Clinical vignette demonstrates seizure recurrence after disconnection in a young girl, implying challenges in EEG interpretation.
Seizure recurrence after hemispheric disconnection requires careful evaluation to distinguish seizures due to incomplete disconnection in the operated hemisphere from independent contralateral hemisphere epileptogenicity.1-4 Interpretation of scalp EEG in this postoperative setting poses unique challenges and may be misleading.5, 6 We present a clinical vignette of Rasmussen encephalitis demonstrating false lateralization on EEG after functional hemispherectomy, with seizure freedom achieved following conversion to anatomic hemispherectomy. A 10-year-old right-handed girl developed epilepsy at age 7 years. Initial seizures consisted of brief left facial and arm clonic jerking. Development and neurological examination were normal. EEG demonstrated right frontotemporal epileptiform discharges (Figure S1A), and MRI revealed focal atrophy around the right inferior precentral sulcus (Figure S2A). Over the next 3 years, seizure burden increased with emergence of atypical absences, drop attacks, and bilateral tonic–clonic seizures, occasionally preceded by left facial clonic activity. Interictal and ictal EEG showed generalized epileptiform discharges, at times preceded by frontocentral discharges (Figure S1B–D). Serial MRI brain demonstrated progressive right hemispheric atrophy (Figure S2B), consistent with Rasmussen encephalitis.7 Brain FDG-PET showed concordant focal hypometabolism (Figure S2C). Autoimmune and genetic testing were unrevealing. Seizures remained refractory despite trails of several antiseizure medications and immunomodulatory therapy with intravenous immunoglobulin for 2 years. At age 11, the patient underwent a functional right hemispherectomy, as the family elected to accept the expected hemiplegia in hopes of seizure freedom. Histopathology showed chronic inflammatory infiltrates, microglial nodules, and gliosis consistent with Rasmussen encephalitis. Four months after surgery, seizures recurred. Interictal EEG showed a higher burden of epileptiform discharges in the left parasagittal region (60–65%) than the right hemisphere (35–40%) (Figure 1A,B). Bilateral asymmetric tonic seizures, with flexion of the paretic left arm and extension of the right arm, were recorded. Ictal EEG showed robust left hemispheric patterns (Figure 1C). Several focal impaired consciousness seizures with left frontocentral ictal patterns were also recorded (Figure 1D). On magnetoencephalogram, interictal and ictal dipoles were localized to the left hemisphere (Figure S3A). Postoperative brain MRI, including diffusion tensor imaging, showed no residual connections (Figure S3B). Given persistent seizures and the known limitations of EEG lateralization after hemispheric disconnection, a right anatomic hemispherectomy was performed. The patient achieved sustained seizure freedom with resolution of contralateral epileptiform activity (Figure S4). Antiseizure medications were successfully discontinued gradually 1 year after surgery, and she has remained seizure-free at the last follow up, 5 years after surgery. Our case illustrates two critical interpretive pitfalls following functional hemispherectomy. First, ictal and interictal EEG lateralization to the contralateral hemisphere does not reliably indicate independent contralateral epileptogenicity. Second, absence of an overt residual connection on postoperative MRI does not exclude incomplete disconnection. Seizure freedom after anatomic hemispherectomy in our patient confirms that there was a residual connection between the hemispheres and the preoperative EEG findings were false lateralizing. Following disconnective surgery, seizures arising from the operated hemisphere may generate insufficient synchronized cortical activity for scalp EEG detection until propagation to the contralateral hemisphere, potentially yielding misleading lateralization. The presence of abundant generalized abnormalities prior to functional hemispherectomy in our patient may be a factor for such paradoxical patterns after hemispheric disconnection. Similar false lateralization of scalp EEG in hemispherectomy candidates with extensive lesions with volume loss has been described.5 Conversely, interictal epileptiform discharges in the disconnected hemisphere are common and in seizure-free patients these findings are clinically inconsequential.6 MRI remains essential for detecting residual connections, commonly involving the genu or splenium of the corpus callosum, temporal stem, insula, and frontobasal regions.2 Successful seizure freedom in our patient after anatomical hemispherectomy shows that the absence of residual connections on postoperative MRI should not be interpreted as definitive evidence of complete disconnection. Reoperative hemispherectomy has been shown to result in seizure freedom in 20–33% of patients and substantial improvement in many others.3, 4, 8, 9 In one series of Rasmussen encephalitis, all five patients who needed reoperation after initial functional hemispherectomy achieved seizure freedom.4 Absence of structural abnormalities in the left hemisphere and rarity of true bilateral disease in Rasmussen encephalitis and high seizure burden in our patient prompted us to proceed with anatomic hemispherectomy. This case underscores the need for cautious interpretation of postoperative EEG and MRI findings after functional hemispherectomy. Recurrent seizures after hemispheric disconnection warrant consideration of an incomplete disconnection even in the absence of supportive evidence on MRI and EEG. The authors declare no conflict of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request. Figure S1. (A–D). Interictal and ictal EEG findings before hemispheric disconnection. (A) Interictal EEG at age 7 years showing focal epileptiform discharges over the right fronto-centro-temporal regions (black arrow). (B) Interictal EEG at age 10 years demonstrating frequent bilateral frontocentral spikes and generalized spike–wave complexes. (C) and (D): Ictal EEG epochs recorded at age 10 years, illustrating distinct seizure types. Generalized ictal pattern with low amplitude fast activity associated with a generalized tonic seizure preceded by a myoclonic jerk (C). Note the bilateral parasagittal spikes preceding the seizure, resembling the interictal discharges.: Generalized spike discharges followed by brief attenuation, corresponding with whole-body myoclonic jerks (D). Data S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. A routine EEG done in an 8-year-old girl who has been seizure free for 2 years after hemispheric disconnection for Rasmussen encephalitis shows frequent repetitive epileptiform discharges in the operated hemisphere. Which of the following is the right next step? A 12-year-old boy began having recurrent seizures 4 months after right disconnective hemispherectomy for Rasmussen encephalitis. Video EEG showed ictal patterns maximum over the left hemisphere. Which of the following is the right next step? Which of the following is not a common feature of Rasmussen encephalitis? Answers may be found in the Supporting Information.
No takes yet. Share an insight, caveat, or question.
Natteru et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: