Limits...
A neuronal network model for context-dependence of pitch change perception.

Huang C, Englitz B, Shamma S, Rinzel J - Front Comput Neurosci (2015)

Bottom Line: We developed a recurrent, firing-rate network model, which detects frequency-change-direction of successively played stimuli and successfully accounts for the context-dependent perception demonstrated in behavioral experiments.The model's network architecture and slow facilitating inhibition emerge as predictions of neuronal mechanisms for these perceptual dynamics.Since the model structure does not depend on the specific stimuli, we show that it generalizes to other contextual effects and stimulus types.

View Article: PubMed Central - PubMed

Affiliation: Courant Institute of Mathematical Sciences, New York University New York, NY, USA.

ABSTRACT
Many natural stimuli have perceptual ambiguities that can be cognitively resolved by the surrounding context. In audition, preceding context can bias the perception of speech and non-speech stimuli. Here, we develop a neuronal network model that can account for how context affects the perception of pitch change between a pair of successive complex tones. We focus especially on an ambiguous comparison-listeners experience opposite percepts (either ascending or descending) for an ambiguous tone pair depending on the spectral location of preceding context tones. We developed a recurrent, firing-rate network model, which detects frequency-change-direction of successively played stimuli and successfully accounts for the context-dependent perception demonstrated in behavioral experiments. The model consists of two tonotopically organized, excitatory populations, E up and E down, that respond preferentially to ascending or descending stimuli in pitch, respectively. These preferences are generated by an inhibitory population that provides inhibition asymmetric in frequency to the two populations; context dependence arises from slow facilitation of inhibition. We show that contextual influence depends on the spectral distribution of preceding tones and the tuning width of inhibitory neurons. Further, we demonstrate, using phase-space analysis, how the facilitated inhibition from previous stimuli and the waning inhibition from the just-preceding tone shape the competition between the E up and E down populations. In sum, our model accounts for contextual influences on the pitch change perception of an ambiguous tone pair by introducing a novel decoding strategy based on direction-selective units. The model's network architecture and slow facilitating inhibition emerge as predictions of neuronal mechanisms for these perceptual dynamics. Since the model structure does not depend on the specific stimuli, we show that it generalizes to other contextual effects and stimulus types.

No MeSH data available.


Related in: MedlinePlus

Biasing effects depend on the spectral distribution of bias tones and tuning width of I units. (A) Mean relative response difference, D (Equation 6, see Materials and Methods), of Eup and Edown for T2 vs. PC of a single bias tone (abscissa, different locations) depends on the tuning width of the inhibitory units (narrow tuning = blue, broad tuning = green). The ambiguous Shepard tone pair is for T1 = 0 st, T2 = 6 st. The footprints of E to E (σee) and E to I (σei) are 2.5 times wider for broad tuning of I units, and the synaptic strength of recurrent excitation (aee) is increased to have comparable firing rates. Parameter values for narrow tuning are σee = 0.02, σei = 0.08 octaves, and aee = 0.7, and those for broad tuning are σee = 0.05, σei = 0.2 octaves, and aee = 1.5. Other parameters are the same as used in Materials and Methods. Narrow tuning is used in other figures. (B) The biasing effect accumulates with the number of bias tones. The buildup depends more steeply on Nbias for broad tuning of I units (green) than for narrow tuning (blue). A faster decay time constant of facilitation τfd leads to lower biasing effects, but does not strongly affect the buildup “rate” (solid: τfd = 2 s; dashed: τfd = 1 s). The percentage of ascending responses, P(up), over trials (each trial is for a sequence of random Shepard tones) is plotted vs. the number of biasing tones Nbias. An “ascending choice” is made if D > 0.1; a threshold value, 0.1, is used for all conditions. The Nbias Shepard tones for a sequence are randomly sampled for ascending bias in the region above T1 and below T2 and for the tritone pair as in (A); there were 400 trials for each Nbias (error bars denote 2 SEM).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4526807&req=5

