Male C57BL/6 mice (8 weeks old and 25~30 g) were housed in an air-conditioned room under a 12 h/12 h light/dark cycle at a controlled room temperature (23 ${}^{\circ }$C $\pm$ 2 ${}^{\circ }$C) and humidity (50%) with free access to food and water. All of the experiments were performed in accordance with the guidelines of the Academia Sinica Institutional Animal Care and Utilization Committee and all NIH guidelines for the care and use of animal subjects were followed.

Pentylenetetrazol, pilocarpine (PIL), and picrotoxin (PTX) were purchased from Sigma-Aldrich (St. Louis, MO, USA). All three drugs were used to induce seizures. Pentylenetetrazol was dissolved in sterile water and injected intraperitoneally at 50 mg/kg as a single injection [20]. Pilocarpine and PTX were dissolved in sterile water and injected intraperitoneally at 300 and 5 mg/kg, respectively, as a single injection [21, 22]. Ethosuximide (ESM) and valproic acid (VPA) were purchased from Sigma-Aldrich (St. Louis, MO, USA) and used to suppress seizures. The doses of ESM were 200 mg/kg (ESM-200 group), 400 mg/kg (ESM-400 group), and 600 mg/kg (ESM-600 group). The doses of VPA were 200 mg/kg (VPA-200 group), 300 mg/kg (VPA-300 group), and 400 mg/kg (VPA-400 group). All of the drugs were dissolved in distilled water and intraperitoneally administered 30 min before PTZ administration to induce seizures. The injection volume for drug administration was 0.1 mL/25 g. The doses of each drug were based on previous studies [20, 21].

Behavior was recorded by a video camera and classified and scored according to criteria from Itoh’s convulsion scale [9, 10]: no. 0 (normal), no. 1 (immobilization), no. 2 (short myoclonic jerks), no. 3 (myoclonic jerks), no. 4 (generalized clonic seizures with a kangaroo posture), and no. 5 (generalized tonic-clonic seizures with a loss of posture tone).

For the magnet-induced current coil system, we referred to the method of the previous study [17]. The NdFeB magnet was attached to the belly of the mouse’s body. The coil was wrapped with a round copper wire (1000 turns) and enclosed by a ring of aluminum in a cylindrical tube (25 cm outer diameter, 45 cm height). Voltage in the coil was induced by a change in the electromagnetic field associated with the magnet’s movement (0.8 cm outer diameter, 0.2 cm thickness). The voltage in the coil was in accordance with Faraday’s law of electromagnetic induction. The voltage depends on the acceleration of the magnet. The voltage was also influenced by the direction of the magnet’s movement. Vertical movements more efficiently changed the voltage compared with horizontal movements. The output voltage signals in the coil were amplified by a differential alternating-current amplifier (Model 1700, A-M system, Carlsborg, WA, USA) and a data acquisition system (Axon Digidata 1322A, Axon Instruments, Union City, CA, USA). The currents that resulted from the animal’s movements were recorded on a hard drive for off-line analysis.

The mice were anesthetized with 4% isoflurane in 100% O${}_2$ and then placed in a stereotaxic apparatus. The animals were maintained under anesthesia with 1.6% isoflurane in 100% O${}_2$, and body temperature was maintained at 36.5–37.5 ${}^{\circ }$C during surgery with a homeothermic blanket system (Model 50-7079, Harvard Apparatus, Holliston, MA, USA). For EEG recordings of the cortex, a small hole was drilled in the skull (the hole was centered 1.5 mm anterior to bregma and 2.0 mm lateral to the midline), and silver-ball electrodes were placed in the hole. The electrodes were implanted in the epidural space. Stainless steel screws were used as the reference and ground electrodes and fixed to the front and back of the skull. All of the electrodes were connected with a connector (catalog no. A79014, Omnetics Connector Corporation, Minneapolis, MN, USA). The self-curing dental acrylic powder was used to affix the Omnetics connector to the skull. After surgery, the mice were placed in cages to recover for 1 week and observed to ensure that no untoward effects of surgery occurred.

One week after surgery, an elastic bandage was used to attach the magnet to the mouse’s body, perpendicular to the peritoneal cavity. The silver-ball electrode was connected with the Omnetics connector to a computer via chronic head stages (RA16CH, TDT, Alachua, FL, USA). Individual mice were then placed inside the test chamber, and the camera was adjusted to ensure that the full chamber could be captured in the videos. Five minutes later, the behavior was videotaped, and EEGs were recorded before and after intraperitoneal PTZ administration. The experimental mice were acutely treated with a convulsive dose of PTZ (50 mg/kg). Control mice received the volume injection of 0.1 mL/25 g for the normal saline solution with an intraperitoneal injection. EEGs and convulsive behavior were recorded for 30 min. All of the data, including brain responses and the video recordings, were stored on a hard disk for off-line analysis.

According to a previous study [22], a box was suspended on a force transducer (Model FT 10, Grass Instruments, Cary, North Carolina, USA). A damper was used to restrict the horizontal movement of the chamber, so vertical movements of the chamber were recorded as the animal’s movement. The output signals from the transducer were connected to a data acquisition system (Axon Digidata 1322A, Axon Instruments, Union City, CA, USA). Vertical displacement of the box that was caused by the animal’s movements was converted to a digital signal and recorded on a hard disk for off-line analysis.

