Article Figures & Data
Figures
- Figure 1.
Histogram of the preferred theta firing phase of 141 identified pyramidal cells recorded along with the interneurons that are the main subject of this paper. The phase angles have been rotated 90° clockwise from the values derived from our local EEG recordings so that they conform to the convention of Csicsvari et al. (1999). The resulting equivalence of the phase convention is visible by comparing this histogram to Figure 5B of Csicsvari et al. (1999).
- Figure 2.
A, Histogram of the durations of the initial negative-going peak of interneuron spike waveforms. Classification according to negative-going peak duration was in excellent agreement with other criteria; there was a discrepancy only for the single cluster whose initial peak duration was >300 μs. B, Histogram of the overall firing rate for interneurons. The mean value was 20.8 spikes/s.
- Figure 3.
A, Autocorrelations for three theta-modulated interneurons. An approximate value for the 8 Hz frequency of local theta is visible from the four clear peaks for the 500 ms interval on each side of t = 0. B, Autocorrelations for three theta-independent interneurons. Although the precise form of the autocorrelation function varies from cell to cell, there is no indication of theta-frequency modulation.
- Figure 4.
Histogram of oscillation score values (defined in text) for 104 interneurons. The distribution is strongly bimodal such that there are no values in the range 0.4–0.5. Choosing a threshold of 0.45 thus separates the sample into two distinct groups such that high values are for theta-modulated cells and low values are for theta-independent cells.
- Figure 5.
Firing rate as a function of running speed for theta-modulated and theta-independent interneurons. The activity of both cell types increases with running speed, but the effect is greater for theta-modulated cells as confirmed by a significant group by running speed interaction in an ANOVA.
- Figure 6.
High-frequency burst activity by theta independent interneurons. A, High-frequency burst generated by a theta-independent interneuron during pellet chasing. Component spikes of the burst were acquired separately in 1.07 ms frames and then properly aligned from the time stamps to reproduce the intervals. B, Autocorrelation for spikes acquired during pellet chasing. There is a clear mode at a delay of ∼3.5 ms (∼285 Hz). C, Autocorrelation for spikes acquired during slow-wave sleep. The first mode is at ∼1.8 ms (∼514 Hz). Additional modes at about 3.6 and 5.4 ms are also present. By inspection of the raw data, many of the 3.6 and 5.4 ms intervals were due to bursts of >2 spikes. The great difference in the scaling of the y axes of B and C indicate that high-frequency activity was much more common during slow-wave sleep than during pellet chasing.
- Figure 7.
Histogram of preferred theta firing phases (two cycles are shown) for 74 theta-modulated interneurons. The color code emphasizes that there are four preferred phase modes. In this code, class 1 is red, class 2 is green, class 3 is orange, and class 4 is blue. The continuous line represents the sum of fits to four von Mises distributions whose means and dispersions are stated in the text. Note that the blue bin between 300° and 320° contains one cell from class 1. A remarkable feature of the distribution is the isolation of class 3 cells from the others.
- Figure 8.
Phase histogram examples for one interneuron selected from each of the four preferred phase classes; the color code is the same as in Figure 7. The histograms are binned at 20°. The top panel shows the filtered, averaged theta activity recorded at the same time as the class 1 cell. The fitted sine wave shows the sine wave that would have been chosen to represent the negative half of the theta cycle. The preferred phase (middle of the preferred bin) and number of spikes for the four examples are as follows: class 1—350°, 590 spikes; class 2—50°, 1082 spikes; class 3—210°, 2110 spikes; class 4—310°, 1532 spikes.
- Figure 9.
Color coded firing rate map examples for each of the four preferred phase classes. The example interneurons were chosen to illustrate the large spatial coherence range for each class, suggesting that there is not a strong differential contribution of the different classes to the great spatial firing specificity of place cells. (Spatial coherence was calculated as the z-transform of the firing rate correlation between each pixel and its neighbors—see Materials and Methods.) The number to the upper right of each map is the coherence (see Materials and Methods); the number to the bottom right of each map is the average firing rate for the entire recording session.
- Figure 10.
Normalized firing rate as a function of running speed for the four theta-modulated classes; the color code is the same as in Figure 7. Firing rate increases with running speed for each class, but the lack of a significant class by rate interaction from an ANOVA (see text) indicates that the functions are indistinguishable.
- Figure 11.
Proposed match between the preferred theta phase classes seen in freely moving rats and those seen in urethane-anesthetized rats; the color code for phase class is the same as in Figure 7. The phase coordinate is 0° at 3:00 and increases counterclockwise. Theta phase classes in freely moving rats are indicated with solid lines; the urethane classes are indicated with dashed lines. The angle subtended by each of the colored sectors reflects the dispersion of the distribution for the corresponding phase preference class in freely moving rats. The urethane theta phase classes are as follows: bistratified, oriens/lacunosum-moleculare (O-LM), cholecystokinin-expressing (CCK), axo-axonic, and parvalbumin basket cells.
In this issue
