Seizures Before Surgery Linked to Sharper Processing Speed in Glioma Patients

When a brain tumor announces itself with a seizure, clinicians have long treated that event as a warning sign of aggressive disease. A new study from Taiwan turns that intuition partly on its head. Researchers examining patients with newly diagnosed glioma found that those whose tumors first declared themselves through epileptic seizures performed significantly better on a key test of processing speed than patients who arrived without seizures, and the advantage persisted even after the researchers statistically accounted for age, education, tumor size, and the molecular profile of the cancer. The finding, published in the Journal of Neuro-Oncology, suggests that tumor-related epilepsy may carry information about the tumor-brain environment that standard clinical measures fail to capture.

The study, led by Fang-Tzu Chuang of Asia University and Ko-Ting Chen of Chang Gung Memorial Hospital, together with colleagues at several Taiwanese institutions, enrolled sixty-five patients who had been newly diagnosed with glioma and had not yet received any treatment. Thirty of them had experienced tumor-related epilepsy, meaning seizures caused directly by the tumor, while thirty-five had not. This treatment-naive design matters enormously for interpreting the results. Once patients undergo surgery, radiation, or chemotherapy, and once they begin taking antiseizure medications, it becomes nearly impossible to disentangle the effects of the tumor itself from the effects of the interventions. By testing cognition before any treatment began, the researchers obtained a snapshot of what the tumor and the brain’s response to it were doing on their own.

The neuropsychological battery was comprehensive. Patients completed measures of attention, processing speed, executive function, memory, language, and visuospatial ability, including the Trail Making Test, the Wechsler Adult Intelligence Scale, the Benton Visual Retention Test, and Nelson’s Modified Card Sorting Test, among others. Crucially, rather than comparing raw scores, the team converted each patient’s performance into percentile ranks adjusted for demographic factors such as age and education. This approach allows a fair comparison between, say, a twenty-five-year-old university graduate and a sixty-year-old patient with fewer years of schooling, both of whom may be functioning normally or abnormally relative to their own demographic peers.

Before any statistical correction, the picture looked striking. Patients with tumor-related epilepsy were younger, more highly educated, had smaller enhancing tumor volumes on imaging, were more likely to have lower-grade tumors, and carried isocitrate dehydrogenase, or IDH, mutations far more often. Each of these characteristics is independently associated with better cognitive outcomes in the glioma literature. IDH-mutant gliomas, in particular, are known to behave more indolently and to be associated with preserved neurological function, and prior work has linked the mutation to a higher incidence of preoperative seizures, possibly through the accumulation of the oncometabolite D-2-hydroxyglutarate, which can promote hyperexcitability in surrounding neural tissue. On several cognitive measures, the seizure group simply outperformed the non-seizure group.

But raw group differences can be deceptive when the groups differ in so many other ways. To address this, the researchers applied the Benjamini-Hochberg false discovery rate correction, a statistical procedure designed to limit the number of spurious positive findings when many tests are run simultaneously. After this correction, only one measure survived: Part A of the Trail Making Test, a timed task in which the patient connects numbered circles in sequence. The test is a sensitive index of visual scanning, attention, and psychomotor speed. Patients with tumor-related epilepsy scored at a median percentile of 83, meaning they performed better than roughly eighty-three percent of demographically matched healthy individuals, while patients without seizures scored at a median percentile of 21. The effect size, expressed as r = .41, is moderate to large by conventional standards, and the corrected significance value was q = 0.04.

The more demanding question was whether the seizure-cognition link would survive adjustment for the confounding variables. In a multivariable regression model that included age and enhancing tumor volume, seizure presentation remained significantly associated with better TMT-A performance, with a standardized coefficient of B = 0.276 and a p-value of .017. In sensitivity analyses that additionally incorporated IDH mutation status and other demographic and clinicopathological characteristics, the association held firm. This persistence is the study’s central message. If the apparent cognitive advantage of seizure patients were merely a proxy for youth, education, small tumors, low grade, or favorable molecular profile, it should have dissolved once those variables entered the model. Instead, it endured, implying that the seizure itself, or the biology that produces it, marks something clinically meaningful about how the tumor interacts with the brain.