All EEG and behavioral recordings were subjected to fast-Fourier transform (FFT) to a spectrum of 0–50 Hz. Amplitude ratio (%) is calculated as the power amplitude of each frequency band divided by the summation of power amplitude for 0–50 Hz. The amplitude ratio (%) was then calculated for each frequency band. Behavioral ratio (%) was calculated by the formulation [100 * (the numbers of the specific behavior responses/the numbers of all behavioral responses)] (%). The measured behavioral responses included the following: baseline: resting and walking; No. 1: immobilization; No. 2: short myoclonic jerk; No. 3: myoclonic jerk; No. 4: generalized clonic seizures with a kangaroo posture; and No. 5: generalized tonic-clonic seizure with loss of posture tone. Note that, “short myoclonic jerk” was the head nodding, facial and forelimb clonus; however, “myoclonic jerk” was the continuous myoclonic jerk tail rigidity; that is, Straube tail [9, 10]. One-way analysis of variance (ANOVA) was performed to compare the amplitude ratio (%) of the 19–21 Hz frequencies of the sham, PTZ, ESM-200, ESM-400, ESM-600, VPA-200, VPA-300, and VPA-400 groups, followed by the Fisher Least Significant Difference (LSD) post hoc test. Eight mice were then randomly selected for 30 min of behavioral recordings. The correlation of amplitude ratio (%) of the 19–21 Hz frequencies was analyzed for different types of epileptic behaviors. Pearson correlations were calculated between brain waves (EEG)-induced voltage and epileptic behavior-induced coil voltage (No. 3 [myoclonic jerks] and No. 4 [generalized clonic seizures with a kangaroo posture]) for different frequencies of seizure activity (all frequencies [0–50 Hz], alpha [8–12 Hz], delta [1–4 Hz], theta [4–8 Hz], and beta [12–30 Hz]).

Many behavioral scoring systems and indices can effectively assess epileptic behaviors in animal models [6, 8, 9, 10, 12, 13, 23, 24, 25, 26]. For example, some studies have measured the number of seizures and the duration of epileptic behaviors to indicate seizure activity after PTZ administration [6, 8, 23, 25]. Such assessments are often used to determine seizure activity, but they are rough estimations and cannot delineate different types of epileptic behaviors. The application of Itoh’s convulsion scale to animal models also sums all scores for different epileptic behaviors to confirm the seizure strength of PTZ-induced epileptic behaviors [9, 13, 27]. Racine’s convulsion scale [12], Malhotra’s convulsion scale [26], and Rathor’s convulsion scale [24] have used similar concepts to define different categories of epileptic behaviors and seizure strength. These scales provide a more detailed characterization and dissociation of different types of epileptic behaviors, but scoring these different types of epileptic behaviors is more complicated.

The present results indicated that 19–21 Hz seizure activity was positively correlated with generalized clonic seizures with a kangaroo posture (no. 4), reflected by Itoh’s convulsion scale, after PTZ administration. The time analysis indicated that 19–21 Hz seizure activity occurred with a delay after PTZ administration. No studies of which we are aware have reported correlations between epileptic behavior and 19–21 Hz seizure activity. Our findings indicated that 19–21 Hz seizure activity, assessed by the coil method, might be a novel behavioral marker of epileptic behaviors.

The present study sought to determine the brain mechanisms that are involved in 19–21 Hz seizure activity. The coil method detected 19–21 Hz seizure activity after PTZ administration. Pentylenetetrazol is a GABA${}_A$ receptor antagonist. Another GABA${}_A$/${}_C$ receptor antagonist, PTX, exerted a similar pattern of the amplitude ratio (%) for all frequencies of seizure activity. However, this pattern of the amplitude ratio (%) after PTZ administration was different from the nonselective muscarinic receptor agonist PIL, particularly in the 19–21 Hz frequency range. VPA reduced 19–21 Hz seizure activity. Overall, these findings suggest that the GABA${}_A$ receptor is just one of the mechanisms probably involved, but other mechanisms may be involved in the pathogenesis of the 19–21 Hz seizure activity [28, 29]. However, the cholinergic system was not involved in the 19–21 Hz seizure activity.

