- Top of page
Purpose: To determine the effects of high-frequency electrical stimulation on electrographic seizure activity during and after stimulation (ON-effect and OFF-effect).
Methods: The modulation and suppression of epileptiform activity during (ON-effect) and after (OFF-effect) high-frequency electrical stimulation was investigated using the high-K+ and picrotoxin hippocampal slice epilepsy models. Uniform sinusoidal fields (50 Hz) were applied with various intensity levels for 1 min across brain slices. Extracellular and intracellular activity were monitored during and after stimulation.
Results: The ON-effects of high-frequency stimulation were highly variable across individual slices and models; ON-effects included modulation of activity, pacing, partial suppression, or activity resembling spreading-depression. On average, epileptic activity, measured as power in the extracellular fields, increased significantly during stimulation. Following the termination of electrical stimulation, a robust poststimulation suppression period was observed. This OFF suppression was observed even at relatively moderate stimulation intensities. The duration of OFF suppression increased with stimulation intensity, independent of ON-effects. Antagonism of GABAAfunction did not directly effect OFF suppression duration.
Conclusions: The present results suggest that “rational” seizure control protocols using intermittent high-frequency electrical stimulation should control for both ON and OFF effects.
Technologies applying electrical stimulation to control pharmacologically intractable epileptic seizures are being actively explored (Velasco et al., 2000c; Cohen-Gadol et al., 2003; Goodman, 2004; Polkey, 2004; Murphy & Patil, 2005; Morrell, 2006; Albensi et al., 2007; Li & Mogul, 2007). A variety of stimulation paradigms including DC or slow-adaptive electric fields (Gluckman et al., 1996; Ghai et al., 2000; Gluckman et al., 2001; Lian et al., 2003), high-frequency stimulation (Bikson et al., 2001; Lian et al., 2003; Feddersen et al. 2007; Jensen & Durand, 2007), and low-frequency pulsed stimulation (Albensi et al., 2004) have been developed. A range of potential suppression mechanisms have been proposed with the specific mechanisms depending on the precise waveform applied (Durand & Bikson, 2001; McIntyre et al., 2004c; Li & Mogul, 2007) and the underlying seizure dynamics (e.g., clinical focus, animal model).
Significant unknowns remain about the mechanisms of electrical-stimulation control of seizures. “Rational” protocols based on quantitative predictive control for optimizing stimulation waveform are lacking (Bikson et al., 2006). Moreover, empirical clinical optimization is limited by the relative infrequency of seizures and safety concerns (Theodore & Fisher, 2004). In vitro epilepsy models provide a preliminary tool to prescreen, characterize, and optimize stimulation paradigms and waveforms (Durand & Bikson, 2001). In this report, we investigated the effects of high-frequency stimulation on high-K+ and picrotoxin hippocampal slice models of epilepsy. We considered effects during stimulation (ON-effects) and poststimulation modulation of activity (OFF-effects). We discuss whether stimulation approaches that balance ON/OFF effects may provide a more robust method for seizure control.
- Top of page
Recordings were obtained from the CA1 or CA3 pyramidal cell regions of hippocampal brain slices (0.35–0.40 mm) prepared from male Sprague–Dawley rats (125–175 g; CCNY-IACUC protocol 0406). A total of 29 animals were used in this study. Slices were superfused in an interface recording chamber at 36°C oxygenated (with 95% O2, 5% CO2) “normal” artificial cerebrospinal fluid (ACSF) consisting of (in mM) 125 NaCl, 3 KCl, 1.25 NaH2PO4, 1.6 CaCl2, 1.5 MgSO4, 26 NaHCO3, and 10 dextrose.
