- Top of page
Summary: Purpose: Fast ripples (FRs) are interictal, pathological, high-frequency oscillations in the 200- to 600-Hz range, which can be recorded from limbic regions capable of generating spontaneous seizures in rodent models of epilepsy and in human mesial temporal lobe epilepsy. To evaluate the spatial stability of FR-generating brain areas over long periods, we monitored interictal FR oscillations in rats with chronic recurrent spontaneous seizures.
Methods: After unilateral intrahippocampal injection of kainic acid, 22 rats were video monitored until spontaneous behavioral seizures occurred, and then implanted with multiple hippocampal, dentate gyrus, and entorhinal cortex microelectrodes. Electrophysiological monitoring of microelectrode sites was carried out during daily 8-h recordings for periods ranging from 6 to 98 days.
Results: Interictal FRs were recorded from discretely localized areas, adjacent to non–FR-generating areas in dentate gyrus and entorhinal cortex. The location of interictal FR oscillations remained fixed, and the electrophysiological pattern of FRs remained the same over the time of our study. For the duration of monitoring, sites initially recording interictal FRs continued to display FR oscillations, and sites that initially did not record FRs never demonstrated FR activity. A direct relation was seen between the total number of electrode contacts recording interictal FRs and the frequency of spontaneous seizure generation (p < 0.0001).
Conclusions: These results suggest that interictal FRs reflect abnormal discharges from a fixed pathologic substrate imbedded within less-epileptogenic tissue, and that spontaneous seizure frequency is dependent on the extent and distribution of this pathologic substrate.
Wide-band electrophysiological recordings from patients with mesial temporal lobe epilepsy (MTLE) and from the intrahippocampal kainate rat model of this disorder have revealed very high-frequency oscillations in the range of 200 to 600 Hz, which we have termed fast ripples (FRs) (1,2). In both patients and rats, FRs can be recorded in dentate gyrus (DG), hippocampus proper, and entorhinal cortex (EC), often in association with other epileptiform events such as interictal spikes and paroxysmal gamma-frequency oscillations. As opposed to these latter nonspecific EEG transients, which can be recorded over wide areas of the epileptic brain, FRs appear only in rats and people with spontaneous seizures and are uniquely associated with tissue capable of generating spontaneous epileptic seizures (3,4). FRs, therefore, may be a surrogate marker of epileptogenicity. Because FRs can be recorded before the appearance of spontaneous seizures during the development of epileptogenesis in the kainic acid (KA)-treated rat (5), and FRs are involved in spontaneous seizure generation (4), these pathologic oscillations may be a cause, rather than a result of epileptic activity, and may reflect neuronal mechanisms responsible for epilepsy in MTLE.
Both in vivo and in vitro studies have suggested that FRs represent hypersynchronous bursts of population spikes generated by neurons localized to discrete areas imbedded within larger areas of less epileptogenic brain tissue (6). In vitro electrophysiological recordings from rat epileptogenic hippocampal slices have revealed that the size of FR-generating areas can be increased with small amounts of bicuculline (6); however, the long-term spatial stability of these interictal FR-generating areas in vivo has not been demonstrated. It is conceivable, therefore, that the neuronal mechanisms underlying FR generation only appear to be discretely localized. In this view, all areas of the epileptogenic dentate, hippocampus proper, and EC could contain exactly the same potential to generate FRs, and although these epileptiform events manifest within discrete areas at any given time, they might change their location randomly over time.
The purpose of this study, therefore, was to determine the spatial stability of discretely localized FR-generating neuron clusters over time.
- Top of page
A total of 124 spontaneous behavioral seizures (mean = 5.6/mo) were recorded from 22 epileptic rats. FRs were consistently found in 16 rats and never seen in six. Table 1 illustrates that FRs were recorded within the hippocampal–entorhinal circuitry ipsilateral to the side of KA injection in 16 of 22 rats. Moreover, in 14 of these 16 rats, FRs occurred only unilaterally in the posterior DG near the site of KA injection and in adjacent EC, but not in the anterior (dorsal) DG. In two rats, FRs were recorded bilaterally in all implanted microelectrodes except the left anterior DG (6 and 16 in Table 1). These rats also had unilateral lesions, but had a much higher frequency of spontaneous seizures than did the others.
