Comparison of fiber tract low frequency stimulation to focal and ANT stimulation in an acute rat model of focal cortical seizures

Abstract

Background:

Current implementations of direct brain stimulation for epilepsy in patients involve high-frequency (HFS) electrical current and targeting of grey matter. Studies have shown that low-frequency (LFS) fiber-tract stimulation may also prove effective. To compare the efficacy of high-frequency grey matter stimulation to the low-frequency fiber tract stimulation technique a well-controlled set of experiments using a single animal model of epilepsy is needed.

Objective:

The goal of this study was to determine the relative efficacy of different direct brain stimulation techniques for suppressing seizures using an acute rat model of focal cortical seizures.

Methods:

4-AP was injected into the S1 region of cortex in rodents over three hours. LFPs were recorded from the seizure focus and mirror focus to monitor seizure frequency during the experiments. CC-LFS, HFS-ANT, Focal-HFS, or a transection of the CC was applied.

Results:

Stimulation of the CC yielded a 65%+/−14% (p = 0.0014) reduction of seizures in the focus and a 97% +/−15% (p = 0.0026) reduction in the mirror focus (n = 7). By comparison transection of the CC produced a 65%+/−18% reduction in the focus and a non-statistically significant reduction of 57% +/−18% (p = 0.1381) in the mirror focus (n = 5). All other methods of stimulation failed to have a statistically significant effect on seizure suppression.

Conclusions:

LFS of the CC is the only method of stimulation to significantly reduce seizure frequency in this model of focal cortical seizures. These results support the hypothesis that LFS of fiber tracts has significant potential for seizure control.

Introduction

Epilepsy is one of the most prevalent neurological disorders in the world with a prevalence of 7 cases per 1,000 persons and an annual cumulative incidence of 67 new cases per 100,000 persons[1,2]. Of those patients with epilepsy between 28–37% suffer from refractory epilepsy[3]. Also, patients that do not respond to antiepileptic medication may not be candidates for resection depending on the number of foci present, the location of those foci, and the ability to determine the location of a seizure focus. Patients with cortical epilepsies constitute 20–40% of patients that are eligible for surgery [4]. Furthermore, of the patients that undergo resection for a focal cortical epilepsy, only 66% of patients experience seizure freedom [5]. For patients that are not candidates for resection the use of another procedure, the corpus callosotomy, is an option for preventing the generalization of seizures. This procedure is traditionally performed by cutting the anterior portion of the corpus callosum. Although the corpus callosum innervates 70–80% of the cortex [6], it is the anterior half which is believed to be responsible for the generalization of tonic-clonic seizures [7]. The response rate of patients to a complete transection is only 10% higher than the rate for a transection of the anterior portion [8]. Corpus callosotomy can be effective in suppressing seizures however it has become less common due to the adverse effects that result from the procedure [9].

Direct brain stimulation has developed as a significantly less invasive alternative to previous surgical techniques. Recently, two direct brain stimulation techniques have been granted FDA approval for use in treating refractory epilepsies. Although many grey matter stimulation targets have been investigated along with different stimulation parameters including frequency only the responsive neurostimulation (RNS) system and stimulation of the anterior nucleus of the thalamus (ANT) have been implemented clinically.

The RNS system utilizes a closed loop system to detect seizures and then applies high frequency electrical current to the seizure focus. In one clinical study it was shown that this technique reduced seizure frequency by 37% after 2 years[10]. In a follow-up study the seizure reduction rate increased to 53% following 2 years of stimulation[11]. Alternatively, ANT stimulation utilizes open loop high-frequency stimulation of the ANT to suppress seizures. Stimulation remains on since seizures begin to reappear after stimulation is ended [12]. In a clinical trial of bilateral stimulation of the ANT a 56% reduction in seizure frequency was observed in patients at 2 years post implantation[13]. These results are promising for patients suffering from intractable seizures however there is still significant room for improvement.

