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

The network model accounts for the influence of the biasing sequence on tritone perception. (A) A randomly drawn sequence of 10 Shepard tones precedes an ambiguous pair (at 4 and 10 st). This bias sequence is restricted to lie between the ambiguous pair. Tone durations are 100 ms and inter-tone pause is 50 ms. The gap between the biasing sequence and the tritone pair is 0.5 s. (B) The firing rate difference of Eup and Edown populations (rup(x, t)-rdown(x, t), see Materials and Methods) for the entire sequence shows the local response to each tone. Eup has a larger response to the final tone, T2, indicating an ascending percept (box, consistent with human perception). (C) The influence of the bias sequence is reflected in the accumulation of the facilitation level F in the biased region. (D) Snapshot of the network activity at 30 ms after the onset of T2 (PC = 10 st). Facilitation level (magenta) has built up in the biasing region, below the pitch class of T2. The firing rate profile for Eup (blue thick) has a higher peak than for Edown (green thick) showing that Eup is winning the competition for the model's perceptual choice. Inhibitory input to the Eup (blue thin) and the Edown (green thin) units spread to the higher frequency side and the lower frequency side, respectively. The Edown unit receives higher inhibition than the Eup unit at PC = 10 st (black vertical line) due to facilitation of the I units below T2. (E) Time courses of the Eup (blue) and Edown (green) units at the pitch class of T2 during T2 presentation. (E1), Inhibitory inputs to the Eup and Edown units; (E2), firing rates of the Eup, Edown, and I (red) units. (F) Tuning curves of Eup and Edown units (at PC = 10 st) are affected differentially by biasing. The tuning curve of the Edown (solid green) unit reduces more than the Eup (solid blue) unit after biasing from below. The tuning curves of Eup (dashed blue) and Edown (dashed green) units without biasing are the same as the solid curves in Figure 4B. The biasing sequence is the same as in (A); the tuning curves are measured after the biasing sequence and the gap (0.5 s).
© Copyright Policy
Related In: Results  -  Collection

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

Figure 5: The network model accounts for the influence of the biasing sequence on tritone perception. (A) A randomly drawn sequence of 10 Shepard tones precedes an ambiguous pair (at 4 and 10 st). This bias sequence is restricted to lie between the ambiguous pair. Tone durations are 100 ms and inter-tone pause is 50 ms. The gap between the biasing sequence and the tritone pair is 0.5 s. (B) The firing rate difference of Eup and Edown populations (rup(x, t)-rdown(x, t), see Materials and Methods) for the entire sequence shows the local response to each tone. Eup has a larger response to the final tone, T2, indicating an ascending percept (box, consistent with human perception). (C) The influence of the bias sequence is reflected in the accumulation of the facilitation level F in the biased region. (D) Snapshot of the network activity at 30 ms after the onset of T2 (PC = 10 st). Facilitation level (magenta) has built up in the biasing region, below the pitch class of T2. The firing rate profile for Eup (blue thick) has a higher peak than for Edown (green thick) showing that Eup is winning the competition for the model's perceptual choice. Inhibitory input to the Eup (blue thin) and the Edown (green thin) units spread to the higher frequency side and the lower frequency side, respectively. The Edown unit receives higher inhibition than the Eup unit at PC = 10 st (black vertical line) due to facilitation of the I units below T2. (E) Time courses of the Eup (blue) and Edown (green) units at the pitch class of T2 during T2 presentation. (E1), Inhibitory inputs to the Eup and Edown units; (E2), firing rates of the Eup, Edown, and I (red) units. (F) Tuning curves of Eup and Edown units (at PC = 10 st) are affected differentially by biasing. The tuning curve of the Edown (solid green) unit reduces more than the Eup (solid blue) unit after biasing from below. The tuning curves of Eup (dashed blue) and Edown (dashed green) units without biasing are the same as the solid curves in Figure 4B. The biasing sequence is the same as in (A); the tuning curves are measured after the biasing sequence and the gap (0.5 s).

Mentions: Slow facilitation of inhibitory synapses integrates spectral information of stimulus history in the model. This slow adaptation thereby biases the model's pitch-change-direction percept of the tritone pair that would be ambiguous if tested alone. During a preceding sequence of Shepard tones, Eup and Edown respond to each tone locally with different activity levels indicating percepts of pitch-change direction. Inhibitory synapses gradually facilitate wherever inhibitory neurons are activated (Equation 2), representing a spectral distribution of recent stimulus history (Figure 5C). The facilitation level decays slowly during the silent gap between the preceding sequence and the tritone pair. The facilitated inhibitory synapses disadvantage Edown during the T2 presentation after a sequence of Shepard tones below T2, resulting in a larger population response difference (box in Figure 5B, red area larger than blue area). This imbalance leads to an ascending percept in the model for the tritone comparison. Population firing rates of Eup (Figure 5D, thick blue) and Edown (Figure 5D, thick green) start to separate at 30 ms after the onset of T2. Inhibitory current on Eup (Figure 5D, thin blue) comes from the higher frequency side and spreads to the lower side, pushing the population peak of Eup above the PC of T2. Eup continues recruiting more units at higher CF's by recurrent excitation while Edown is suppressed due to the facilitated inhibition from lower CF units. Hence, the model predicts an ascending percept for a tritone pair after a preceding sequence of tones within +6 st from T1. This context dependence of the model is consistent with psychophysical results (Repp, 1997; Englitz et al., 2013).