What could that something be? The authors and the broader literature point toward several candidate mechanisms. One possibility relates to tumor location and growth pattern. Slow-growing, diffuse tumors that infiltrate functional cortex gradually may be more likely to irritate neural circuits into seizure activity while leaving cognitive networks relatively intact, whereas rapidly expanding masses in non-eloquent regions may reach considerable size silently, compressing and disrupting tissue without ever triggering epileptiform discharges. Work on glioblastoma has shown that patients presenting with seizures differ microstructurally and volumetrically from those without, and recent research has even linked seizure presentation in IDH-wildtype glioblastoma to the upregulation of a synaptic gene signature, hinting that epileptogenic tumors may inhabit a different biological niche than their non-epileptogenic counterparts.

The metabolic dimension adds further intrigue. IDH-mutant gliomas produce D-2-hydroxyglutarate, an oncometabolite detectable by magnetic resonance spectroscopy, and elevated levels of this molecule have been associated with preoperative seizures. Animal and cellular studies suggest that D-2-hydroxyglutarate promotes epileptogenesis through hyperactivation of the mTOR signaling pathway, which regulates synaptic plasticity. In this framework, the very biochemistry that makes a glioma epileptogenic also tends to make it slower-growing and less aggressive, which could explain why seizure patients arrive at the clinic with better-preserved processing speed. The seizure, in this reading, is not a cause of better cognition but a visible symptom of a tumor biology that is, paradoxically, gentler to the mind.

The clinical implications are worth spelling out. Preoperative cognitive performance is one of the strongest predictors of quality of life, employment status, and even survival in glioma patients; studies have shown that initial cognitive impairment predicts shorter survival in glioblastoma. If seizure presentation reliably flags a distinct cognitive and biological profile, neurosurgeons and neuro-oncologists could use that information when planning the extent of resection, counseling patients about prognosis, and designing cognitive rehabilitation programs. A patient who presents with seizures and strong processing speed may tolerate more aggressive surgical strategies aimed at long-term control, whereas a patient arriving with silent tumor growth and already-impaired cognition may require a different calculus. The finding also cautions against the reflexive assumption that seizures in brain tumor patients are uniformly bad news for the brain.

The researchers are careful to frame the work as preliminary. Sixty-five patients is a modest sample, and the single-center design in Taiwan raises questions about generalizability, although the use of locally validated normative data for the neuropsychological instruments strengthens the demographic adjustments. The study also did not generate or analyze external datasets, and the authors note that the mechanisms linking epileptogenesis to cognitive preservation remain to be tested directly, perhaps through longitudinal follow-up, functional connectivity imaging, or metabolic spectroscopy. Still, the core result is clean and statistically disciplined: after correcting for multiple comparisons and adjusting for the obvious confounders, seizure presentation retained a significant association with better processing speed. That suggests tumor-related epilepsy is not just a symptom to be suppressed but a signal to be read, a clinical clue that the tumor-brain context of one patient may differ in kind, not merely in degree, from that of another. For a disease as heterogeneous and unforgiving as glioma, every such clue counts.

Subject of Research: The association between tumor-related epilepsy and preoperative cognitive performance in patients with glioma

Article Title: Tumor-related epilepsy and preoperative cognitive performance in glioma: a preliminary study

Article References: Chuang, F.-T., Chen, K.-T., Wei, K.-C., Chen, Y.-C., Wang, W.-H., Liao, Y.-C., & Chang, H.-T. (2026). Tumor-related epilepsy and preoperative cognitive performance in glioma: a preliminary study. Journal of Neuro-Oncology, 180(1), Article 11. https://doi.org/10.1007/s11060-026-05820-z

Image Credits: AI Generated

DOI: 10.1007/s11060-026-05820-z

Keywords: glioma, tumor-related epilepsy, neuropsychological assessment, Trail Making Test, IDH mutation, processing speed, cognitive function, brain tumor, epileptogenesis, D-2-hydroxyglutarate, neuro-oncology, preoperative cognition