“High-potassium” (high-K+) ACSF consisted of (in mM): 125 NaCl, 8.0 KCl, 1.25 NaH2PO4, 0.5 CaCl2, 1.5 MgSO4, 26 NaHCO3, and 10 dextrose. “Picrotoxin ACSF” was made by adding 100 μM picrotoxin (GABAA-receptor antagonist) to normal ACSF. “High-potassium plus picrotoxin” ACSF was made by adding 100 μM picrotoxin to high-potassium ACSF. Perfusion with these solutions resulted in spontaneous epileptiform activity (electrographic seizures) in the CA1 or CA3 regions of the hippocampus; epileptiform activity was characterized by spontaneous bursts of population-spike trains. Slices in which spreading depression-like activity was observed in the absence of stimulation were excluded. Individual slices were superfused with only a single epileptiform solution. All chemicals were obtained from Sigma-Aldrich (St. Louis, MO, U.S.A.).
Recordings of extracellular field potentials were obtained using glass micropipettes (10–15 MΩ, pulled on a P-97; Sutter Instruments, Novato, CA, U.S.A.) filled with 125–150 mM NaCl. One recording electrode was positioned in the somatic layer of the CA1 or CA3 region. A second electrode was positioned at an isopotential site in the bath (Ghai et al., 2000; Durand & Bikson, 2001). For intracellular recording, the potential from a field electrode positioned next to the intracellular electrode (50–100 MΩ, filled with 3 mM KCl) was subtracted (Bikson et al., 2004; Radman et al., 2007).
Uniform 50 Hz sinusoidal electric fields were generated across individual slices by passing current (A-M Systems stimulus isolator Model 2200, Carlsborg, WA, U.S.A.) between two parallel AgCl-coated silver wires placed on the surface of the ACSF in the interface chamber (Bikson et al., 2004). Stimulation was applied for 1 min. Stimulation intensity (mV/mm) ranged from 70 to 414 mV/mm.
Suppression during ON or OFF periods was defined as reduction in population spike activity to 20% of prestimulation values. ON and OFF power ratios (dB) were quantified by comparing field power during stimulation (1 min) or immediately poststimulation (first 1 min) relative to the power of the 1-min field base-line before stimulation (all 100 Hz high-passed). “Spreading depression-like” events were defined extracellularly as slow shifts in the extracellular field potential ≤−10 mV for >10 s (Haglund & Schwartzkroin, 1990; Tong & Chesler, 2000; Bikson et al., 2003) that were followed by a refractory period (absence of spontaneous or evoked population spikes). “Spreading depression-like” events were defined intracellulary as shifts in membrane potential to ≥15 mV for >10 s (Haglund & Schwartzkroin, 1990). In the analysis, we excluded cases in which spreading depression-like events were induced by stimulation.
Signals were amplified and filtered with an Axoclamp-2B (Axon Instruments, Union City, CA, U.S.A.) and FLA-01 amplifiers (Cygnus Technology, Delaware Water Gap, PA, U.S.A.); then digitized and processed using a Power 1401 and Signal software (CED, Cambridge Electronic Design, Cambridge, U.K.). Additional filtering (including 50 Hz band-stop), statistical analysis, and figure generation were implemented using MATLAB R14 (Mathworks, Inc., Natick, MA, U.S.A.). Results are reported as mean ± standard error. After combining data from all slices for each epilepsy model, Pearson's correlation coefficient was used to determine the total correlation (rt) between electric field intensity, OFF suppression period, and ON power ratio. Significance of correlation (p-value) was calculated using Student's t-test on the combined data; p < 0.05 reported as significant. In addition, we calculated correlation coefficients for each slice and averaged across slices to obtain a pooled within-slice covariance (rw).
- Top of page
The effects of high-intensity 50 Hz sinusoidal electric fields on epileptiform activity were evaluated in the high-K+ and picrotoxin epilepsy models. Modulation during stimulation (ON-effects) and suppression after stimulation (OFF-effects) were quantified. For all three models, ON-effects were classified as: (1) “modulation/pacing”: epileptiform activity remained and population spikes could occur in phase with stimulation; (2) “partial suppression”: epileptiform activity was suppressed for more than 10 s, but less than the 1-min duration of stimulation; (3) “spreading depression-like event”: during stimulation a spreading depression-like event (see Methods) is triggered.
