Examples of Hippocampal EEG Patterns and accompanying Power Spectral Analyses


FIGURE 1. Rhythmical Slow Wave Activity (RSA) recorded from the dentate gyrus of the hippocampus. Traces recorded simultaneously from the same dentate gyrus with the top trace recorded from the granular/hilar region (red), and the bottom trace recorded from the molecular layer. Both sweeps exhibit the regular amplitude (approx 1mV peak to peak) theta frequency (approx 7Hz) wave form. However, note that the EEG epoch recorded from the granular layer also exhibits a 15-30Hz frequency small amplitude signal riding on top of the dominant theta pattern. This low amplitude fast activity is characteristic of cell body layers, and also occurs in CA1 pyramidal layer.


Power Spectral Analyses are a way of expressing dominant frequencies within a wave form. Using a Fast Fourrier Transform (FFT), the wave form is mathematically decomposed into its constituent parts of varying frequencies and the area beneath the curves calculated. Thus, the area or "power" is plotted at each of the frequencies or "spectra". Overall the Power Spectral analysis can be read much like a frequency histogram, with the various frequencies of the wave form laid out systematically along the ordinate, and their dominance or proportional occurrence plotted on a log scale along the abscissa, since area is measured in "squared" units.

FIGURE 2. Power Spectral analysis of RSA bearing epochs, including samples displayed in Figure 1. Transforms averaged from the same dentate gyrus recordings with the red function originally recorded from the granular/hilar region, and the green function originally recorded from the molecular layer. As the animal was awake and behaving freely during EEG recording, subsequent theta has a bandwidth of approximately 6-9Hz, with an apparent peak frequency of about 7Hz. These characteristics can also be seen in the raw traces displayed in Figure 1. In the power Spectral Analyses the area beneath the curve, or power at these frequencies is readily apparent as the dominant FFT peak occurring from 5 to 10Hz, with the peak frequency, or most power, at 7Hz. Also as noted in Figure 1., the red trace from the granular layer carried a smaller amplitude fast activity pattern, which was not present in the green trace recorded from the molecular layer. This distinction can also be seen in the power spectral analyses, as the red function shows more power in the higher frequencies then does the green function. However, even though the presence of the granular fast activity bandwidth is detectable, it has no obvious peak frequency, and further, it is of nowhere near the same power as the theta bandwidth. This is simply a function of the fast activity's small amplitude, since the FFT calculates the area beneath the curve, it integrates both the proportional presence of a frequency and the amplitude at which that frequency occurs. As a side note power spectral analyses typically include a procedure for tapering off the ends of a record, in order to avoid cutting a waveform abruptly and potentially aliasing a wave form of a much lower frequency. Procedures such as the "Hanning Window" involve integrating trace ends with a diminutive cosine function. However, as can be seen by the amount of power at the extremely low frequencies, the FFT at left did not include such a procedure, and the power at these frequencies may be deemed spurrious.


FIGURE 3. Non-RSA pattern recorded the dentate gyrus of the hippocampus. Traces recorded simultaneously from the same dentate gyrus with the top trace recorded from the granular/hilar region (red), and the bottom trace recorded from the molecular layer (green). Historically, much of the hippocampal EEG work has been done in the CA1 region of the dorsal hippocampus, and it is here that most of the slow wave activity patterns were described and defined. As such, even though typically Large amplitude Irregular Activity (LIA) is the other dominant pattern in CA1 of the awake animal, it is more appropriate to identify the similar pattern in the dentate gyrus simply as non-theta, since some of the defining characteristics of CA1 LIA are not present in the dentate.


FIGURE 4. Power Spectral analyses of non-RSA bearing epochs, including sample displayed in Figure 3. Note, in contrast to Figure 2, which depicts the spectral analyses for theta, the functions here appear to have no discernable peaks and simply taper off as the frequencies increase. However, as in Figure 3, the red function originally recorded from the granular layer exhibits greater power than the green function at the higher frequencies.