Table 1. Spatial distribution of the recording sites
| 1|| 3||−−||++||++||−−||−−||−−||30 (36)|
| 2|| 1||−−||−−||−−||−−||−−||−−|| (21)|
| 3|| 3||−−||++||−−||−−||−−||−−||11 (11)|
| 4|| 6||−−||++||−−||−−||−−||−−||9 (9)|
| 5|| 1||−−||+−||−+||−−||−−||−−||60 (60)|
| 6||12||−+||++||++||−−||++||++||98 (98)|
| 7|| 2||−−||−−||−−||−−||−−||−−|| (43)|
| 8|| 4||−−||++||++||−−||−−||−−||11 (11)|
| 9|| 5||−−||++||−−||−−||−−||−−||5 (6)|
|10|| 3||−−||+−||+−||−−||−−||−−||6 (6)|
|11|| 1||−−||++||++||−−||−−||−−||15 (15)|
|12|| 3||−−||−−||−−||−−||−−||−−|| (58)|
|13|| 5||−−||++||++||−−||−−||−−||21 (21)|
|14|| 2||−−||−−||−−||−−||−−||−−|| (19)|
|15|| 3||−−||++||++||−−||−−||−−||26 (26)|
|17|| 4||−−||++||++||−−||−−||−−||28 (28)|
|18|| 2||−−||−−||−−||−−||−−||−−|| (28)|
|19|| 3||−−||++||++||−−||−−||−−||8 (8)|
|20|| 2||−−||−−||−−||−−||−−||−−|| (30)|
|21|| 5||−−||++||++||−−||−−||−−||12 (12)|
|22|| 3||−−||++||++||−−||−−||−−||6 (6)|
|Total||124|| || || || || || ||375|
|Mean|| 5.6|| || || || || || || 17|
|Probability|| ||0.05||0.68||0.54||0||0.09||0.09|| |
FRs were recorded for periods of 5–98 days (mean, 17 days). In all but two of 16 rats with FRs, these oscillations were recorded from the same location on every experimental day. In rats 1 and 5, FRs disappeared for several days and then reoccurred at the same locations. A specific example of the degree of spatial stability of FRs during long-term recording from one rat is presented in Fig. 1. Part A illustrates examples of averaged FRs recorded from three locations in the DG and two in the EC 47 days (thick lines) and 61 days (thin lines) after KA injection. Although the amplitude and shape of the slow waves recorded in DG2, DG3, EC1, and EC2 changed over the 2-week period between the two recordings, the high-frequency oscillations remained evident, and changes in FR amplitude were relatively independent of slow-wave amplitude or polarity. DG1 never displayed FRs, although it was only 3.5 mm from DG2. In both DG2 and DG3, a decrement in amplitude of the oscillations occurred on day 61. However, the duration and frequency of the oscillations were undiminished in EC1 and EC2. Power spectral analysis (Fig. 1B) shows two separate peaks within the frequency band 100 to 600 Hz, one within 100 to 200 Hz (Ripples) and another between 300 and 600 Hz, with a peak ∼400 Hz (FRs). Ripples are normal oscillations and are present at all recording sites, whereas FRs are abnormal and seen at only in four recording sites. Within this 2-week period, the 400-Hz peak appeared consistently in the same electrodes with no significant difference in FR frequency at any recording site. Cross-correlation analysis revealed the continued presence of a temporal relation between areas generating FRs (Fig. 1C).
