Figure 3.

Firing of CA3 PV-expressing basket and putative pyramidal cells is phase locked to field gamma oscillations, both recorded locally in CA3 and downstream in area CA1. A, Basket cell J31a fired preferentially (bottom trace) at the trough of gamma oscillations in CA1 SP (top trace) and at the peak of CA3 gamma (middle trace). For this recording, CA1 SP gamma was approximated by reversal of gamma recorded from CA1 SR (see Materials and Methods). In both A and B, gray vertical lines indicate spike times. B, A putative pyramidal cell (B36o; bottom line) also fired phase locked to both CA1 (top trace) and CA3 (middle trace) gamma oscillations. Calibrations: 0.2 mV, 50 ms. C, D, Phase distribution histograms of normalized spike counts from the cells shown in A and B. The key in C is valid for C–E. E, CA3 gamma troughs were phase locked to those in CA1. For this recording (black line), most CA3 gamma troughs occurred 212° out of phase with CA1 gamma troughs, consistent with the fact that this cell tended to fire on the troughs of CA1 (C), but on the peaks of CA3 gamma oscillations (D; relative phase difference, 249°). Similarly, for the putative pyramidal cell recording (gray line), the fact that CA3 gamma troughs mostly occurred out of phase with CA1 gamma troughs is consistent with the different phase preferences of this cell's firing relative to gamma in CA1 (C) and CA3 (D). F, G, The preferred firing phase of cells with significantly coupled firing relative to CA1 (F) and CA3 gamma (G; in some cases selected periods to avoid spike artifacts). Numbers in parentheses indicate the number of significantly modulated cells per total number of analyzed cells per cell type. Note the homogeneous phase preferences relative to local gamma oscillations (G) in contrast to CA1 gamma (F). H, The differences in preferred firing phase relative to CA3 and CA1 gamma oscillations roughly match the phase differences between CA1 and CA3 gamma oscillations. Recordings were only used when there was significant coupling of firing to both CA1 and CA3 gamma, and between CA1 and CA3 gamma, and spike artifact did not influence the result in CA3 (5 of 6 basket cells and 12 of 15 pyramidal cells). A black line along the diagonal is for reference. I, The coupling of spikes from basket (black) and putative pyramidal (gray) cells to CA1 gamma was significantly different from their coupling to CA3 gamma (exact paired Wilcoxon signed ranks tests, *, basket, p = 0.031, n = 6; pyr, p = 0.015, n = 15). Error bars indicate SD. J, K, Mean phase angle spectra (J; only significant values are plotted) and mean vector length (K; gray, all values; black, significant values at p < 0.05; Rayleigh test) of CA3 basket cell spikes during periods of theta oscillations, relative to the LFP recorded by the local electrode in CA3.