however, with different cells having different overall coherence levels, subtracting the mean coherence from each cell resulted in a strong anticorrelation between gamma-50 and gamma-80 coherence (raw coherence: r = −0.47, p = 0.0004; z-scored coherence: r = −0.91, p < 0.0001), indicating that cells' relative preference for cohering at 50 Hz and 80 Hz was anticorrelated. Furthermore, comparison of the observed 50–80 Hz coherence cross-frequency correlation with that obtained from the shuffled bootstrap samples revealed the correlation between 50 Hz and 80 Hz coherence to be much smaller than expected (compare Figures 14C,D; white arrows, z = −7.33, p < 0.0001). Taken together, these results indicate that overall, FSIs cohered at 50 Hz or 80 Hz independently. It is important to note that this result is not implied by the previously shown relationship between power in these frequency bands (Figure 4C); in theory it is possible for single neurons to fire coherently in both these ranges (and some do), but the data in Figure 14 indicate that overall, at least partially distinct populations of FSIs cohered at gamma-50 and gamma-80 frequencies.
