A 2-month-old male domestic shorthair cat was presented to a local veterinary clinic with a complaint of recurrent generalized convulsions. The local veterinarian noticed the cat also had frequent facial twitching, ataxia, visual impairment, and slightly decreased mentation at the first consultation. Video clips recorded by the owner (Supplementary Video 1) showed generalized myoclonic seizures, which was momentary but clustered for 1 to a few minutes, and GTCS lasting about several tens of seconds to 1 minute. These seizures were initially observed several times per week; therefore, antiseizure drug (ASD) therapy with 1.0 mg/kg of phenobarbital (PB) every 12 h per os was initiated. The dose of PB was increased to 2.0 mg/kg q12h PO (achieved static serum level: 14.8 μg/ml) at 3 months of age due to insufficient seizure control. At this time, complete blood count, serum chemistry including fasted and postprandial ammonia and total bile acid, urinalysis, screening test of feline immunodeficiency virus and feline leukemia virus, cerebrospinal fluid (CSF) analysis including PCR for Toxoplasma gondii and feline corona virus, and low-field (0.4T) magnetic resonance imaging (MRI) of the brain were performed. All complementary tests were unremarkable. Due to severely increased seizure frequency to > 40 times daily (uncountable), the dose of PB was further increased (4.0 mg/kg q12 h PO), and zonisamide (8.0 mg/kg q12h PO), diazepam (DZP; 0.5 mg/kg q12 h PO) as well as levetiracetam (LEV; 20 mg/kg q8 h PO) were added. However, zonisamide was discontinued 3 days after the initiation due to the appearance of aphagia.

At 5 months of age, the cat was referred to a veterinary neurologist (SK) at the Veterinary Specialists Emergency Center (Saitama, Japan). Neurological examination revealed severe mental dullness, disorientation, difficulty in standing and walking, decreased postural reactions in all limbs, and bilateral menace response deficits with normal pupillary light reflexes. Appetite remained normal but the cat needed assisted feedings. Levetiracetam had been discontinued by the owner because the cat seemed to develop stupor after its administration. Doses of PB and DZP were maintained at 5.0 mg/kg q12 h PO (achieved static serum level: 27.1 μg/ml) and 0.8 mg/kg q12 h PO, respectively. Although the frequency and duration of GTCS decreased, myoclonic seizures and short tonic seizures still occurred 1–6 times per day (sz/d). Although other differential diagnoses at this point including inborn errors of metabolism, inherited neurodegenerative conditions, or immune-mediated limbic encephalitis could not be completely ruled out, further diagnostics such as the Tandem mass spectrometry and/or investigation of known antibodies such as anti-voltage-gated potassium channel and/or anti-leucine-rich glioma inactivated 1 receptor antibodies for limbic encephalitis were unavailable locally then. At 7 months of age, the cat was presumptively diagnosed as DRE of unknown cause with severely impaired quality of life (QOL), when the potential candidacy of the epilepsy surgery research project was discussed and the owner requested the candidacy to be evaluated.

The project team of veterinary epilepsy surgery (Supplementary Data 1.1) discussed this case based on the abovementioned history and data. All members considered this cat as a candidate for epilepsy surgery. Electroencephalography (EEG) and high-field MRI as presurgical evaluations were performed to determine the type of surgical procedure.

At 8 months of age, the cat was further referred to the Veterinary Medical Teaching Hospital of Nippon Veterinary and Life Science University (Tokyo, Japan) to undergo EEG and high-field MRI. At this point, seizure frequency increased reaching up to 7–10 sz/d, including GTCS. The cat was non-ambulatory and nearly stuporous but remained responsive to forced feeding, loud sounds, and nociceptive stimulation. Postural reaction deficits and bilateral menace deficits were noticed on neurological examination; no other abnormalities were recorded. Feeding, drinking, and urination had been fully assisted by the owner. Complete blood count and serum chemistry profile were re-evaluated and revealed mild dehydration, which was easily improved later by increasing the water intake. Scalp EEG under sedation (Supplementary Data 1.2) revealed frequent spikes and multiple spike complexes were observed on the central regions (C3, C4, Cz) dominantly but synchronized generally (Figure 1A). High-field MRI (3.0T) under general anesthesia (Supplementary Data 1.3) showed no appreciable structural abnormalities in the whole brain (Figure 2A). Cerebrospinal fluid analysis including cell count (0/μl; ref 0–5/μl), cytology (no cells observed), protein concentration (12.4 mg/dl; ref 0–25 mg/dl), glucose (126 mg/dl; ref 35–140 mg/dl), and electrolytes (Na⁺ 158; K⁺ 2.9; Cl⁻ 135 mEq/l) was also unremarkable. The serum level of PB at this point was 33 μg/ml. With these repeated MRI and CSF analysis, inherited neurodegenerative conditions were considered unlikely. Although other differentials could not be completely ruled out, DRE of unknown etiology was considered the most likely.

