Limits...
High Stimulus-Related Information in Barrel Cortex Inhibitory Interneurons.

Reyes-Puerta V, Kim S, Sun JJ, Imbrosci B, Kilb W, Luhmann HJ - PLoS Comput. Biol. (2015)

Bottom Line: We also demonstrate that ensembles of INH cells jointly provide as much information about such stimuli as comparable ensembles containing the ~20% most informative EXC neurons, however presenting less information redundancy - a result which was consistent when applying both theoretical information measurements and linear discriminant analysis classifiers.These results suggest that a consortium of INH neurons dominates the information conveyed to the neocortical network, thereby efficiently processing incoming sensory activity.This conclusion extends our view on the role of the inhibitory system to orchestrate cortical activity.

View Article: PubMed Central - PubMed

Affiliation: Institute of Physiology, University Medical Center of the Johannes Gutenberg University, Mainz, Germany.

ABSTRACT
The manner in which populations of inhibitory (INH) and excitatory (EXC) neocortical neurons collectively encode stimulus-related information is a fundamental, yet still unresolved question. Here we address this question by simultaneously recording with large-scale multi-electrode arrays (of up to 128 channels) the activity of cell ensembles (of up to 74 neurons) distributed along all layers of 3-4 neighboring cortical columns in the anesthetized adult rat somatosensory barrel cortex in vivo. Using two different whisker stimulus modalities (location and frequency) we show that individual INH neurons--classified as such according to their distinct extracellular spike waveforms--discriminate better between restricted sets of stimuli (≤6 stimulus classes) than EXC neurons in granular and infra-granular layers. We also demonstrate that ensembles of INH cells jointly provide as much information about such stimuli as comparable ensembles containing the ~20% most informative EXC neurons, however presenting less information redundancy - a result which was consistent when applying both theoretical information measurements and linear discriminant analysis classifiers. These results suggest that a consortium of INH neurons dominates the information conveyed to the neocortical network, thereby efficiently processing incoming sensory activity. This conclusion extends our view on the role of the inhibitory system to orchestrate cortical activity.

No MeSH data available.


Influence of network size on decoding performance and ensemble-based stimulus location information for a representative experiment.(A1) Effect of network size on decoding performance when neurons were chosen in ascending order, i.e. firstly selecting those neurons carrying a lower amount of information (up to n = 74 neurons recorded in this experiment). Three different traces are presented for networks containing only excitatory (red triangles), only inhibitory (blue circles), or both classes of neurons (grey line). Dashed horizontal line represents the chance level (50% since two whiskers were stimulated in this illustrative experiment). (A2) Effect of network size on performance when neurons were selected in descending order, i.e. firstly those neurons carrying a higher amount of information. Due to the asymptotic form of the resulting curves, only the values for network sizes of up to 15 neurons are presented in this case. Otherwise same as in A1. (B1) Ensemble-based mutual information when neurons are selected in ascending order. The maximum value for the resulting information is 1 bit, i.e. the quantity necessary to distinguish unequivocally between two whiskers. (B2) Mutual information when neurons are selected in descending order. Only the values for network sizes of up to 15 neurons are plotted in this case. Green dashed squares represent the highest sized group of INH neurons, and the lowest sized group of EXC neurons carrying at least the same amount of total information than the ensemble of INH cells (0.87 bits, see panel C2). (C1) Information redundancy (equivalent to negative values of synergy) in networks selected in ascending order. (C2) Information redundancy when neurons are selected in descending order. Note the shorter scale of the y-axis as compared to C1. The redundancy values found in the two equally informative groups (n = 8 INH vs. n = 9 EXC neurons) were directly compared in Fig 6A, along with analogous data from the remaining experiments.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4476555&req=5