Figure 1. Characteristics of fast ripples (FRs) recorded with five fixed microelectrodes within three different brain areas of the hippocampal–entorhinal circuitry ipsilateral to the kainic acid lesion in a single rat. DG1, anterior dentate gyrus; DG2, posterior dentate gyrus (3.5 mm from DG1); DG3, posterior dentate gyrus 1.5 mm away from DG2; EC1 and EC2, entorhinal cortex with 1.5 mm separation between the recording sites. A, B: FRs recorded 47 days (heavy lines) and FRs recorded 61 days (light lines) after kainic acid injection. A: Averages of FRs (n = 300) in each recording site. B: Power spectral analysis of averaged FRs (second peak and arrow; see text for details; n = 300). C: Cross-correlograms between FRs recorded at the EC1 site and FRs recorded at EC2, DG2, and DG3. The upward black cross-correlograms represent FRs recorded at 47 days, and the downward, gray cross-correlograms represent FRs recorded at 61 days.
Download figure to PowerPoint
Figure 2 illustrates the results of a Spearman nonparametric correlation analysis between the frequency of seizures/month and the number of recording sites where FRs were observed. A significant positive correlation was found between the number of sites generating FRs and the frequency of spontaneous seizures. The correlation remains significant with removal of the two rats with highest and the three rats with lowest seizure rates. No trends were detected that suggested increases or decreases in the rate of FRs or seizures during the period of monitoring.
Figure 2. Correlation between the numbers of microelectrodes recording fast ripples (FRs) and the average probability of seizure occurrence per day. Each circle represents one of the 22 epileptic rats. The number of electrodes implanted was always six pairs per animal. The coefficient of correlation (r) was 0.64, with p value <0.0001.
Download figure to PowerPoint
- Top of page
The main finding of these experiments is the demonstration that FR occurrence in discrete areas of the brains of rats with intrahippocampal KA-induced epilepsy is spatially stable over periods ranging from days to months. Although electrophysiological features of FR oscillations may change over longer periods in rats where seizures became more severe or spontaneously remit, this aspect of the evolution of epileptogenesis was not studied here. Electrode contacts that recorded FRs initially did so on almost every recording day, whereas electrode contacts that failed to record FRs initially never recorded them. This suggests that the neuronal substrate generating FRs is fixed, and argues for the existence of enduring structural disturbances within discrete areas, rather than a more diffuse pathologic substrate that manifests as discrete areas of FR generation that randomly change location over time. Although previous in vivo and in vitro studies led us to suggest that FR-generating tissue consists of small, pathologically interconnected neuron (PIN) clusters (5,6), neither these experiments nor our previous studies have clearly defined the size of FR-generating tissue in three dimensions. From the data reported here, it may be reasonable also to consider a mosaic of epileptogenic neuron clusters of varying sizes imbedded in less-epileptogenic tissue, with the density, or total volume, of pathologic substrate in DG being greatest near the KA lesion. We did not specifically examine for a correlation between the size of the KA lesion area and the probability or rate of FR generation; however, it may be that if the KA lesion is too big, no seizures and no FRs occur. Bilateral lesions were not necessary for bilateral seizures and FRs to occur.
That the number of contacts recording FRs correlated with the frequency of spontaneous seizures suggests that seizure generation is dependent on the spatial extent and distribution of neurons involved in the synchronous bursting discharges responsible for FRs. No contacts recorded FRs in six animals with infrequent seizures, suggesting that FR-generating areas, if present, were small or few in number, whereas the two rats with the most frequent seizures showed FRs not only at all contacts in ipsilateral posterior DG and EC, but also in these structures contralaterally, and in anterior ipsilateral DG, indicating large and widespread areas of epileptogenic pathologic substrate. In vitro experiments have demonstrated that the size of these areas can be enlarged by γ-aminobutyric acid (GABA) antagonists (6), providing a mechanism by which nearby areas may eventually coalesce and become synchronized to give rise to spontaneous ictal events. The demonstration of discrete fixed interictal areas of electrophysiologically identified epileptogenic substrate imbedded within less epileptogenic tissue has important implications for designing future experiments to elucidate the basic mechanisms underlying spontaneous seizure generation.