Based on these results, the epileptogenic zone of this case was not determined; therefore, the project team selected CCT as a suitable surgical procedure based on the following reasons: (1) epileptiform discharges distributed bilaterally; (2) unremarkable MRI; (3) severe seizure frequency that might result in decreased mental status; and (4) decreased QOL of the patient and the owner (Supplementary Data 1.1). As the final checkpoint to detect an identifiable epileptogenic focus, an intraoperative electrocorticography (ECoG) recording was planned immediately following the craniotomy. If the intraoperative ECoG recording clearly identified the epileptogenic focus, a different procedure such as a focal resection and/or multiple subpial transections would have been considered more suitable (Supplementary Data 1.1).

The owner was requested to record seizures including the types, frequency, and duration, as well as the visual analog scales (VAS; Supplementary Data 2) before and 3, 6, and 12 months after surgery. An attending neurologist (DH) evaluated the activities of daily living (ADL) (Supplementary Data 3) and performed follow-up EEG and MRI during the same period.

Corpus callosotomy (CCT) was performed at 12 months of age. General anesthesia was induced by propofol (6 mg/kg IV) and maintained with inhalation of isoflurane and oxygen. The cat was positioned in sternal recumbency in a Sphinx-like position. During surgery, electrocardiogram, oxygen saturation, end-tidal CO₂ concentration, indirect blood pressure, rectal temperature, and urine production volume were monitored. Lactated Ringer's fluid was infused at 3–5 ml/kg/h, and a constant rate infusion (CRI) of remifentanil as intraoperative analgesia was performed at 20 μg/kg/h. After shaving and sterilization, an H-shaped scalp incision was made with the skin opened like a double door. Bilateral temporal muscles were incised along to the external sagittal crest and detached from the parietal bones, and then retracted bilaterally. A 3 cm (left–right) × 4 cm (rostral–caudal) rectangular-shaped craniotomy straddling the left and right of the parietal bone was performed. The first four burr holes were made on each corner of the rectangle with the rostral pair at the level of the caudal end of the frontal bone and the caudal pair approximately 1.5 cm rostral to the external occipital protuberance. The craniotomy was completed by connecting these burr holes to expose the underlying cerebral surface of the parietal lobes. In the process of connecting the burr holes, the craniotomy lines connecting the two rostral and the two caudal burr holes, i.e., transverse craniotomy lines, had widened areas across the midline in order to visualize the underlying dorsal sagittal sinus well. The removed bone fragment was wrapped with saline-filled gauze and saved for the closure.

After the craniotomy, epidural ECoG was recorded. A 1.5 × 1.5 cm grid-type silicon-sheet electrode with nine exploration electrodes was placed on the dura covering each hemisphere's parietal lobe. On ECoG, bilaterally synchronized spike-containing burst-suppression was recorded (Supplementary Data 1.4).

As ECoG did not aid to identify the epileptogenic zone, a sagittal sinus-based U-shaped dural incision was made on the left hemisphere and inverted to the right side. The left hemisphere was gently retracted laterally with neurosurgical putties, spatula, and shafts of suction to widen the longitudinal fissure. By peeling off the cingulate gyrus on both sides under the falx gently, the bright white CC was exposed (Supplementary Video 2).

The CC was bisected using bipolar cautery and suction from the caudal, the splenium, to the rostral, the genu. During the bisecting of the CC, the great cerebral vein, where the confluence of the vein of the corpus callosum and internal cerebral vein, was observed at the caudal end of the splenium. The third ventricle was also observed under the body. During the approach of the genu, a cortical vessel on the cingulate gyral surface was damaged, and the hemostasis manipulation and the subsequent brain swelling made it difficult to identify the genu. Therefore, the rostral dissection was abandoned (Supplementary Video 2). Due to the brain swelling, 1 g/kg of mannitol was intravenously administered over 15 minutes.