For determining “maximal” on stimulation effects, stimulation intensity was increased until either: (1) partial suppression was observed; (2) a spreading depression-like event was induced (at intensities below spreading depression-like events, modulation/pacing was observed); or (3) the maximal field amplitude tested (320–414 mV/mm) continued to induce modulation/pacing. Unless othwerise stated, the high-intensity ON results reported below refer to these “maximal” stimulation effects.
ON-effects of stimulation during high-K+ epileptifom activity
The effects of stimulation on high-K+ electrographic field activity was evaluated in 24 slices. Low-intensity stimulation resulted in modulation/pacing of activity in all slices tested. High-intensity stimulation (between 200 and 414 mV/mm) resulted in modulation/pacing in 16 slices (Fig. 1A; 16.8 ± 1.0 dB ON power ratio), partial suppression of activity for a portion of the stimulation period in three slices (average field threshold 223 mV/mm; 14.5 ± 1.5 dB ON power ratio; Fig. 1B,C), and a spreading depression-like event in five slices (Fig. 1D). Partial suppresion could be associated with a characteristic slow-field shift; this slow-field shift (Fig. 1C) was distinct from spreading depression-like events as it was relativly short and activity (pacing or spontaneous epileptiform activity) resumed immediately after return to baseline (see Methods).
Figure 1. Effects of high-intensity sinusoidal (50 Hz) stimulation on high-K+-induced epileptiform extracellular field activity. Typical modulation/pacing of activity (A) and examples of partial suppression during stimulation (B, C), and a stimulation induced spreading depression-like event (D); traces from different slices, see text for classification scheme. All signals were 50 Hz band-stop filtered, removing the stimulation artifact, but leaving spontaneous and paced activity. Note that in all cases, particularly at the initiation of stimulation, episodes of population spike pacing were observed; the interspike interval was generally (a multiple of) the stimulation period (20 ms = 1/50 Hz). Episodes of suppression (including during spreading depression-like events) were characterized by the absence of synchronized population activity. In all cases, a poststimulation OFF suppression period was observed.
Download figure to PowerPoint
Intracellular recording confirmed observations with field electrodes (Fig. 2). Intracellularly high-intensity fields resulted in cell modulation/pacing (n = 1; Fig. 2A), transient suppresion (n = 2; Fig. 2B), or a spreading depression-like event (n = 2; Fig. 2C).
Figure 2. Effects of high-intensity sinusoidal (50 Hz) stimulation on high-K+-induced epileptiform intracellular activity. Typical modulation/pacing of activity (A) and example of partial suppression (B) and spreading depression-like event (C) during stimulation. All signals were 50 Hz band-stop filtered, removing the stimulation artifact, but leaving spontaneous and paced action potentials. During stimulation pacing of action potentials generally occurred at the frequency of stimulation (50 Hz) or subharmonic. During stimulation episodes of action potential suppression or action potential attenuation could clearly be observed. Note that in all cases, a poststimulation OFF suppression period was observed.
Download figure to PowerPoint
ON-effects of stimulation during picrotoxin epileptifom activity
High-intensity stimulation (160–400 mV/mm) of picrotoxin induced activity (Fig. 3) resulted in partial suppression of activity (Bikson et al., 2001) in four of eight slices tested (average field threshold 235 mV/mm; 4.2 ± 0.7 dB ON power ratio) and modulation/pacing in three slices (8.3 ± 2.0 dB ON power ratio), with one slice showing a spreading depression-like event. Lower-intensity stimulation (80–100 mV/mm) resulted in activity modulation/pacing in a total of 12 slices tested.