Both techniques rely on electrical stimulation of grey matter targets with frequencies >100 Hz [12]. However, several studies have shown that low-frequency stimulation of white matter tracts may also have a substantial effect on seizures[14–18]. In one study in humans with intractable epilepsy, low-frequency stimulation of the fornix during a 4 hour session yielded a 90% reduction in seizure odds 2 days after the stimulation ended[14]. In an acute focal cortical seizure model in rodents it was shown that low-frequency (20 Hz) stimulation of another fiber tract (the corpus callosum) gave rise to a 95% reduction in seizures[18].

Although there are significantly more studies documenting the effects of focal and ANT stimulation for seizure suppression and the white matter stimulation studies have yet to be implemented in a double blind clinical trial, there are now two different paradigms of direct brain stimulation for epilepsy. The grey matter high-frequency stimulation techniques have been shown to be effective in suppressing seizures and the low frequency white matter stimulation techniques have shown preliminary results suggesting that they could also be effective. Unfortunately, studies that have documented seizure suppression have been performed in different animal models or different patient populations. There have been no studies performed to examine the differences in efficacy between these two paradigms under the same conditions. Here we test each direct brain stimulation technique in a single model acutely in rodents. To compare the effect of seizure suppression to a non-stimulation based treatment that is clinically useful we compare fiber tract stimulation to a callosotomy of fibers innervating the focal cortical region. Each therapeutic technique for seizure suppression is evaluated for its suppressive capacity. The results of this study should help to inform the development of new and improved brain stimulation methods for seizure suppression in patients with refractory epilepsy.

Materials and Methods

Surgical Procedure

All animal procedures were conducted in accordance with the guidelines reviewed and approved by the institutional animal care and use committee of Case Western Reserve University. Forty-eight adult male Sprague-Dawley rats (150–300 g; Charles River) were used in this study. Isoflurane was administered at a concentration of between 1–3% and vitals were monitored while the animals were under anesthesia. The rat’s head was secured using a stereotactic frame prior to any further manipulation. A small incision was made along the rostrocaudal axis to expose the skull and all connective tissue was removed with manual abrasion with hydrogen peroxide. Subsequently, several burr holes were drilled into the skull for either electrode placement, microsyringe insertion, or transection. Depending on the experiment there were different numbers of holes created in the skull to accommodate the stimulation electrodes (figure 1A) or transection (figure 1B). In all animals two burr holes were made in the skull over right somatosensory cortex (S1), one was made over the left S1, and another in in the most posterior bone sutures of the skull. All coordinates were determined through the use of a rat brain atlas[19]. One electrode was placed in one of the 2 holes over the right S1 (anteroposterior [AP] = −0.4 mm, lateral = 2.3 mm, depth = −1.4 mm) another electrode was placed in the contralateral S1 (AP = −1.6 mm, lateral = −2.0 mm, depth = −1.4 mm). A stainless-steel screw electrode was placed in the posterior bone sutures and used as a reference for recording. A microsyringe was inserted into the second burr hole over the right S1 (AP = −1.6 mm, lateral = 2.0 mm, depth = −1.4 mm). For corpus callosum stimulation burr holes were drilled in the midline and electrodes were placed in the two holes (AP = −1.3 mm, lateral = +/−0.6 mm, depth −3.0 mm). For focal stimulation an additional hole was made over the right S1 along with a hole over the right posterior bone suture for a screw electrode to be used as a return for stimulation current. The stimulation electrode was placed in the third burr hole over the right S1 (AP = −1.1 mm, lateral = 2.2 mm, depth = −1.4 mm). For stimulation of the ANT a screw electrode was also placed in a burr hole over the right posterior bone suture. Also, a burr hole was drilled over the right ANT and an electrode was inserted (AP = −1.5 mm, lateral = 1.5 mm, depth = −5.3 mm). To transect a portion of the corpus callosum 3 large adjacent burr holes were drilled along the midline of the rostrocaudal axis of the skull to form a continuous opening for the insertion of a knife (AP = 0 to −3 mm, lateral = 0 mm) as shown in figure 1B. All electrodes were fixed to the surface of the skull with dental cement.

Figure 1: Experimental setup.