pcbi.1004121.g005: Influence of network size on decoding performance and ensemble-based stimulus location information for a representative experiment.(A1) Effect of network size on decoding performance when neurons were chosen in ascending order, i.e. firstly selecting those neurons carrying a lower amount of information (up to n = 74 neurons recorded in this experiment). Three different traces are presented for networks containing only excitatory (red triangles), only inhibitory (blue circles), or both classes of neurons (grey line). Dashed horizontal line represents the chance level (50% since two whiskers were stimulated in this illustrative experiment). (A2) Effect of network size on performance when neurons were selected in descending order, i.e. firstly those neurons carrying a higher amount of information. Due to the asymptotic form of the resulting curves, only the values for network sizes of up to 15 neurons are presented in this case. Otherwise same as in A1. (B1) Ensemble-based mutual information when neurons are selected in ascending order. The maximum value for the resulting information is 1 bit, i.e. the quantity necessary to distinguish unequivocally between two whiskers. (B2) Mutual information when neurons are selected in descending order. Only the values for network sizes of up to 15 neurons are plotted in this case. Green dashed squares represent the highest sized group of INH neurons, and the lowest sized group of EXC neurons carrying at least the same amount of total information than the ensemble of INH cells (0.87 bits, see panel C2). (C1) Information redundancy (equivalent to negative values of synergy) in networks selected in ascending order. (C2) Information redundancy when neurons are selected in descending order. Note the shorter scale of the y-axis as compared to C1. The redundancy values found in the two equally informative groups (n = 8 INH vs. n = 9 EXC neurons) were directly compared in Fig 6A, along with analogous data from the remaining experiments.

Mentions: As expected, when applying LDA classifiers to ensembles of neurons selected in ascending order, the decoding performance of both INH and EXC neurons grew supralinearly (Fig 5A1). Remarkably, a small group of INH cells (8 in the illustrative experiment) reached a high decoding performance (83.7%) comparable to a much larger group of EXC cells (65 neurons). This result suggests that while a substantial proportion of EXC neurons conveyed no or little information about the stimulus location, most of INH neurons were highly informative. In contrast, when neurons were selected in descending order, the decoding performance grew promptly and asymptotically for both groups of INH and EXC neurons (Fig 5A2). The 8 most informative EXC neurons presented a decoding performance (85.4%) similar to that reached by INH neurons, indicating that only a small proportion of EXC neurons can be considered as informative as INH neurons, while the majority of EXC cells have a much smaller decoding capacity. Independently of the neuronal type we found that the most informative 9 cells (12.2% of the total) reached a decoding performance (93%) close to that reached by the whole population (93.5%) (Fig 5A2). This finding strongly supports a sparse stimulus encoding scheme in which a minority of cells convey the majority of stimulus-related.


High Stimulus-Related Information in Barrel Cortex Inhibitory Interneurons.

Reyes-Puerta V, Kim S, Sun JJ, Imbrosci B, Kilb W, Luhmann HJ - PLoS Comput. Biol. (2015)

Influence of network size on decoding performance and ensemble-based stimulus location information for a representative experiment.(A1) Effect of network size on decoding performance when neurons were chosen in ascending order, i.e. firstly selecting those neurons carrying a lower amount of information (up to n = 74 neurons recorded in this experiment). Three different traces are presented for networks containing only excitatory (red triangles), only inhibitory (blue circles), or both classes of neurons (grey line). Dashed horizontal line represents the chance level (50% since two whiskers were stimulated in this illustrative experiment). (A2) Effect of network size on performance when neurons were selected in descending order, i.e. firstly those neurons carrying a higher amount of information. Due to the asymptotic form of the resulting curves, only the values for network sizes of up to 15 neurons are presented in this case. Otherwise same as in A1. (B1) Ensemble-based mutual information when neurons are selected in ascending order. The maximum value for the resulting information is 1 bit, i.e. the quantity necessary to distinguish unequivocally between two whiskers. (B2) Mutual information when neurons are selected in descending order. Only the values for network sizes of up to 15 neurons are plotted in this case. Green dashed squares represent the highest sized group of INH neurons, and the lowest sized group of EXC neurons carrying at least the same amount of total information than the ensemble of INH cells (0.87 bits, see panel C2). (C1) Information redundancy (equivalent to negative values of synergy) in networks selected in ascending order. (C2) Information redundancy when neurons are selected in descending order. Note the shorter scale of the y-axis as compared to C1. The redundancy values found in the two equally informative groups (n = 8 INH vs. n = 9 EXC neurons) were directly compared in Fig 6A, along with analogous data from the remaining experiments.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4476555&req=5