Figure 6: Biasing effects depend on the spectral distribution of bias tones and tuning width of I units. (A) Mean relative response difference, D (Equation 6, see Materials and Methods), of Eup and Edown for T2 vs. PC of a single bias tone (abscissa, different locations) depends on the tuning width of the inhibitory units (narrow tuning = blue, broad tuning = green). The ambiguous Shepard tone pair is for T1 = 0 st, T2 = 6 st. The footprints of E to E (σee) and E to I (σei) are 2.5 times wider for broad tuning of I units, and the synaptic strength of recurrent excitation (aee) is increased to have comparable firing rates. Parameter values for narrow tuning are σee = 0.02, σei = 0.08 octaves, and aee = 0.7, and those for broad tuning are σee = 0.05, σei = 0.2 octaves, and aee = 1.5. Other parameters are the same as used in Materials and Methods. Narrow tuning is used in other figures. (B) The biasing effect accumulates with the number of bias tones. The buildup depends more steeply on Nbias for broad tuning of I units (green) than for narrow tuning (blue). A faster decay time constant of facilitation τfd leads to lower biasing effects, but does not strongly affect the buildup “rate” (solid: τfd = 2 s; dashed: τfd = 1 s). The percentage of ascending responses, P(up), over trials (each trial is for a sequence of random Shepard tones) is plotted vs. the number of biasing tones Nbias. An “ascending choice” is made if D > 0.1; a threshold value, 0.1, is used for all conditions. The Nbias Shepard tones for a sequence are randomly sampled for ascending bias in the region above T1 and below T2 and for the tritone pair as in (A); there were 400 trials for each Nbias (error bars denote 2 SEM).

Mentions: With a single Shepard tone as context that precedes a tritone pair, the impact of biasing depends on the PC of the bias tone, B, and on the tuning width of I units. If the tuning width is narrow (about 3 st for our default parameter settings, not shown explicitly), biasing is most effective when it occurs about 1 st from T2 (Figure 6A, blue). If the tuning of an I unit is broad (say, about 6 st), the most effective bias tone is shifted to midway between T1 and T2 (Figure 6A, green). The response difference of Eup and Edown depends on the facilitation level difference from above and below T2. On the one hand, B needs to be close enough to T2 so that the I units activated by B partially overlap those activated by T2; the biasing effect depends on accumulated facilitation level, more on one side than the other, so that inhibition affects Eup and Edown units differentially. On the other hand, when B is too close to T2, the facilitation level is maximal but flat around the PC of T2, showing little difference between the two sides of T2. Therefore, the dependence of the tritone comparison on the PC of B scales with the tuning widths of inhibitory units.


A neuronal network model for context-dependence of pitch change perception.

Huang C, Englitz B, Shamma S, Rinzel J - Front Comput Neurosci (2015)

Biasing effects depend on the spectral distribution of bias tones and tuning width of I units. (A) Mean relative response difference, D (Equation 6, see Materials and Methods), of Eup and Edown for T2 vs. PC of a single bias tone (abscissa, different locations) depends on the tuning width of the inhibitory units (narrow tuning = blue, broad tuning = green). The ambiguous Shepard tone pair is for T1 = 0 st, T2 = 6 st. The footprints of E to E (σee) and E to I (σei) are 2.5 times wider for broad tuning of I units, and the synaptic strength of recurrent excitation (aee) is increased to have comparable firing rates. Parameter values for narrow tuning are σee = 0.02, σei = 0.08 octaves, and aee = 0.7, and those for broad tuning are σee = 0.05, σei = 0.2 octaves, and aee = 1.5. Other parameters are the same as used in Materials and Methods. Narrow tuning is used in other figures. (B) The biasing effect accumulates with the number of bias tones. The buildup depends more steeply on Nbias for broad tuning of I units (green) than for narrow tuning (blue). A faster decay time constant of facilitation τfd leads to lower biasing effects, but does not strongly affect the buildup “rate” (solid: τfd = 2 s; dashed: τfd = 1 s). The percentage of ascending responses, P(up), over trials (each trial is for a sequence of random Shepard tones) is plotted vs. the number of biasing tones Nbias. An “ascending choice” is made if D > 0.1; a threshold value, 0.1, is used for all conditions. The Nbias Shepard tones for a sequence are randomly sampled for ascending bias in the region above T1 and below T2 and for the tritone pair as in (A); there were 400 trials for each Nbias (error bars denote 2 SEM).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4526807&req=5

