Limits...
Effects of isoflurane anesthesia on ensemble patterns of Ca2+ activity in mouse v1: reduced direction selectivity independent of increased correlations in cellular activity.

Goltstein PM, Montijn JS, Pennartz CM - PLoS ONE (2015)

Bottom Line: As compared to anesthesia, populations of V1 neurons coded more mutual information on opposite stimulus directions during wakefulness, whereas information on stimulus orientation differences was lower.Increases in correlations of calcium activity during visual stimulation were correlated with poorer population coding, which raised the hypothesis that the anesthesia-induced increase in correlations may be causal to degrading directional coding.The population-wide decrease in coding for stimulus direction thus occurs independently of anesthesia-induced increments in correlations of spontaneous activity.

View Article: PubMed Central - PubMed

Affiliation: Center for Neuroscience, Swammerdam Institute for Life Sciences, University of Amsterdam, Amsterdam, The Netherlands; Research Priority Program Brain and Cognition, University of Amsterdam, Amsterdam, The Netherlands.

ABSTRACT
Anesthesia affects brain activity at the molecular, neuronal and network level, but it is not well-understood how tuning properties of sensory neurons and network connectivity change under its influence. Using in vivo two-photon calcium imaging we matched neuron identity across episodes of wakefulness and anesthesia in the same mouse and recorded spontaneous and visually evoked activity patterns of neuronal ensembles in these two states. Correlations in spontaneous patterns of calcium activity between pairs of neurons were increased under anesthesia. While orientation selectivity remained unaffected by anesthesia, this treatment reduced direction selectivity, which was attributable to an increased response to the -direction. As compared to anesthesia, populations of V1 neurons coded more mutual information on opposite stimulus directions during wakefulness, whereas information on stimulus orientation differences was lower. Increases in correlations of calcium activity during visual stimulation were correlated with poorer population coding, which raised the hypothesis that the anesthesia-induced increase in correlations may be causal to degrading directional coding. Visual stimulation under anesthesia, however, decorrelated ongoing activity patterns to a level comparable to wakefulness. Because visual stimulation thus appears to 'break' the strength of pairwise correlations normally found in spontaneous activity under anesthesia, the changes in correlational structure cannot explain the awake-anesthesia difference in direction coding. The population-wide decrease in coding for stimulus direction thus occurs independently of anesthesia-induced increments in correlations of spontaneous activity.

Show MeSH

Related in: MedlinePlus

Population coding: Mutual information of stimulus orientation and direction as a function of ensemble size.(A) Mutual information of single trial ΔF/F responses to stimuli that differed 45° in movement direction. Mutual information was quantified as mean bits per trial across differently sized groups of neurons (ranging from 1 to 64 cells), averaged (± SEM) across unique stimulus combinations for the awake (red line) and anesthetized (blue line) condition. (B) Idem, for stimuli differing 90° in direction. (C) Idem, for stimuli differing 135°. (D) Idem, for stimuli differing 180° (opposite movement directions). (E) The mean mutual information on stimulus direction for a sample size of 64 neurons. Mutual information for each unique stimulus combination was determined by averaging across 2000 unique samples of 64 neurons from the entire population. The mean (± SEM) was taken across unique stimulus combinations (n = 8 for 45°, 90° and 135°; n = 4 for 180°). Data from awake (red) and anesthetized (blue) mice are displayed separately as a function of the angular difference in stimulus combinations (* p < 0.05, *** p < 0.001).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4338011&req=5

pone.0118277.g004: Population coding: Mutual information of stimulus orientation and direction as a function of ensemble size.(A) Mutual information of single trial ΔF/F responses to stimuli that differed 45° in movement direction. Mutual information was quantified as mean bits per trial across differently sized groups of neurons (ranging from 1 to 64 cells), averaged (± SEM) across unique stimulus combinations for the awake (red line) and anesthetized (blue line) condition. (B) Idem, for stimuli differing 90° in direction. (C) Idem, for stimuli differing 135°. (D) Idem, for stimuli differing 180° (opposite movement directions). (E) The mean mutual information on stimulus direction for a sample size of 64 neurons. Mutual information for each unique stimulus combination was determined by averaging across 2000 unique samples of 64 neurons from the entire population. The mean (± SEM) was taken across unique stimulus combinations (n = 8 for 45°, 90° and 135°; n = 4 for 180°). Data from awake (red) and anesthetized (blue) mice are displayed separately as a function of the angular difference in stimulus combinations (* p < 0.05, *** p < 0.001).

