Journal: eLife

Article Title: Stochastic and deterministic dynamics of intrinsically irregular firing in cortical inhibitory interneurons

doi: 10.7554/eLife.16475

Figure Lengend Snippet: ( a ) Naturalistic stimulus protocol. The cell was stimulated with a constant step of AMPA conductance (g AMPA , reversing at 0 mV) with added conductance Ornstein-Uhlenbeck noise (standard deviation 2% of the step amplitude, τ = 5 ms), combined with a sinusoidal GABA A conductance ( g GABA , reversing at −60 mV) and introduction of positive, zero or negative g Kt . g AMPA was adjusted so that the cell fired close to the frequency of the g GABA inhibitory oscillation. ( b ) Effect of adding g Kt on a slightly irregular-firing cell at 10 Hz g GABA . ( i ) Step current response (black), response to oscillatory conductance stimulus with (green) or without (blue) addition of 3.57 nS g Kt . ( ii ) Spike entrainment synchrony (see Materials and methods, Spike Analysis) to the 10 Hz g GABA oscillation as a function of the oscillation amplitude. Synchrony rises progressively with oscillation amplitude in control (blue), and is depressed by the addition of 3.57 nS g Kt (green). ( iii ) Spike phase histogram for pooled responses to lower amplitude g GABA oscillations (up to 1 nS), showing a reduction in the sharpness of synchrony. ( c ) Example of subtracting g Kt in another irregular-firing cell at 10 Hz g GABA . ( i ) example responses. ( ii ) Subtraction of g Kt (red) increased synchrony to g GABA oscillation over a wide range of amplitudes, when compared to control (blue). ( iii ) spike phase histograms for pooled responses up to 1 nS g GABA oscillations. Subtraction of g Kt enhances the phase preference. ( d ) Lack of effect of g Kt on synchronization to 40 Hz (gamma) oscillation. ( e ) Summary of effects of g Kt perturbation on synchrony in different cells. Each symbol denotes an experiment on an individual cell, showing the ratio of synchrony, evaluated at ≈1/3 of the maximum g GABA amplitude applied in each case, during g Kt injection, normalized to its control value with no injection. 10 Hz: g Kt addition (n = 10, green) and subtraction, (n = 10, red); 40 Hz: g Kt addition (n = 6, green) or subtraction (n = 7, red). At 10 Hz, but not 40 Hz, g Kt perturbation has a significant effect. Wilcoxon nonparametric rank sum test, p=6.5 × 10 –5 for both positive and negative g Kt , and p=0.36 and 0.69 at 40 Hz for positive and negative g Kt respectively. DOI: http://dx.doi.org/10.7554/eLife.16475.023 10.7554/eLife.16475.024 Figure 8—source data 1. Numerical values for . DOI: http://dx.doi.org/10.7554/eLife.16475.024

Article Snippet: To test for differences between two conditions, the two-sided Wilcoxon rank sum test (Matlab Statistics Toolbox function ranksum), equivalent to the Mann-Whitney U test, was used. n , the number of samples, and p, the probability of observing the two distributions under the null hypothesis that they have equal medians, are given in all cases, and p<0.05 is taken as the significance level.

Techniques: Standard Deviation, Control, Injection