The dura was sutured with 6-0 polydioxanone after confirming normalized brain volume and no bleeding from the intracranial surgical field. Then, a post-callosotomy intraoperative ECoG was recorded; Although large spikes were still synchronized on both hemispheres, small independent spikes, not seen pre-CCT, were observed from each hemisphere (Supplementary Data 1.3). After recording ECoG, the stored bone fragment was returned and fixed with polymethylmethacrylate. Temporal muscles and scalp were sutured routinely. A percutaneous endoscopic gastrostomy (PEG) tube was placed to facilitate postoperative feeding.

The cat recovered from anesthesia uneventfully, and the patient's conscious level had moderately improved compared with the preoperative state. However, facial twitching was observed intermittently until the next day. As postoperative analgesia, CRI of fentanyl (2 μg/kg/h to tapered) was infused for 12 h after surgery, and a fentanyl patch (4.2 mg/head) was applied for 3 days. Although cefmetazole (25 mg/kg q2–12h IV) had been used from intra- to postoperative 12 h, the cat showed fever (~40°C) with mild neutrophilia (WBC 22,000 /μl) on the next day. Because bacterial meningitis was a potential concern, meropenem (25 mg/kg, CRI for 30 min q24h) for 3 days and enrofloxacin (5 mg/kg q24h PO) for 5 days were administered. The fever ceased within 4 h after initiation of meropenem, after that fever and leukocytosis had never been observed remaining the cause being unidentified.

The day after surgery, the cat was able to maintain a sitting position and eat wet food on his own, with the nutritional requirement maintained using a PEG tube. Furthermore, on the postoperative day (POD) 3, even though mild right-sided hemiparesis was observed, the cat became ambulatory and was able to voluntarily urinate and defecate. The PEG tube was removed at POD 10, and the cat was discharged from the hospital at POD 16. During the postoperative hospitalization, 5 myoclonic seizures were observed, while there were 12 seizure-free days. Neurological findings at the discharge time were slightly somnolence, mild disorientation, ambulate but generalized ataxia with postural reaction deficits in all limbs, and loss of biliteral menace response. Phenobarbital (3.5 mg/kg q12 h PO) and DZP (0.5 mg/kg q12 h PO) were continued throughout the hospitalization.

After discharge, the cat was routinely consulted at a local clinic every 2–4 weeks, and presented to the teaching hospital at 3, 6, and, 12 months after surgery for rechecks and follow-up EEG and MRI. Antiseizure drug therapy with PB and DZP were maintained for the 1-year follow-up period. The neurological status, mild cognitive impairment, ataxia with decreased postural reactions, loss of bilateral menace response, remained unchanged for the 1-year follow-up period. Mental status fluctuated from alert to somnolent depending on the time of the visit or seizure frequency. Changes of seizure types were observed postoperatively: GTCS had completely disappeared, and generalized myoclonic seizures changed to momentary facial myoclonic seizures and shorter tonic seizures. On the other hand, a new seizure type manifesting as the initial left arm-extended fencing posture (fencing posture seizure; FPS) followed by contralateral (right) myoclonic or clonic seizures (Supplementary Video 3) emerged 7 months postoperatively. Sequential changes in seizure frequency, VAS, and ADL from the preoperative time to the 1-year follow-up are shown in Figure 3, Table 1, and Supplementary Data 3, respectively. Seizure frequencies of all seizure types were markedly decreased (7.2–1.7 sz/d; 76% reduction); however, 1–2 sz/d in average either myoclonic seizures or FPS, or both remained unchanged through the 1-year follow-up period.

Follow-up scalp EEGs showed clearly decreased spike frequency. Although a few spikes were observed in a hemisphere only, most were still synchronized bilaterally (Figures 1B–D). Follow-up MRI at 3, 6, and 12 months after surgery revealed the bisected CC with the remaining genu, loss of caudal part of the left cingulate gyrus, and old hemorrhagic and a necrotic lesion in the left cingulate gyrus likely related to intraoperative tissue handling as well as vascular damage (Figure 2B).