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

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

The network model accounts for the influence of the biasing sequence on tritone perception. (A) A randomly drawn sequence of 10 Shepard tones precedes an ambiguous pair (at 4 and 10 st). This bias sequence is restricted to lie between the ambiguous pair. Tone durations are 100 ms and inter-tone pause is 50 ms. The gap between the biasing sequence and the tritone pair is 0.5 s. (B) The firing rate difference of Eup and Edown populations (rup(x, t)-rdown(x, t), see Materials and Methods) for the entire sequence shows the local response to each tone. Eup has a larger response to the final tone, T2, indicating an ascending percept (box, consistent with human perception). (C) The influence of the bias sequence is reflected in the accumulation of the facilitation level F in the biased region. (D) Snapshot of the network activity at 30 ms after the onset of T2 (PC = 10 st). Facilitation level (magenta) has built up in the biasing region, below the pitch class of T2. The firing rate profile for Eup (blue thick) has a higher peak than for Edown (green thick) showing that Eup is winning the competition for the model's perceptual choice. Inhibitory input to the Eup (blue thin) and the Edown (green thin) units spread to the higher frequency side and the lower frequency side, respectively. The Edown unit receives higher inhibition than the Eup unit at PC = 10 st (black vertical line) due to facilitation of the I units below T2. (E) Time courses of the Eup (blue) and Edown (green) units at the pitch class of T2 during T2 presentation. (E1), Inhibitory inputs to the Eup and Edown units; (E2), firing rates of the Eup, Edown, and I (red) units. (F) Tuning curves of Eup and Edown units (at PC = 10 st) are affected differentially by biasing. The tuning curve of the Edown (solid green) unit reduces more than the Eup (solid blue) unit after biasing from below. The tuning curves of Eup (dashed blue) and Edown (dashed green) units without biasing are the same as the solid curves in Figure 4B. The biasing sequence is the same as in (A); the tuning curves are measured after the biasing sequence and the gap (0.5 s).
© Copyright Policy
Related In: Results  -  Collection

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

Figure 5: The network model accounts for the influence of the biasing sequence on tritone perception. (A) A randomly drawn sequence of 10 Shepard tones precedes an ambiguous pair (at 4 and 10 st). This bias sequence is restricted to lie between the ambiguous pair. Tone durations are 100 ms and inter-tone pause is 50 ms. The gap between the biasing sequence and the tritone pair is 0.5 s. (B) The firing rate difference of Eup and Edown populations (rup(x, t)-rdown(x, t), see Materials and Methods) for the entire sequence shows the local response to each tone. Eup has a larger response to the final tone, T2, indicating an ascending percept (box, consistent with human perception). (C) The influence of the bias sequence is reflected in the accumulation of the facilitation level F in the biased region. (D) Snapshot of the network activity at 30 ms after the onset of T2 (PC = 10 st). Facilitation level (magenta) has built up in the biasing region, below the pitch class of T2. The firing rate profile for Eup (blue thick) has a higher peak than for Edown (green thick) showing that Eup is winning the competition for the model's perceptual choice. Inhibitory input to the Eup (blue thin) and the Edown (green thin) units spread to the higher frequency side and the lower frequency side, respectively. The Edown unit receives higher inhibition than the Eup unit at PC = 10 st (black vertical line) due to facilitation of the I units below T2. (E) Time courses of the Eup (blue) and Edown (green) units at the pitch class of T2 during T2 presentation. (E1), Inhibitory inputs to the Eup and Edown units; (E2), firing rates of the Eup, Edown, and I (red) units. (F) Tuning curves of Eup and Edown units (at PC = 10 st) are affected differentially by biasing. The tuning curve of the Edown (solid green) unit reduces more than the Eup (solid blue) unit after biasing from below. The tuning curves of Eup (dashed blue) and Edown (dashed green) units without biasing are the same as the solid curves in Figure 4B. The biasing sequence is the same as in (A); the tuning curves are measured after the biasing sequence and the gap (0.5 s).
Mentions: Slow facilitation of inhibitory synapses integrates spectral information of stimulus history in the model. This slow adaptation thereby biases the model's pitch-change-direction percept of the tritone pair that would be ambiguous if tested alone. During a preceding sequence of Shepard tones, Eup and Edown respond to each tone locally with different activity levels indicating percepts of pitch-change direction. Inhibitory synapses gradually facilitate wherever inhibitory neurons are activated (Equation 2), representing a spectral distribution of recent stimulus history (Figure 5C). The facilitation level decays slowly during the silent gap between the preceding sequence and the tritone pair. The facilitated inhibitory synapses disadvantage Edown during the T2 presentation after a sequence of Shepard tones below T2, resulting in a larger population response difference (box in Figure 5B, red area larger than blue area). This imbalance leads to an ascending percept in the model for the tritone comparison. Population firing rates of Eup (Figure 5D, thick blue) and Edown (Figure 5D, thick green) start to separate at 30 ms after the onset of T2. Inhibitory current on Eup (Figure 5D, thin blue) comes from the higher frequency side and spreads to the lower side, pushing the population peak of Eup above the PC of T2. Eup continues recruiting more units at higher CF's by recurrent excitation while Edown is suppressed due to the facilitated inhibition from lower CF units. Hence, the model predicts an ascending percept for a tritone pair after a preceding sequence of tones within +6 st from T1. This context dependence of the model is consistent with psychophysical results (Repp, 1997; Englitz et al., 2013).

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