(A) Recording electrode placement in the seizure focus (right) and mirror focus (left) in blue, with a screw reference electrode in black. Microsyringe is shown inserted into the right S1 loaded with a 4-AP cocktail that is injected hourly in all experiments. Electrode placement for electrical stimulation animal groups is illustrated beneath the sham group setup depiction: corpus callosum stimulation group (purple), seizure focus stimulation group (Yellow), and anterior nucleus of the thalamus stimulation group (Green). (B) Positioning of the blade for the corpus callosum transection group.

Focal Cortical Model

We used a model similar to what was reported in [18]. Briefly, 1 μL injections of a cocktail of 4-aminopyridine (4-AP) were administered once per hour at the beginning of each of the three hours of the experiment in order to maintain spontaneous seizures. During the intervention hour both stimulation and the injection were performed simultaneously at the top of the hour. The cocktail consisted of 30 mM 4-AP in an artificial cerebrospinal fluid solution containing a reduced calcium and magnesium concentration (1.2 mM Ca2+ and 0.6 mM Mg2+). The injections were made in the right S1 region to create a seizure focus in the right somatosensory cortex. Anesthetic depth was decreased by lowering the concentration of Isoflurane to <2% to limit the effect of anesthesia on neural excitability.

Data Acquisition and Seizure Identification

Local field potential (LFP) recordings were sampled at 40 kHz and amplified by 100. LFPs were monitored during the entirety of every experiment to determine seizure frequency in the electroencephalogram (EEG). Seizures were identified using the criteria developed by Nissinen et. Al, wherein an EEG segment must have an amplitude greater than 2 times the baseline, the majority of the spectral power must be > 5 Hz, and this segment must last for longer than 5 seconds[20]. All EEG segments meeting these criteria were classified as seizures. The duration of these seizures was recorded for purposes of comparison between different experimental time periods. EEG recorded during stimulation was cleaned using template subtraction and a median filter to remove the stimulation artifact using the method utilized in [18].

Direct Brain Stimulation

For CC stimulation the amplitude was determined by finding 50% of the maximum evoked potential amplitude in the S1. Based on this the current for all experiments involving stimulation was set to 4 mA. Electrical current was applied in the form of a 2-mA biphasic (4 mA peak-peak) current pulse with a 100-microsecond pulse width (each phase was 100 microseconds, giving a total biphasic pulse width of 200 microseconds). The current pulses were delivered continuously for 1 hour at either a high-frequency (200 Hz) or a low-frequency (20 Hz). For grey matter stimulation a screw electrode over the posterior ipsilateral cortex was used as a return for the current. CC stimulation was applied between a pair of electrodes positioned parallel to the longitudinal axis of the callosal axons.

Transection

A partial transection was made in some animals along a portion of the rostrocaudal axis. A small blade was inserted into the most anterior position in the skull opening down to 4 mm beneath the surface. The knife was then moved caudally until it reached the most posterior position in the opening (figure 1B). The knife was then pulled directly up out of the brain to complete the transection of a 3 mm section of the corpus callosum. Only the region of the corpus callosum innervating the focal region was cut (as determined by evoked potentials). The transection was confirmed by monitoring the evoked potential in the cortical electrodes. Once the evoked potentials disappeared the transection of the portion of CC was considered complete.