pcbi.1004121.g005: Influence of network size on decoding performance and ensemble-based stimulus location information for a representative experiment.(A1) Effect of network size on decoding performance when neurons were chosen in ascending order, i.e. firstly selecting those neurons carrying a lower amount of information (up to n = 74 neurons recorded in this experiment). Three different traces are presented for networks containing only excitatory (red triangles), only inhibitory (blue circles), or both classes of neurons (grey line). Dashed horizontal line represents the chance level (50% since two whiskers were stimulated in this illustrative experiment). (A2) Effect of network size on performance when neurons were selected in descending order, i.e. firstly those neurons carrying a higher amount of information. Due to the asymptotic form of the resulting curves, only the values for network sizes of up to 15 neurons are presented in this case. Otherwise same as in A1. (B1) Ensemble-based mutual information when neurons are selected in ascending order. The maximum value for the resulting information is 1 bit, i.e. the quantity necessary to distinguish unequivocally between two whiskers. (B2) Mutual information when neurons are selected in descending order. Only the values for network sizes of up to 15 neurons are plotted in this case. Green dashed squares represent the highest sized group of INH neurons, and the lowest sized group of EXC neurons carrying at least the same amount of total information than the ensemble of INH cells (0.87 bits, see panel C2). (C1) Information redundancy (equivalent to negative values of synergy) in networks selected in ascending order. (C2) Information redundancy when neurons are selected in descending order. Note the shorter scale of the y-axis as compared to C1. The redundancy values found in the two equally informative groups (n = 8 INH vs. n = 9 EXC neurons) were directly compared in Fig 6A, along with analogous data from the remaining experiments.
Mentions: As expected, when applying LDA classifiers to ensembles of neurons selected in ascending order, the decoding performance of both INH and EXC neurons grew supralinearly (Fig 5A1). Remarkably, a small group of INH cells (8 in the illustrative experiment) reached a high decoding performance (83.7%) comparable to a much larger group of EXC cells (65 neurons). This result suggests that while a substantial proportion of EXC neurons conveyed no or little information about the stimulus location, most of INH neurons were highly informative. In contrast, when neurons were selected in descending order, the decoding performance grew promptly and asymptotically for both groups of INH and EXC neurons (Fig 5A2). The 8 most informative EXC neurons presented a decoding performance (85.4%) similar to that reached by INH neurons, indicating that only a small proportion of EXC neurons can be considered as informative as INH neurons, while the majority of EXC cells have a much smaller decoding capacity. Independently of the neuronal type we found that the most informative 9 cells (12.2% of the total) reached a decoding performance (93%) close to that reached by the whole population (93.5%) (Fig 5A2). This finding strongly supports a sparse stimulus encoding scheme in which a minority of cells convey the majority of stimulus-related.

Bottom Line: We also demonstrate that ensembles of INH cells jointly provide as much information about such stimuli as comparable ensembles containing the ~20% most informative EXC neurons, however presenting less information redundancy - a result which was consistent when applying both theoretical information measurements and linear discriminant analysis classifiers.These results suggest that a consortium of INH neurons dominates the information conveyed to the neocortical network, thereby efficiently processing incoming sensory activity.This conclusion extends our view on the role of the inhibitory system to orchestrate cortical activity.

View Article: PubMed Central - PubMed

Affiliation: Institute of Physiology, University Medical Center of the Johannes Gutenberg University, Mainz, Germany.

ABSTRACT
The manner in which populations of inhibitory (INH) and excitatory (EXC) neocortical neurons collectively encode stimulus-related information is a fundamental, yet still unresolved question. Here we address this question by simultaneously recording with large-scale multi-electrode arrays (of up to 128 channels) the activity of cell ensembles (of up to 74 neurons) distributed along all layers of 3-4 neighboring cortical columns in the anesthetized adult rat somatosensory barrel cortex in vivo. Using two different whisker stimulus modalities (location and frequency) we show that individual INH neurons--classified as such according to their distinct extracellular spike waveforms--discriminate better between restricted sets of stimuli (≤6 stimulus classes) than EXC neurons in granular and infra-granular layers. We also demonstrate that ensembles of INH cells jointly provide as much information about such stimuli as comparable ensembles containing the ~20% most informative EXC neurons, however presenting less information redundancy - a result which was consistent when applying both theoretical information measurements and linear discriminant analysis classifiers. These results suggest that a consortium of INH neurons dominates the information conveyed to the neocortical network, thereby efficiently processing incoming sensory activity. This conclusion extends our view on the role of the inhibitory system to orchestrate cortical activity.

No MeSH data available.