Hidalgo Monroy Lerma, Bruno

Abstract Objective Resective surgery achieves seizure freedom in approximately 60%–80% of patients with drug‐resistant temporal lobe epilepsy (TLE), yet the extent to which seizure cessation permits recovery of disrupted functional brain networks remains unclear. Increasing evidence suggests that recurrent seizures alter subcortical–cortical communication involved in arousal and cognition, contributing to chronic interictal deficits, such as in executive function. Prior work demonstrates that the nucleus basalis of Meynert (NBM), the primary source of cortical cholinergic input to the cortex, exhibits aberrant functional connectivity in patients with TLE. We examined NBM connectivity in patients with TLE before and after surgery using patient‐specific NBM segmentations. We sought to investigate whether seizure freedom is associated with recovery of NBM functional connectivity, consistent with previous observations of interictal network dysfunction reversal in other subcortical arousal structures. Methods Resting‐state functional magnetic resonance imaging was analyzed in 75 patients with TLE, including 34 patients postoperatively, and 106 controls. Patient‐specific NBM segmentations were used to compute functional connectivity between the NBM and frontoparietal association cortex (FPAC) and across the whole brain. Subcortical arousal network community structure was assessed using community detection. Postoperative network changes were evaluated relative to seizure outcome. Results Prior to surgery, patients exhibited reduced bilateral NBM connectivity to the FPAC and whole brain compared to controls, indicating widespread cholinergic network disruption beyond the ipsilateral hemisphere. Following surgery, NBM connectivity increased and no longer differed from controls. Seizure‐free patients demonstrated greater postoperative increases in ipsilateral NBM‐FPAC connectivity than non‐seizure‐free patients, which was not explained by time to postoperative scan or volumetric change. Subcortical arousal community structure remained altered postoperatively. Significance Successful TLE surgery is associated with partial restoration of cholinergic arousal network connectivity, supporting the idea that seizure cessation enables recovery of brain networks disrupted by recurrent seizures. These findings position the NBM as a potential biomarker of network recovery following epilepsy surgery.

BACKGROUND In patients with drug-resistant epilepsy (DRE) who are not candidates for resective surgery, thalamic neuromodulation targeting structures such as the anterior nucleus of the thalamus (ANT), centromedian nucleus (CM), and pulvinar (PUL) has emerged as a promising therapy. The clinical investigation of these targets’ efficacy and side effect profiles is ongoing, complicating discussions between neurosurgeons, neurologists, and patients. The implementation of single-pulse and continuous stimulation protocols to conduct patient-specific side effect screens and generate prognostic biomarkers could offer guidance for target selection. OBSERVATIONS A 32-year-old male with extensive polymicrogyria and DRE underwent stereo-electroencephalography (SEEG) with multisite thalamic sampling to inform neuromodulation target selection. Single-pulse stimulation was conducted to screen for acute sensorimotor effects, followed by continuous (12- to 24-hour) stimulation using clinically relevant parameters. Intolerable paresthesia occurred during CM single-pulse stimulation, leading to its exclusion as a target. Single-pulse and continuous stimulation of the ANT and PUL were well tolerated. PUL stimulation offered superior suppression of epileptiform discharges during sleep. LESSONS These findings meaningfully informed the selection of the PUL as a neuromodulation target. This case illustrates the safety, feasibility, and utility of patient-specific single-pulse and continuous stimulation assessments during SEEG monitoring to evaluate efficacy and tolerability prior to chronic neuromodulation, particularly for novel thalamic targets. https://thejns.org/doi/10.3171/CASE25947

Neuromodulation is becoming a common treatment strategy for patients with drug-resistant epilepsy who are not candidates for resective surgery. Responsive neurostimulation can effectively reduce seizure burden, however it remains unclear which neural signals are most important for triggered stimulation. Furthermore, the high-dimensional parameter space remains systematically underexamined. As more patients become candidates for neurostimulation, clinical teams will need to decide on the optimal site and stimulation program for therapeutic neuromodulation. In this review, we highlight how recent insight into the biological rhythms of epilepsy may pair with personalization protocols in the emerging field of neuropsychiatric neuromodulation. Finally, we discuss future directions for epilepsy neuromodulation, specifically how machine learning can offer clinical teams actionable feedback on neuromodulation efficacy.

AbstractObjectiveEpilepsy is a debilitating disorder affecting more than 50 million people worldwide, and one third of patients continue to have seizures despite maximal medical management. If patients' seizures localize to a discrete brain region, termed a seizure onset zone, resection may be curative. Localization is often confirmed with stereotactic electroencephalography; however, this may require patients to stay in the hospital for weeks to capture spontaneous seizures. Automated localization of seizure onset zones could therefore improve presurgical evaluation and decrease morbidity.MethodsUsing more than 1 000 000 interictal stereotactic electroencephalography segments collected from 78 patients, we performed five‐fold cross‐validation and testing on a multichannel, multiscale, one‐dimensional convolutional neural network to classify seizure onset zones.ResultsAcross held‐out test sets, our models achieved a seizure onset zone classification sensitivity of .702 (95% confidence interval [CI] = .549–.805), specificity of .741 (95% CI = .652–.835), and accuracy of .738 (95% CI = .687–.795), which was significantly better than models trained on random labels. The models performed well across the entire brain, with top five region performance demonstrating accuracies between 70.0% and 88.4%. When split by outcomes, the models performed significantly better on patients with favorable Engel outcomes after resection or who were responsive neurostimulation responders. Finally, SHAP (Shapley Additive Explanation) value analysis on median‐normalized input data assigned consistently high feature importance to interictal spikes and large deflections, whereas similar analyses on histogram‐equalized data revealed differences in feature importance assignments to low‐amplitude segments.SignificanceThis work serves as evidence that deep learning on brief interictal intracranial data can classify seizure onset zones across the brain. Furthermore, our findings corroborate current understandings of interictal epileptiform discharges and may help uncover novel interictal morphologies. Clinical application of our models may reduce dependence on recorded seizures for localization and shorten presurgical evaluation time for drug‐resistant epilepsy patients, reducing patient morbidity and hospital costs.