Experimental Design

Experiments were carried out in 48 male Sprague Dawley rats. In every experiment we recorded one hour of baseline activity followed by one hour during which either stimulation was applied, a transection was made, or no action was taken (sham condition). Subsequently, we recorded for one additional hour to observe any after-effect. For the initial set of experiments 4 groups of 7 animals each (28 total, n=7) were divided into a corpus callosum low-frequency (CC-LFS) group, a focal high-frequency stimulation (Focal-HFS) group, an anterior nucleus of the thalamus high-frequency stimulation (ANT-HFS) group, and a sham group. The CC-LFS group received 1 hour of 4-mA 20 Hz stimulation, both the Focal-HFS and ANT-HFS group received 1 hour of 4-mA 200 Hz stimulation, and the sham group received nothing. In another set of experiments 3 additional groups of animals were added to switch the frequency parameters between white matter and grey matter targets. Each group contained 5 animals and was compared against 5 animals from the previous group of 7 sham animals (15 additional animals, n=5). The animals were split into either the corpus callosum high-frequency (CC-HFS) group, the focal low-frequency (Focal-LFS) group, the anterior nucleus of the thalamus low-frequency (ANT-LFS) group, a corpus callosum group (CC-CUT) or the 5 sham animals from the previous experiments (SHAM). With these experiments the CC group received 1 hour of 4-mA 200 Hz stimulation while both the Focal-LFS and ANT-LFS received 1 hour of 4-mA 20 Hz stimulation. In a final set of experiments a group of 5 animals was subjected to a partial transection (CC-CUT) of the corpus callosum during the second hour of the experiment. For all experiments the total time spent seizing during each hour was normalized to the total time in each period (1 hour). The percent time spent seizing during each hour was compared between each group and the sham group. In order to make these comparisons we used a Friedman test with Dunn multiple comparisons post hoc test with a significance level of 0.05. A nonparametric analysis of variance was utilized due to the non-normal distribution of these data as determined by a D’Agostino-Pearson omnibus normality test (P < 0.05).

Results

Comparison of CC-LFS, Focal-HFS, and ANT-HFS

We first compared the efficacy of CC-LFS, focal-HFS, and ANT-HFS in suppressing seizures by determining the percent time spent seizing in each group of animals and comparing results to the time-matched sham group. During the first hour of recording seizures began at the site of injection in the S1 after about 10 minutes and spread to the contralateral S1 after another 5 to 10 minutes. The seizures were typically larger in amplitude in the seizure focus (figure 2A) than in the mirror focus (figure 2B). Seizures demonstrated the same characteristics as noted in [18] with high-frequency and amplitude segments that follow discrete patterns and repeat frequently.

Figure 2: Stimulation paradigm comparison.

Sample recording from (A) the seizure focus and (B) the contralateral somatosensory cortex (mirror focus) during baseline and CC LFS. Comparison in percent time spent seizing during baseline, stimulation, and post-stimulation between the different stimulation conditions and sham in (C) the seizure focus (* p = 0.0014, n = 7) and in (D) the contralateral cortex (** p = 0.0026, n = 7).

The corpus callosum low-frequency stimulation group was the only stimulation group that experienced a reduction in seizures. During CC-LFS, activity consisted of short bursts of spikes (<5 seconds) in the mirror focus (figure 2B) and occasional brief seizures in the cortical focus itself (figure 2A). CC-LFS reduced seizures by 65% (p = 0.0014, n = 7) in the seizure focus (figure 2C) and by 97% (p = 0.0026, n = 7) in the contralateral mirror focus (figure 2D). There were no significant differences between the other stimulation techniques and the sham group. Although non-significant, focal-HFS and ANT-HFS generated an increase instead of an expected decrease in seizure duration. Recordings in the focal region showed that focal-HFS increased the percent change in time spent seizing during stimulation by 6.6% +/− 17% while ANT-HFS generated a 9.1% +/−18% increase. In the mirror focus focal-HFS and ANT-HFS produced a 25% +/−19% and 18% +/−22% increase in time spent seizing respectively.

Effect of location vs frequency on efficacy

To determine if this disparity in effect might be due to a difference in only one of the parameters rather than the combination of both location and frequency, we reversed the parameter pairings in several additional groups of animals. We applied high-frequency stimulation to the corpus callosum and low-frequency stimulation to the seizure focus and anterior nucleus of the thalamus. Surprisingly, none of these pairings resulted in a decrease in seizures in either the seizure focus (figure 3A) or the mirror focus (figure 3B).

Figure 3: Reversed stimulation parameters comparison.

Percent time spent seizing during baseline, stimulation, and post-stimulation in sham animals, corpus callosum high-frequency stimulation, seizure focus low-frequency stimulation, and anterior nucleus of the thalamus low-frequency stimulation in (A) the seizure focus and (B) the mirror focus (n = 5).

