the area above the MTF (on logarithmic axes) that corresponds to the attenuation of low frequencies, as shown in figure 3e. Across the population, the LF areas of the VN MTFs were significantly smaller than those of the V MTFs, as shown in figure 3f (paired t-test, p<0.001, n = 63). The average decrease in MTF LF area was 60% (see histogram in figure 3g), indicating that, as a population, IC neurons exhibit a strong shift from bandpass modulation tuning under noise-free conditions to lowpass modulation tuning with the addition of noise.

Bottom Line: We find that the onset of ambient noise evokes a change in receptive field dynamics that corresponds to a change from bandpass to lowpass temporal filtering.We show that these changes occur within a few hundred milliseconds of the onset of the noise and are evident across a range of overall stimulus intensities.Using a simple model, we illustrate how these changes in temporal processing exploit differences in the statistical properties of vocalizations and ambient noises to increase the information in the neural response in a manner consistent with the principles of efficient coding.

View Article: PubMed Central - PubMed

Affiliation: Department of Biology II, Ludwig-Maximilians-University Munich, Martinsried, Germany. lesica@zi.biologie.uni-muenchen.de

ABSTRACT
In this study, we investigate the ability of the mammalian auditory pathway to adapt its strategy for temporal processing under natural stimulus conditions. We derive temporal receptive fields from the responses of neurons in the inferior colliculus to vocalization stimuli with and without additional ambient noise. We find that the onset of ambient noise evokes a change in receptive field dynamics that corresponds to a change from bandpass to lowpass temporal filtering. We show that these changes occur within a few hundred milliseconds of the onset of the noise and are evident across a range of overall stimulus intensities. Using a simple model, we illustrate how these changes in temporal processing exploit differences in the statistical properties of vocalizations and ambient noises to increase the information in the neural response in a manner consistent with the principles of efficient coding.

Show MeSH