Extended Data Figure 3: Isolation of dmPFC neuronal assemblies. | Nature

Extended Data Figure 3: Isolation of dmPFC neuronal assemblies.

From: Prefrontal neuronal assemblies temporally control fear behaviour

Extended Data Figure 3

a, Top: example of a binarized rate histogram matrix computed from 55 PNs recorded in a single animal. Bottom: examples of correlation matrices computed for particular time bins from the same 55 PNs (bin size: 150 ms). b, Corresponding example of principal component scores for the same population of 55 PNs and freezing behaviour as a function of time. For the displayed component (first principal component), the distribution of scores is different for freezing and no freezing periods. This shows that freezing episodes are associated with a specific profile of activity for the simultaneously recorded neurons taken as a population. c, Averaged Mahalanobis distance between neuronal population profiles correlated with freezing and no freezing epochs as a function of the length of the sliding window used to produce the binarized rate histograms (see Methods). The optimal separation for neuronal population profiles correlated with freezing and no freezing epochs was obtained with sliding window lengths between 150 and 250 ms (n = 7 mice, Friedman repeated-measures one-way ANOVA on ranks test, *P < 0.05). Error bars, mean ± s.e.m. d, Corresponding example distribution of principal component scores for neuronal population profiles correlated with freezing (black dots) and no freezing (white dots) epochs (Wilcoxon signed rank test, P < 0.05). e, Corresponding example of freezing probability as a function of neuronal population principal component score. Freezing probability associated with each score is compared with surrogate data (black dots) for which freezing intervals have been shuffled. Plain and open red dots represent neuronal population profiles that are significantly over-represented compared with chance during freezing and no freezing periods, respectively (one sample t-test, surrogate data versus actual freezing probability for each score, for scores equal to or above 1.15, freezing probability superior to chance level and P < 0.05). The red dotted line represents the boundary between neuronal population profiles correlated with freezing and no freezing epochs. f, Top: coactivation matrix averaged over freezing epochs from the same 55 dmPFC PNs recorded in a single animal. The strength of the coactivation between pairs of neurons is expressed as the percentage of coactivation compared with chance. Bottom: surrogate data. Coactivation matrix averaged over freezing epochs from the same 55 dmPFC PNs recorded in a single animal for which spike trains have been shuffled 50 times. g, Average percentage of ANs (AP index > 1) and ONs (AP index < 1) to freezing-activated neuronal assemblies (n = 7 mice, ON: 59.6 ± 4.5%; AN: 40.4 ± 4.5%). hm, Left: example of freezing probability as a function of neuronal population principal component score for individual mice as in e. Right: normalized coactivation matrix averaged over freezing predicting epochs from dmPFC PNs recorded in individual animals (n = 29, 40, 46, 26, 45 and 49 neurons). The strength of the coactivation between pairs of neuron is expressed as the percentage of coactivation compared with chance. Error bars, mean ± s.e.m.

Back to article page