Anticonvulsant drugs have two main mechanisms of action for treating epileptic behavior [30, 31]. One mechanism involves GABAergic neurotransmission. For example, VPA has been shown to increase GABAergic neurotransmission and reduce PTZ-induced seizures in animal models. Valproic acid also suppresses patients’ partial seizures and generalized seizures [32]. Another mechanism involves the blockade of neuronal ion channels. For example, ESM has decreased T-type calcium channel currents in the thalamus to induce seizure activity [33]. Ethosuximide has been reported to ameliorate symptoms of generalized seizures in epilepsy patients [32]. Previous studies have suggested that ESM and VPA might share a similar mechanism of action that decreases the threshold of T-type calcium channel currents to reduce epileptic seizures [33, 34]. The mechanism of action of ESM and VPA might partially overlap to treat symptoms of epileptic seizures. A previous human study of childhood absence seizures suggested that ESM and VPA were more effective than lamotrigine to ameliorate symptoms. Moreover, ESM had fewer adverse effects on attention function [35]. Thus, many mechanisms of action appear to mediate symptoms of epileptic seizures. The present study indicated that the highest dose of ESM (600 mg/kg) and VPA (400 mg/kg) decreased 19–21 Hz seizure activity that was induced by PTZ. The present results are consistent with a previous study that showed that ESM and VPA reduced other epileptic seizure behaviors, such as PTZ-induced clonic seizures [36]. On the other hand, the present result showed that the low dose of VPA (200 mg/kg) facilitated 19–20 Hz seizure activity induced by PTZ. Obviously, this evidence conflicts with the data of the highest dose of VPA (400 mg/kg). In particular, the present results showed that the low dose of VPA enhanced PTZ-induced seizure activity. The data were consistent with the previous clinical findings [37, 38]. For example, the EEG study appeared that VPA administrations facilitated absence epilepsy [37]. Moreover, fewer patients with absence seizure increased the frequency of seizure activity after VPA treatments [38]. Therefore, some cases with VPA administrations revealed opposite results. Therefore, it is a crucial issue that why the different doses of the VPA administration show opposite results for affecting PTZ-induced seizure activity? Does VPA appear to have a dose-dependent effect on PTZ-induced clonic seizures? This emerging issue should be scrutinized in further studies.

The specific brain waves that are associated with epileptic behaviors remain unknown. For example, patients with temporal lobe seizures exhibit 1–2 Hz slow-wave delta activity in bilateral frontal and parietal cortices. Moreover, 1–2 Hz slow-wave delta activity in bilaterally parietal cortices was correlated with temporal lobe seizures [39]. However, Jan et al. [40], suggested a different brain wave pattern in temporal lobe epilepsy patients. These authors suggested the occurrence of irregular 2–5 Hz lateralized activity, 5–10 Hz sinusoidal waves, and continuous epileptic discharges in start-stop-start ictal rhythms in temporal lobe epilepsy patients [40]. Extra-temporal seizures (including in frontal, central, parietal, occipital, and middle brain areas) have been shown to be associated with diverse types of brain waves. These brain waves and frequencies include beta, alpha, theta, and delta frequency ranges [41]. A focal beta frequency discharge was linked to the early stage of frontal lobe epilepsy [42]. Serafini et al. [43], suggested that ~20 Hz frequency brain waves (rhythmic EEG spike potentials) were associated with myoclonic jerks, which is consistent with the present results. In the present study, 19–21 Hz seizure activity was positively associated with 20-30 Hz high-frequency beta waves that were related to myoclonic jerks (no. 3) and generalized clonic seizures with a kangaroo posture (no. 4). The relationship between specific brain waves and epileptic behaviors needs to be further clarified.

The present study compared the coil method and force transducer method with regard to their ability to detect epileptic seizures. The coil method utilizes a magnetometer coil to detect small and fine behavioral responses, based on changes in electromagnetic fields. The coil method requires that a magnet be attached to the mice. When the animal’s movement causes the induction of a magnetic field, the coil device can detect voltage changes, and then these signals are transformed by an amplifier. The coil method provides high temporal and spatial resolution in real time and, consequently, detects smaller and finer movements [15, 18, 19, 44]. In the present study, the coil method detected 19–21 Hz seizure activity after PTZ administration. The force transducer method did not detect 19–21 Hz seizure activity. These findings are consistent with previous studies [15, 18, 19, 44]. Overall, the coil method detected a novel behavioral marker of epileptic seizures that could have clinical implications.

The present study developed a coil method to detect seizure activity in 19–21 Hz frequencies. The 19–21 Hz seizure activity is a critical behavioral marker for detecting the occurrence of seizure behavior. Some limitations and issues should be addressed in further studies. For example, the magnet size should be kept in miniature, and the magnet weight was required to be minimum to avoid interfering with the animal’s behaviors. The diameter of the coil can be increased to allow more space for the animal to move. But it may reduce the sensitivity of the magnetic field detection. There is the possibility that other behavioral frequencies are associated with different types of seizure behaviors. The coil method may detect another early behavioral symptom in neurodegeneration diseases or neurological disorders such as Parkinson’s disease or Alzheimer’s disease. Thus, further studies need to investigate how to apply the coil method in detecting other neurological diseases.

On the other hand, the present result showed that the low dose of VPA (200 mg/kg) enhanced PTZ-induced 19–20 Hz seizure activity compared to the PTZ group. The result was opposite to the anticipatory effect. The reason may be that VPA activated various mechanisms but not the specific mechanism of the GABA${}_A$ receptor. For example, these mechanisms included several voltage-operated channels, GABA clearing enzymes and GABA transporters, the IP3 (inositol triphosphate) pathway, etc. [29]. Therefore, the crucial issue has emerged: one of these mechanisms in the lower dose of VPA facilitated PTZ-induced 19–20 Hz seizure activity. It should be investigated in further studies.