Figure 3. Effects of high-intensity sinusoidal (50 Hz) stimulation on picrotoxin-induced epileptiform extracellular field activity. Lower intensity stimulation resulted in modulation/pacing (A) while higher intensity stimulation resulted in partial suppression of activity (B) or spreading depression-like event (C). All signals were 50 Hz band-stop filtered, removing the stimulation artifact, but leaving spontaneous and paced activity. Note that lower-intensity stimulation could modulate/aggravate activity. In all cases, a poststimulation OFF suppression period was observed.
Download figure to PowerPoint
ON-effects of stimulation during high-K+ plus pictroxin epileptifom activity
In the high-K+ model, inhibitory synaptic function is intact (Jensen & Yaari, 1997). We tested the role of GABA-ergic function during high-frequency stimulation of high-K+ by adding picrotoxin (0.1 mM) during high-K+ bursting. Stimulation (75–414 mV/mm) of high-K+ plus picrotoxin activity resulted in modulation/pacing in five of six slices tested (Fig. 4; 15.6 ± 0.8 dB ON power ratio) and a spreading depression-like event in the remaining slice.
Figure 4. Effects of high-intensity sinusoidal (50 Hz) stimulation on high-K+ plus picrotoxin-induced epileptiform extracellular field activity. Both lower intensity and higher intensity stimulation resulted in modulation/pacing (A, B) or could trigger a spreading depression-like event (C). All signals were 50 Hz band-stop filtered, removing the stimulation artifact, but leaving spontaneous and paced activity. Note that in all cases, particularly at the initiation of stimulation, episodes of population spike pacing were observed; the interspike interval was generally (a multiple of) the stimulation period (20 ms = 1/50 Hz). In all cases, a poststimulation OFF suppression period was observed.
Download figure to PowerPoint
OFF-effects of stimulation
Succesful OFF suppression was defined as a poststimulation suppression period greater than twice the baseline (prestimulation) interelectrographic seizure period. High-intensity stimulation (160–414 mV/mm) resulted in successful OFF suppression in all 19 slices tested in the high-K+ model, six of seven slices tested in the picrotoxin model, and all five slices tested in the high-K+ plus picrotoxin model. The minimum field amplitudes required for OFF suppression were on average 147, 134, and 125 mV/mm in the high-K+, picrotoxin, and high-K+ plus picrotoxin models, respectively. The average poststimulation OFF power ratios, at the minimum stimulation intensities proceeding succesful OFF suppression, were −4.6 ± 0.9 dB, −0.6 ± 0.3 dB, and −0.9 ± 0.5 dB in the high-K+, picrotoxin, and high-K+ plus picrotoxin models, respectively.
In each model, the duration of poststimulation OFF suppression increased with stimulation intensity (Fig. 5, left; significant correlation in all three models with p < 0.01). For the high-K+ and picrotoxin models, no correlation between ON power ratio and poststimulation OFF suppression duration was observed (Fig. 5, right top and center). However, the high-K+ plus pictroxin model did show a correlation (Fig. 5, right bottom). This correlation remained even when a linear effect of stimulus intensity on each of these variables was subtracted (residuals remain correlated with r = 0.57, p < 0.02).
Figure 5. Summary of ON and OFF stimulation effects for each epilepsy model. The relationship between pairs of three metrics were compared (postsuppression duration, stimulation intensity, and ON power ratio) separately for each model (high-K+, picrotoxin, high-K+ plus picrotoxin). Each point represents a field application (with multiple field amplitudes tested in specific slices). rt is the correlation coefficient of combined data; p-values indicate significance of this total correlation. rw is the within-slice correlation coefficient averaged across slices (see Methods), and reflects predictability for a given slice. Symbols indicate pacing (filled circle), partial suppression (open circle), or spreading depression-like events (filled triangle) resulted from given stimulations. Spreading depression-like events were excluded from calculation of correlation coefficients. Note that poststimulation duration increased consistently with stimulation intensity. ON power ratio did not correlate consistently with stimulation intensity or with postsuppression duration (see text).
Download figure to PowerPoint