Figure 6: Biasing effects depend on the spectral distribution of bias tones and tuning width of I units. (A) Mean relative response difference, D (Equation 6, see Materials and Methods), of Eup and Edown for T2 vs. PC of a single bias tone (abscissa, different locations) depends on the tuning width of the inhibitory units (narrow tuning = blue, broad tuning = green). The ambiguous Shepard tone pair is for T1 = 0 st, T2 = 6 st. The footprints of E to E (σee) and E to I (σei) are 2.5 times wider for broad tuning of I units, and the synaptic strength of recurrent excitation (aee) is increased to have comparable firing rates. Parameter values for narrow tuning are σee = 0.02, σei = 0.08 octaves, and aee = 0.7, and those for broad tuning are σee = 0.05, σei = 0.2 octaves, and aee = 1.5. Other parameters are the same as used in Materials and Methods. Narrow tuning is used in other figures. (B) The biasing effect accumulates with the number of bias tones. The buildup depends more steeply on Nbias for broad tuning of I units (green) than for narrow tuning (blue). A faster decay time constant of facilitation τfd leads to lower biasing effects, but does not strongly affect the buildup “rate” (solid: τfd = 2 s; dashed: τfd = 1 s). The percentage of ascending responses, P(up), over trials (each trial is for a sequence of random Shepard tones) is plotted vs. the number of biasing tones Nbias. An “ascending choice” is made if D > 0.1; a threshold value, 0.1, is used for all conditions. The Nbias Shepard tones for a sequence are randomly sampled for ascending bias in the region above T1 and below T2 and for the tritone pair as in (A); there were 400 trials for each Nbias (error bars denote 2 SEM).
Mentions: With a single Shepard tone as context that precedes a tritone pair, the impact of biasing depends on the PC of the bias tone, B, and on the tuning width of I units. If the tuning width is narrow (about 3 st for our default parameter settings, not shown explicitly), biasing is most effective when it occurs about 1 st from T2 (Figure 6A, blue). If the tuning of an I unit is broad (say, about 6 st), the most effective bias tone is shifted to midway between T1 and T2 (Figure 6A, green). The response difference of Eup and Edown depends on the facilitation level difference from above and below T2. On the one hand, B needs to be close enough to T2 so that the I units activated by B partially overlap those activated by T2; the biasing effect depends on accumulated facilitation level, more on one side than the other, so that inhibition affects Eup and Edown units differentially. On the other hand, when B is too close to T2, the facilitation level is maximal but flat around the PC of T2, showing little difference between the two sides of T2. Therefore, the dependence of the tritone comparison on the PC of B scales with the tuning widths of inhibitory units.

Bottom Line: We developed a recurrent, firing-rate network model, which detects frequency-change-direction of successively played stimuli and successfully accounts for the context-dependent perception demonstrated in behavioral experiments.The model's network architecture and slow facilitating inhibition emerge as predictions of neuronal mechanisms for these perceptual dynamics.Since the model structure does not depend on the specific stimuli, we show that it generalizes to other contextual effects and stimulus types.

View Article: PubMed Central - PubMed

Affiliation: Courant Institute of Mathematical Sciences, New York University New York, NY, USA.

ABSTRACT
Many natural stimuli have perceptual ambiguities that can be cognitively resolved by the surrounding context. In audition, preceding context can bias the perception of speech and non-speech stimuli. Here, we develop a neuronal network model that can account for how context affects the perception of pitch change between a pair of successive complex tones. We focus especially on an ambiguous comparison-listeners experience opposite percepts (either ascending or descending) for an ambiguous tone pair depending on the spectral location of preceding context tones. We developed a recurrent, firing-rate network model, which detects frequency-change-direction of successively played stimuli and successfully accounts for the context-dependent perception demonstrated in behavioral experiments. The model consists of two tonotopically organized, excitatory populations, E up and E down, that respond preferentially to ascending or descending stimuli in pitch, respectively. These preferences are generated by an inhibitory population that provides inhibition asymmetric in frequency to the two populations; context dependence arises from slow facilitation of inhibition. We show that contextual influence depends on the spectral distribution of preceding tones and the tuning width of inhibitory neurons. Further, we demonstrate, using phase-space analysis, how the facilitated inhibition from previous stimuli and the waning inhibition from the just-preceding tone shape the competition between the E up and E down populations. In sum, our model accounts for contextual influences on the pitch change perception of an ambiguous tone pair by introducing a novel decoding strategy based on direction-selective units. The model's network architecture and slow facilitating inhibition emerge as predictions of neuronal mechanisms for these perceptual dynamics. Since the model structure does not depend on the specific stimuli, we show that it generalizes to other contextual effects and stimulus types.

No MeSH data available.


Related in: MedlinePlus