Mentions: Depending on recording conditions, direction selectivity may be highly sensitive to noise in tuning curves: A spurious large-amplitude response may create the apparent notion of direction selectivity in a neuron that is merely orientation-tuned. Because the awake state is associated with more sources of stimulus independent fluctuations in calcium signals than the anesthetized state (e.g. movement artifacts, eye movements), we quantified to what extent the difference in direction selectivity is dependent on the signal-to-noise ratio of the cells included in our analysis. Signal-to-noise ratio was quantified by dividing the mean response to the preferred direction by the standard deviation of the individual trial responses for this direction. Subsequently, we calculated tuning curve parameters for the entire group of cells while, in each iteration, removing the cell with the lowest signal-to-noise ratio from the remaining sample. This resulted in a curve that displays the difference between the awake and anesthetized state for increasing signal-to-noise ratios of the included group of cells (S4 Fig.). If the larger direction selectivity index in the awake condition would be explained by increased noise, this difference should be reduced with each removal of a low signal-to-noise cell from the group, until no difference in direction selectivity index would be visible when considering the group containing only the highest signal-to-noise cells. As expected, the response amplitude to the preferred- and direction increased with rising signal-to-noise ratio, which resulted in increasing values for the orientation selectivity index, depending on which data were used for the SNR quantification (i.e., an increasing OSI under anesthesia when cells were selected on SNR in the same condition, and similarly for the awake condition; S4B-D and G-I Fig.). Tuning width and direction selectivity index, however, remained largely unchanged (Fig. 4A,E,F and J), which indicates that the awake-anesthesia difference in direction selectivity index is not due to increased noise (or reduced SNR) in the awake condition.


Effects of isoflurane anesthesia on ensemble patterns of Ca2+ activity in mouse v1: reduced direction selectivity independent of increased correlations in cellular activity.

Goltstein PM, Montijn JS, Pennartz CM - PLoS ONE (2015)

Population coding: Mutual information of stimulus orientation and direction as a function of ensemble size.(A) Mutual information of single trial ΔF/F responses to stimuli that differed 45° in movement direction. Mutual information was quantified as mean bits per trial across differently sized groups of neurons (ranging from 1 to 64 cells), averaged (± SEM) across unique stimulus combinations for the awake (red line) and anesthetized (blue line) condition. (B) Idem, for stimuli differing 90° in direction. (C) Idem, for stimuli differing 135°. (D) Idem, for stimuli differing 180° (opposite movement directions). (E) The mean mutual information on stimulus direction for a sample size of 64 neurons. Mutual information for each unique stimulus combination was determined by averaging across 2000 unique samples of 64 neurons from the entire population. The mean (± SEM) was taken across unique stimulus combinations (n = 8 for 45°, 90° and 135°; n = 4 for 180°). Data from awake (red) and anesthetized (blue) mice are displayed separately as a function of the angular difference in stimulus combinations (* p < 0.05, *** p < 0.001).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4338011&req=5

