Limits...
The role of inhibition in a computational model of an auditory cortical neuron during the encoding of temporal information.

Bendor D - PLoS Comput. Biol. (2015)

Bottom Line: Using a computational neuronal model, we find that stimulus-locked responses are generated when sound-evoked excitation is combined with strong, delayed inhibition.In contrast to this, a non-synchronized rate representation is generated when the net excitation evoked by the sound is weak, which occurs when excitation is coincident and balanced with inhibition.Together these data suggest that feedforward inhibition provides a parsimonious explanation of the neural coding dichotomy observed in auditory cortex.

View Article: PubMed Central - PubMed

Affiliation: Institute of Behavioural Neuroscience, Department of Experimental Psychology, University College London, London, United Kingdom.

ABSTRACT
In auditory cortex, temporal information within a sound is represented by two complementary neural codes: a temporal representation based on stimulus-locked firing and a rate representation, where discharge rate co-varies with the timing between acoustic events but lacks a stimulus-synchronized response. Using a computational neuronal model, we find that stimulus-locked responses are generated when sound-evoked excitation is combined with strong, delayed inhibition. In contrast to this, a non-synchronized rate representation is generated when the net excitation evoked by the sound is weak, which occurs when excitation is coincident and balanced with inhibition. Using single-unit recordings from awake marmosets (Callithrix jacchus), we validate several model predictions, including differences in the temporal fidelity, discharge rates and temporal dynamics of stimulus-evoked responses between neurons with rate and temporal representations. Together these data suggest that feedforward inhibition provides a parsimonious explanation of the neural coding dichotomy observed in auditory cortex.

No MeSH data available.


Temporal fidelity of synchronized and mixed neurons.Only simulated neurons with an excitatory input strength between 3–6 nS were used in this analysis, such that synchronized and mixed neurons had a similar distribution of excitatory levels. a. Max vector strength distribution for acoustic pulse train responses in simulated neurons. Mean: sync = 0.93, mixed = 0.79, Wilcoxon rank sum test: P < 3.3 x 10-52. b. IPI synchronization limit distribution for acoustic pulse train responses in simulated neurons. Mean: sync = 10.2 ms, mixed = 7.7 ms, Wilcoxon rank sum test: P < 1.3 x 10-43. c. Max vector strength distribution for acoustic pulse train responses in real neurons. Mean: sync = 0.68, mixed = 0.60, Wilcoxon rank sum test: P = 0.16. d. IPI synchronization limit distribution for acoustic pulse train responses in real neurons. Mean: sync = 25.7 ms, mixed = 13.4 ms, Wilcoxon rank sum test: P < 0.02.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4400160&req=5

pcbi.1004197.g008: Temporal fidelity of synchronized and mixed neurons.Only simulated neurons with an excitatory input strength between 3–6 nS were used in this analysis, such that synchronized and mixed neurons had a similar distribution of excitatory levels. a. Max vector strength distribution for acoustic pulse train responses in simulated neurons. Mean: sync = 0.93, mixed = 0.79, Wilcoxon rank sum test: P < 3.3 x 10-52. b. IPI synchronization limit distribution for acoustic pulse train responses in simulated neurons. Mean: sync = 10.2 ms, mixed = 7.7 ms, Wilcoxon rank sum test: P < 1.3 x 10-43. c. Max vector strength distribution for acoustic pulse train responses in real neurons. Mean: sync = 0.68, mixed = 0.60, Wilcoxon rank sum test: P = 0.16. d. IPI synchronization limit distribution for acoustic pulse train responses in real neurons. Mean: sync = 25.7 ms, mixed = 13.4 ms, Wilcoxon rank sum test: P < 0.02.