During focal-LFS there was a non-significant 7.4% +/−15% decrease and 14% +/−15% increase in seizure duration in the focus and mirror focus respectively. When LFS was applied to the ANT a non-significant 57% +/−17% and a 43% +/−18% decrease in percent time spent seizing in the focus and mirror focus respectively were observed. Neither reduction was statistically significant although the effect in the focus was close to the 5% significance threshold (p = 0.0592, n = 5). Applying HFS to the corpus callosum resulted in a non-significant 5% +/−16% reduction and a 12% +/−16% increase in seizure duration in the focus and mirror focus respectively.

Comparison between CC transection and CC-LFS

Lastly, instead of applying electrical stimulation we made a transection of the corpus callosum within the region responsible for reciprocally innervating the seizure focus and mirror focus (figure 4A). Transecting this region of the corpus callosum enabled us to directly compare the efficacy of a corpus callosotomy to CC-LFS in seizure suppression.

Figure 4: Partial corpus callosum transection comparison.

(A) Depiction of partial corpus callosum transection through CC tracts connecting the seizure focus (right, purple) to the mirror focus (left, purple). (B) Recording in the seizure focus before and immediately after the transection was completed. Comparison in percent time spent seizing during each hour between sham group, CC cut group, and CC-LFS group in (C) the seizure focus (* p = 0.0165, ** p = 0.0048, n = 5) and (D) the mirror focus (* p = 0.0048, n = 5).

Following a one-hour baseline period, a blade was lowered from the surface of the brain to a region below the CC and moved along the anteroposterior axis to cut only those fibers innervating the focus. Immediately following the transection, activity in both the focus and mirror focus decreased dramatically. However, the seizure activity gradually returned as shown in figure 4B. Typically, after about 40 minutes seizure activity returned to sham levels. During the first hour following the transection only the seizure focus showed a 65% +/−18% reduction in seizures (p = 0.016, n = 5). As shown in figure 4C, the reduction in seizures occurring in the seizure focus caused by a CC transection was comparable to CC-LFS with no significant difference between the CC-transection group and the stimulation group. There was a 57% +/−18% non-statistically significant reduction in seizures in the contralateral cortex (p = 0.1381, n = 5). In comparison, the CC-LFS group demonstrated a seizure suppression of 97% in the contralateral cortex that was significantly more effective than callosotomy of the same fibers used for stimulation (figure 4D).

Discussion

Previous randomized controlled studies of vagal nerve stimulation (VNS) and brain stimulation have reported a seizure suppression rate of around 50% in 50% of patients [10,11,13,21,22]. Many other stimulation targets have been studied with either LFS or HFS in animal studies or pilot studies in patients. The stimulation targets have included the ANT[13,23–28], cerebellum[29–32], hippocampus [33,34], cortex [35,36], amygdala [37], nucleus caudatus [38], ventral hippocampal commissure (VHC) [14–17], and the corpus callosum [18]. Of these targets the only white matter targets studied are the VHC and CC. The VHC and CC have shown similar results in seizure suppression (>90%) within mTLE and focal cortical seizure animal models respectively [15,16,18].

Previous studies have evaluated the various brain stimulation paradigms in different patient populations or animal models such as kainic acid, pilocarpine, alumina cream, penicillin, amygdaloid kindling, genetic absence model, flurothyl, and pentylenetetrazol [39]. However, in this study we attempted to normalize the experimental conditions to compare each technique’s ability to suppress seizures in the same acute model of seizures. We chose to compare two popular direct brain stimulation technologies in patients against a new stimulation paradigm shown to have early success in both animals and patients with refractory epilepsy[14–18]. Low-frequency fiber tract stimulation was tested against focal high-frequency stimulation and anterior nucleus of the thalamus high-frequency stimulation. Additionally, stimulation location and frequency pairings were switched to evaluate the effect of these variables in isolation on efficacy.