pone.0118277.g004: Population coding: Mutual information of stimulus orientation and direction as a function of ensemble size.(A) Mutual information of single trial ΔF/F responses to stimuli that differed 45° in movement direction. Mutual information was quantified as mean bits per trial across differently sized groups of neurons (ranging from 1 to 64 cells), averaged (± SEM) across unique stimulus combinations for the awake (red line) and anesthetized (blue line) condition. (B) Idem, for stimuli differing 90° in direction. (C) Idem, for stimuli differing 135°. (D) Idem, for stimuli differing 180° (opposite movement directions). (E) The mean mutual information on stimulus direction for a sample size of 64 neurons. Mutual information for each unique stimulus combination was determined by averaging across 2000 unique samples of 64 neurons from the entire population. The mean (± SEM) was taken across unique stimulus combinations (n = 8 for 45°, 90° and 135°; n = 4 for 180°). Data from awake (red) and anesthetized (blue) mice are displayed separately as a function of the angular difference in stimulus combinations (* p < 0.05, *** p < 0.001).
Mentions: Depending on recording conditions, direction selectivity may be highly sensitive to noise in tuning curves: A spurious large-amplitude response may create the apparent notion of direction selectivity in a neuron that is merely orientation-tuned. Because the awake state is associated with more sources of stimulus independent fluctuations in calcium signals than the anesthetized state (e.g. movement artifacts, eye movements), we quantified to what extent the difference in direction selectivity is dependent on the signal-to-noise ratio of the cells included in our analysis. Signal-to-noise ratio was quantified by dividing the mean response to the preferred direction by the standard deviation of the individual trial responses for this direction. Subsequently, we calculated tuning curve parameters for the entire group of cells while, in each iteration, removing the cell with the lowest signal-to-noise ratio from the remaining sample. This resulted in a curve that displays the difference between the awake and anesthetized state for increasing signal-to-noise ratios of the included group of cells (S4 Fig.). If the larger direction selectivity index in the awake condition would be explained by increased noise, this difference should be reduced with each removal of a low signal-to-noise cell from the group, until no difference in direction selectivity index would be visible when considering the group containing only the highest signal-to-noise cells. As expected, the response amplitude to the preferred- and direction increased with rising signal-to-noise ratio, which resulted in increasing values for the orientation selectivity index, depending on which data were used for the SNR quantification (i.e., an increasing OSI under anesthesia when cells were selected on SNR in the same condition, and similarly for the awake condition; S4B-D and G-I Fig.). Tuning width and direction selectivity index, however, remained largely unchanged (Fig. 4A,E,F and J), which indicates that the awake-anesthesia difference in direction selectivity index is not due to increased noise (or reduced SNR) in the awake condition.

Bottom Line: As compared to anesthesia, populations of V1 neurons coded more mutual information on opposite stimulus directions during wakefulness, whereas information on stimulus orientation differences was lower.Increases in correlations of calcium activity during visual stimulation were correlated with poorer population coding, which raised the hypothesis that the anesthesia-induced increase in correlations may be causal to degrading directional coding.The population-wide decrease in coding for stimulus direction thus occurs independently of anesthesia-induced increments in correlations of spontaneous activity.

View Article: PubMed Central - PubMed

Affiliation: Center for Neuroscience, Swammerdam Institute for Life Sciences, University of Amsterdam, Amsterdam, The Netherlands; Research Priority Program Brain and Cognition, University of Amsterdam, Amsterdam, The Netherlands.

ABSTRACT
Anesthesia affects brain activity at the molecular, neuronal and network level, but it is not well-understood how tuning properties of sensory neurons and network connectivity change under its influence. Using in vivo two-photon calcium imaging we matched neuron identity across episodes of wakefulness and anesthesia in the same mouse and recorded spontaneous and visually evoked activity patterns of neuronal ensembles in these two states. Correlations in spontaneous patterns of calcium activity between pairs of neurons were increased under anesthesia. While orientation selectivity remained unaffected by anesthesia, this treatment reduced direction selectivity, which was attributable to an increased response to the -direction. As compared to anesthesia, populations of V1 neurons coded more mutual information on opposite stimulus directions during wakefulness, whereas information on stimulus orientation differences was lower. Increases in correlations of calcium activity during visual stimulation were correlated with poorer population coding, which raised the hypothesis that the anesthesia-induced increase in correlations may be causal to degrading directional coding. Visual stimulation under anesthesia, however, decorrelated ongoing activity patterns to a level comparable to wakefulness. Because visual stimulation thus appears to 'break' the strength of pairwise correlations normally found in spontaneous activity under anesthesia, the changes in correlational structure cannot explain the awake-anesthesia difference in direction coding. The population-wide decrease in coding for stimulus direction thus occurs independently of anesthesia-induced increments in correlations of spontaneous activity.

Show MeSH
Related in: MedlinePlus