Mentions: We also hypothesized that the strength of a neuron’s inhibitory input would impact its temporal fidelity. Delayed inhibition suppresses random spiking that can occur between acoustic pulses; the resulting stimulus-locked response has a higher vector strength, thus providing a better temporal representation of the acoustic pulse train. As synchronized neurons have stronger delayed inhibition than mixed response neurons, synchronized neurons should therefore have higher vector strengths. However, this comes at a cost, as delayed inhibition prevents the response to a second acoustic pulse during the brief period that the neuron is suppressed. Thus mixed response neurons (which have less inhibition) should be able to stimulus synchronize at shorter IPIs than synchronized neurons. As predicted, we observed that for our simulated neuronal population, synchronized neurons had higher maximum vector strengths than mixed response neurons (Fig. 8a, sync = 0.93, mixed = 0.79; Wilcoxon rank sum test, P< 3.3 x 10-52, see Methods), while mixed response neurons had lower stimulus synchronization limits (Fig. 8b, sync = 10.2 ms, mixed = 7.7 ms; Wilcoxon rank sum test, P< 1.3 x 10-43). We observed a similar trend for our real neuron population- synchronized neurons had higher maximum vector strengths than mixed response neurons (Fig. 8c, sync = 0.68, mixed = 0.60; Wilcoxon rank sum test, P = 0.16), while mixed response neurons had lower stimulus synchronization limits (Fig. 8d, sync = 25.7 ms, mixed = 13.4 ms; Wilcoxon rank sum test, P< 0.02). Although we only observed a statistically significant difference between synchronized and mixed response neurons for the stimulus synchronization limit and not maximum vector strength, this may be due to the limited number of mixed response neurons that we were able to record from (n = 14). These results suggest that blocking inhibition (e.g. by adding a GABA-A antagonist such as Gabazine [28]), effectively decreasing the I/E ratio, should decrease a neuron’s vector strength while increasing its stimulus synchronization limit. However, the relationship between inhibition and temporal fidelity is more complex. While for simulated neurons with the same excitatory input strength, the stimulus synchronization limit of synchronized neurons decreased as the I/E ratio increased (P<0.001, Spearman correlation coefficient), we did not observe a statistically significant trend between the stimulus synchronization limit and I/E ratio in mixed response neurons (P>0.05, Spearman correlation coefficient).


The role of inhibition in a computational model of an auditory cortical neuron during the encoding of temporal information.

Bendor D - PLoS Comput. Biol. (2015)

Temporal fidelity of synchronized and mixed neurons.Only simulated neurons with an excitatory input strength between 3–6 nS were used in this analysis, such that synchronized and mixed neurons had a similar distribution of excitatory levels. a. Max vector strength distribution for acoustic pulse train responses in simulated neurons. Mean: sync = 0.93, mixed = 0.79, Wilcoxon rank sum test: P < 3.3 x 10-52. b. IPI synchronization limit distribution for acoustic pulse train responses in simulated neurons. Mean: sync = 10.2 ms, mixed = 7.7 ms, Wilcoxon rank sum test: P < 1.3 x 10-43. c. Max vector strength distribution for acoustic pulse train responses in real neurons. Mean: sync = 0.68, mixed = 0.60, Wilcoxon rank sum test: P = 0.16. d. IPI synchronization limit distribution for acoustic pulse train responses in real neurons. Mean: sync = 25.7 ms, mixed = 13.4 ms, Wilcoxon rank sum test: P < 0.02.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4400160&req=5

