Fig. 4: Cross-temporal decoding and stability of cell types for novel and familiar associations. | Nature Communications

Fig. 4: Cross-temporal decoding and stability of cell types for novel and familiar associations.

From: Static and dynamic coding in distinct cell types during associative learning in the prefrontal cortex

Fig. 4

Cross-temporal decoding, stable data point classification, and stability index for novel associations in NovelMap (a, b, e) and familiar associations in FamMap (c, d, f) for the BW and NW populations. The alignment refers to the presentation of the instructor stimulus. a, c Cross-temporal decoding, where the y-axis shows the time bins used for training and the x-axis shows the time bins used for testing the classifier. Normalized classification accuracy values are color-coded. b, d The maps show the classification of each data point (except for the on-diagonal data points, not considered in this analysis and left in blue by convention) of the cross-temporal decoding as static (yellow) or not (blue) (see Methods). e, f Stability indices quantify the stability of the coding over time for each cell type (see Methods). High index values indicate strong stability. Shadow areas represent ±1 SD of indices computed independently by repeating decoding by resampling neurons with replacement for each cell type population (see Methods). Horizontal black lines show the time points with a significant difference in stability between the two populations (cluster-based permutation test p < 0.001). Source data are provided as a Source Data file.

Back to article page