To compare each stimulation technique and a transection of the CC we chose an acute focal cortical model of seizures identical to the one reported in Couturier et al. with the exception that we injected the 4-AP solution in the S1 instead of the primary motor cortex (M1) [18]. The seizures shown in figure 2A and 2B are identical to those produced in the M1 and are consistent with our previous results but in a different part of the cortex. In this model, the percent time spent seizing increases each hour in both the focus and mirror focus from between 30–40% in the first hour to between 60–70% percent by the third hour.

The 4-AP model of focal cortical seizures in rodents has been shown to elicit spontaneous seizures under anesthesia however the seizures can dissipate over time [40–44]. In our model we injected 4-AP hourly in a cocktail with reduced magnesium and calcium concentrations to maintain the seizure focus. Similar to previously reported 4-AP neocortical seizure models rodents demonstrated inter-seizure intervals no longer than 20–30 minutes. Unlike models that rely on a single injection of 4-AP, repeated injections of our 4-AP cocktail generated steady seizure rates throughout the entire 3 hours of experimentation. This focal model produced tonic-clonic seizures that increase in severity over time unlike chronic models of focal cortical epilepsy such as tetanus toxin [45,46]. Chronic models of epilepsy give rise to sporadic and spontaneous seizures for an indefinite period of time after the substance used to induce seizures has been cleared by the body. Our acute model of focal cortical seizures relies on repeated doses of 4-AP to generate repetitive seizures that secondarily generalize from the injection site.

As was reported in [18], 20 Hz stimulation of the corpus callosum yielded a seizure suppression of >90% in the contralateral cortex. Our similar finding in the primary somatosensory cortex suggests that the corpus callosum exhibits a substantial potential for stimulation targeted to the region of the cortex containing the seizure focus. Unlike the previous study in the motor cortex we also recorded from the seizure focus itself. CC-LFS was also effective at suppressing seizures in the focus with a suppression rate of 65%.

However, we were not able to suppress seizures with any other technique in either the mirror focus or the seizure focus. The percent time spent seizing under other stimulation protocols was comparable to the time-matched sham group. The only group that showed any noticeable effect on seizures was the ANT-LFS group. Although the reduction was not statistically significant (p = 0.0592, n = 5) it is close enough to warrant further investigation. This effect was only observed in the seizure focus likely because ANT-LFS was applied to the ipsilateral side. A previous study noted that low-frequency stimulation of the anterior nucleus of the thalamus demonstrated a suppressive effect in a kainic acid chronic model of epilepsy while the conventional ANT-HFS failed to suppress seizures [25]. The reason for this discrepancy could be explained by the proximity of the anterior nucleus of the thalamus to several prominent fiber tracts including the mammillothalamic tract and the fornix.

It is possible that either focal-HFS or ANT-HFS may have an improved efficacy within a different frequency range. Typically ANT stimulation frequency is adjusted to a value that provides the greatest seizure suppression for each patient but is usually above 100 Hz [12,13,26]. However one study using 200 Hz stimulation of the ANT found that seizure frequency could be reduced only if stimulation was applied 1 hour prior to administration of the convulsant pilocarpine [47]. Stimulation of the seizure focus in the responsive neurostimulation system (RNS) is usually fixed at 200 Hz however clinicians similarly make adjustments to suit the needs of their patients [11]. Future studies should determine the specific range of frequencies above 100 Hz that are most effective in suppressing focal cortical seizures. It is also likely that the HFS grey matter stimulation techniques are not as effective in this particular acute model of cortical seizures. Since we have not evaluated the use of CC-LFS in a chronic epilepsy model it is possible that the grey matter techniques may prove to be more effective since chronic models have been shown to impact white matter connectivity [48].

We also included a group of animals with a partial transection of the corpus callosum since CC fibers are known to facilitate the generation of bilateral seizures and cutting these fibers is an effective surgical technique for suppressing seizures. We therefore compared the efficacy of transecting the CC to electrical stimulation of the brain for seizure suppression. Corpus callosotomy is the most aggressive procedure available to neurosurgeons to prevent seizures from spreading and would therefore serve as a good standard for which to compare our technique. Furthermore, a clinically proven surgical technique should aid in evaluating a broader spectrum of therapies that are useful in suppressing seizures. We chose not to perform a complete transection of the corpus callosum from genu to splenium since in a previous study we found CC stimulation to be spatially limited to <1–2mm from the stimulation electrodes [18].

