Comparing these two correlations provided a measure determining which assembly (i.e., old or new) has been expressed in a given theta cycle during learning (see Experimental Procedures). Positive assembly expression values indicate times at which the pyramidal activity patterns preferentially expressed the new cell assemblies during

learning (i.e., more similar to the postprobe), while negative ones point to the expression ERK inhibitor of the old assemblies (i.e., more similar to the preprobe). The instantaneous assembly expression values indicated that within many earlier trials, both the old and the new pyramidal assembly representations were expressed in nonoverlapping theta cycles, with later trials dominated by the new patterns (Figures 2 and S3A–S3D). Moreover, the expression strength of the new assemblies improved during the course of learning, suggesting their refinement. Similar expression of the new and old assemblies can be observed when measured within gamma oscillatory cycles (30–80 Hz; see Experimental Procedures), and the assembly expression scores measured during gamma oscillations correlated significantly (p < 0.00001) with those measured in the overlaying theta cycles (Figures S3E–S3G). These temporal fluctuations between distinct assemblies were not

merely resulting from a change in the animal’s trajectory Carnitine dehydrogenase as no such reorganization of place cell click here assemblies occurred in the cued version of the task (Dupret et al., 2010). The switching between old and new assemblies observed here is similar to previous studies in which cell assembly patterns rapidly flicker between

distinct representations of the same location (Jackson and Redish, 2007; Jezek et al., 2011; Kelemen and Fenton, 2010). The firing rate of many interneurons also fluctuated on a fast time scale that followed this assembly flickering (Figure 3A). As suggested by data from the cued task, these rate fluctuations of interneurons associated with allocentric learning were bigger than those that could be expected due to changes in locomotor, spatial behavior or by natural intrinsic variability (Figures S2D and S2E). Moreover, 72% of our CA1 interneurons exhibited a significant correlation (p < 0.05) between their instantaneous firing rate and the theta-paced expression strength of new pyramidal assemblies. Those that exhibited significant positive correlations—referred to as “pInt” – increased their instantaneous rate at times when the new representation was preferentially expressed ( Figures 3B and 3C; n = 86 interneurons) while the ones with negative correlation – referred to as “nInt”—decreased their firing during the same moments ( Figures 3B and 3D; n = 131 interneurons).