The analysis of neuronal ensembles often suffers from low spiking activity. As statistical analysis relies on sufficiently large samples, rare joint events in neuronal ensembles usually fail to reach statistical significance. We show that statistical power can be enhanced considerably by careful preselection of specific, separated temporal stripes distributed throughout the whole time series in which spiking is more likely to occur. In a data set of two-photon calcium imaging from the visual cortex of mice (Cleppien et al., 2024+), we identified short temporal stripes of interest from the underlying population slow wave activity using a specifically designed multi scale statistical peak detection algorithm (Messer et al., 2020). As individual spiking activity could only be observed during these peaks, we focused the statistical analysis on the peak times, thus excluding time periods with near-zero firing rates. Choosing only one time bin around every peak also resolved questions concerning temporal binning. As a consequence, the estimated firing rates of ensembles were considerably enhanced. In this way, even rare higher order spiking events can reach statistical significance.