It is known that fibers crossing between hemispheres can promote synchronization and seizure generation [49–51]. Therefore, we expected that the callosotomy would decrease seizures on both sides and have a greater effect on the mirror focus. Contrary to our expectation, the transection reduced seizures by 65% in the seizure focus but had a much smaller effect in the mirror focus. This could be explained by the development of an independent focus on the contralateral side. It is possible that the mirror focus had actually become a fully independent focus at some point given the near zero delay between seizure onsets in both hemispheres. Only during the first hour did the mirror focus show a consistent delay in seizure onset from the focus in the contralateral hemisphere. For the remainder of the time the seizures almost always occurred nearly simultaneously. Interestingly, despite a very clear reduction in activity immediately following the cut all activity eventually returned on both sides after an hour. This temporary reduction in seizure activity may reflect a global decrease in cortical excitability due to significant bleeding and tissue damage. The return of seizures may be the result of the tissue stabilizing post-injury (blood clotting, heart rate and blood pressure stabilization). These results show that CC-LFS is more efficient at reducing seizures in this focal model compared to a callosotomy. A limitation of this study is the lack of histological examination to quantify the extent of damage to the CC during the transection. However through monitoring the evoked potentials we ensured that the fibers stimulated during CC-LFS were severed.

Functionally the corpus callosum has been shown to cause inhibition in the cortex [52]. In a previous study of low-frequency stimulation of the hippocampal commissure it was demonstrated that seizure suppression was the result of a GABA_B driven inhibition [53]. The suppressive effect observed with low-frequency stimulation of the corpus callosum may be the result of a similar mechanism. If the callosal fibers are indeed facilitating inhibition then it makes sense that cutting those fibers could result in an increase in excitability over time. The ability of fiber tract LFS to suppress seizures through widespread lowering of excitability could allow clinicians to stimulate different tracts in order to control regions of the brain believed to be involved in seizure generation. The feasibility of controlling seizures in a chronically epileptic brain remains to be determined. Our model only induces spontaneous seizures acutely and therefore provides no means by which to evaluate the effect of stimulating the CC after potential epileptogenic remodeling of the brain. For patients with mTLE the hippocampal commissure has been shown to be a promising target whereas our work in the corpus callosum has suggested that patients with seizures in either the M1 or S1 may benefit significantly from CC-LFS. It is likely that other portions of the CC could be stimulated at low-frequencies to suppress seizures in other portions of the cortex such as the auditory cortex or pre-motor cortex. CC-LFS could be used to suppress seizures originating in cortical regions that cannot be safely resected as is often the case in patients with cortical epilepsies. Furthermore, stimulation of fiber tracts at low-frequencies may provide a unique opportunity for patients suffering from refractory epilepsies that are non-responsive to current high-frequency grey matter stimulation technologies.

Conclusion

Low-frequency stimulation of the corpus callosum was shown to be the only brain stimulation technique capable of suppressing seizures in an acute model of focal cortical seizures in the somatosensory cortex. The effectiveness of stimulation is not due to either the frequency (low) or location (corpus callosum fibers) but rather to the combination of applied low-frequency electrical current and fiber-tract targeting. Furthermore, we determined that CC-LFS is superior to a partial transection of the corpus callosum in suppressing seizures and is completely reversible.

Highlights.

• CC-LFS was compared to Focal-HFS and ANT-HFS in a focal cortical model of seizures

• CC-LFS was compared to CC-HFS, Focal-LFS, (ANT-LFS), and transection of the CC

• CC-LFS suppressed seizures by 65% in the focus and 97% in the mirror focus

• Transection of the CC suppressed seizures by 65% in the focus

• No other technique was able to suppress seizures significantly