pcbi.1004197.g008: Temporal fidelity of synchronized and mixed neurons.Only simulated neurons with an excitatory input strength between 3–6 nS were used in this analysis, such that synchronized and mixed neurons had a similar distribution of excitatory levels. a. Max vector strength distribution for acoustic pulse train responses in simulated neurons. Mean: sync = 0.93, mixed = 0.79, Wilcoxon rank sum test: P < 3.3 x 10-52. b. IPI synchronization limit distribution for acoustic pulse train responses in simulated neurons. Mean: sync = 10.2 ms, mixed = 7.7 ms, Wilcoxon rank sum test: P < 1.3 x 10-43. c. Max vector strength distribution for acoustic pulse train responses in real neurons. Mean: sync = 0.68, mixed = 0.60, Wilcoxon rank sum test: P = 0.16. d. IPI synchronization limit distribution for acoustic pulse train responses in real neurons. Mean: sync = 25.7 ms, mixed = 13.4 ms, Wilcoxon rank sum test: P < 0.02.
Mentions: We also hypothesized that the strength of a neuron’s inhibitory input would impact its temporal fidelity. Delayed inhibition suppresses random spiking that can occur between acoustic pulses; the resulting stimulus-locked response has a higher vector strength, thus providing a better temporal representation of the acoustic pulse train. As synchronized neurons have stronger delayed inhibition than mixed response neurons, synchronized neurons should therefore have higher vector strengths. However, this comes at a cost, as delayed inhibition prevents the response to a second acoustic pulse during the brief period that the neuron is suppressed. Thus mixed response neurons (which have less inhibition) should be able to stimulus synchronize at shorter IPIs than synchronized neurons. As predicted, we observed that for our simulated neuronal population, synchronized neurons had higher maximum vector strengths than mixed response neurons (Fig. 8a, sync = 0.93, mixed = 0.79; Wilcoxon rank sum test, P< 3.3 x 10-52, see Methods), while mixed response neurons had lower stimulus synchronization limits (Fig. 8b, sync = 10.2 ms, mixed = 7.7 ms; Wilcoxon rank sum test, P< 1.3 x 10-43). We observed a similar trend for our real neuron population- synchronized neurons had higher maximum vector strengths than mixed response neurons (Fig. 8c, sync = 0.68, mixed = 0.60; Wilcoxon rank sum test, P = 0.16), while mixed response neurons had lower stimulus synchronization limits (Fig. 8d, sync = 25.7 ms, mixed = 13.4 ms; Wilcoxon rank sum test, P< 0.02). Although we only observed a statistically significant difference between synchronized and mixed response neurons for the stimulus synchronization limit and not maximum vector strength, this may be due to the limited number of mixed response neurons that we were able to record from (n = 14). These results suggest that blocking inhibition (e.g. by adding a GABA-A antagonist such as Gabazine [28]), effectively decreasing the I/E ratio, should decrease a neuron’s vector strength while increasing its stimulus synchronization limit. However, the relationship between inhibition and temporal fidelity is more complex. While for simulated neurons with the same excitatory input strength, the stimulus synchronization limit of synchronized neurons decreased as the I/E ratio increased (P<0.001, Spearman correlation coefficient), we did not observe a statistically significant trend between the stimulus synchronization limit and I/E ratio in mixed response neurons (P>0.05, Spearman correlation coefficient).

Bottom Line: Using a computational neuronal model, we find that stimulus-locked responses are generated when sound-evoked excitation is combined with strong, delayed inhibition.In contrast to this, a non-synchronized rate representation is generated when the net excitation evoked by the sound is weak, which occurs when excitation is coincident and balanced with inhibition.Together these data suggest that feedforward inhibition provides a parsimonious explanation of the neural coding dichotomy observed in auditory cortex.

View Article: PubMed Central - PubMed

Affiliation: Institute of Behavioural Neuroscience, Department of Experimental Psychology, University College London, London, United Kingdom.

ABSTRACT
In auditory cortex, temporal information within a sound is represented by two complementary neural codes: a temporal representation based on stimulus-locked firing and a rate representation, where discharge rate co-varies with the timing between acoustic events but lacks a stimulus-synchronized response. Using a computational neuronal model, we find that stimulus-locked responses are generated when sound-evoked excitation is combined with strong, delayed inhibition. In contrast to this, a non-synchronized rate representation is generated when the net excitation evoked by the sound is weak, which occurs when excitation is coincident and balanced with inhibition. Using single-unit recordings from awake marmosets (Callithrix jacchus), we validate several model predictions, including differences in the temporal fidelity, discharge rates and temporal dynamics of stimulus-evoked responses between neurons with rate and temporal representations. Together these data suggest that feedforward inhibition provides a parsimonious explanation of the neural coding dichotomy observed in auditory cortex.

No MeSH data available.