Focal to bilateral tonic-clonic seizures (FBTCS) is a severe form of seizure associated with various adverse events. This study aimed to characterize abnormalities in the resting-state brain network related to FBTCS and use those findings to fit machine learning models for individual-level identification of patients with FBTCS. T1-weighted and resting-state functional magnetic resonance imaging (rfMRI) data were acquired from 84 patients with FBTCS (FBTCS+), 47 patients without FBTCS (FBTCS-), and 81 matched healthy controls (HCs). Amplitude of low-frequency fluctuations (ALFF), regional homogeneity (ReHo), and degree centrality (DC) were calculated across whole brain and compared among 3 groups. Brain regions with significant differences between FBTCS+ and FBTCS- groups were seeded for resting-state functional connectivity (rs-FC) analysis. Four models were employed to classify FBTCS+ from FBTCS- patients at the individual level. Compared to HCs, both FBTCS+ and FBTCS- patients exhibited diffuse alterations in ALFF, ReHo, and DC, with similar patterns but more significant and widespread in FBTCS+ patients. Direct comparison demonstrated significant increase of DC in the ipsilateral temporal pole, with rs-FC increase to the ipsilateral lingual gyrus and the contralateral temporal pole and superior temporal gyrus, in the FBTCS+ patients relative to FBTCS- patients. Using significant differences as features, four classifiers performed well to distinguish FBTCS+ patient from FBTCS- patient, achieving an average AUC of 0.76. Ipsilateral temporal pole showed increased neural activity and hyper-connection to the temporo-occipital regions in FBTCS+ patients, which provide additional insights for FBTCS and carry individual-level information for sensitive identification of FBTCS+ patient.

ObjectiveCatamenial epilepsy (CE) is a neuroendocrine disorder characterized by seizure exacerbation during specific phases of the menstrual cycle. The pathophysiology of CE remains elusive. This study investigates structural and functional brain alterations in women with Type-I CE, focusing on the interaction between hormonal fluctuations and epileptic networks.MethodsThirty-three CE women with Type-I CE and 27 healthy controls (HCs) underwent longitudinal multimodal MRI scans during the perimenstrual and midluteal phases. Voxel-based morphometry was used to detect abnormal gray matter volume (GMV), while amplitude of low-frequency fluctuation (ALFF) and fractional ALFF (fALFF) evaluated local signal changes. Functional connectivity was explored with brain regions showing abnormal GMV, fALFF, or ALFF as seed regions. Mediation analysis evaluated structural-functional relationships between hormonal changes and seizure frequency.ResultsReduced GMV was found in the CE group within the superior frontal gyrus, mediodorsal thalamus, insula, and limbic regions. Resting-state functional MRI demonstrated cycle-related fluctuations in the anterior cingulate cortex and the inferior temporal gyrus that were specific to women with CE relative to the HC group. In the seed-based FC analysis, the CE group exhibited hypoconnectivity within the temporo-limbic regions compared to HCs. In contrast to the temporal stability observed in HCs, women with CE displayed extensive hyperconnectivity involving the thalamo-striatal and fronto-parietal regions during the perimenstrual phase relative to the midluteal phase. Correlation analysis indicated that perimenstrual estradiol levels were negatively associated with ALFF value in the inferior temporal gyrus. Mediation analysis suggested mediodorsal thalamus GMV atrophy as a critical substrate linking estradiol fluctuations to perimenstrual seizure exacerbation.ConclusionsThese findings may implicate a hormone-sensitive network with widespread cortical alterations anchored in the thalamo-temporolimbic hub as central to CE pathophysiology. The structural vulnerability of the mediodorsal thalamus acts as a modulator where estradiol withdrawal might trigger functional network instability, suggesting a potential precision target for neuromodulation.

BackgroundPeople with stroke face a high mortality risk, and an accurate prediction model is essential to the guidance of clinical decision-making in this population. Recently, with growing attention paid to machine learning (ML) in stroke care, some researchers have investigated the effectiveness of ML in predicting the mortality risk in stroke. However, systematic evidence is still lacking for its effectiveness. ObjectiveThis systematic review aims to evaluate the value of ML in predicting the stroke mortality risk. The findings are expected to offer an evidence-based basis for developing and assessing clinical risk prediction tools. MethodsA search was made in Cochrane Library, PubMed, Embase, and Web of Science up to June 23, 2025, and studies that reported a complete performance of ML in predicting stroke mortality were included. Studies with only risk factors analyzed were excluded. The risk of bias of the included studies was assessed using PROBAST (Prediction model Risk of Bias Assessment Tool). Pooled risk ratios with 95% CIs and prediction intervals (PIs) were derived using the Hartung-Knapp-Sidik-Jonkman method under a random-effects model. Subgroup analyses were also conducted by model type, stroke type, patient source, and treatment background. Moreover, a metaregression was conducted on the C-index for out-of-hospital mortality at different time points to explore the influence of time factors on the model’s predictive performance. ResultsSixty-eight studies were included (23 predicting in-hospital mortality and 45 predicting out-of-hospital mortality), describing the development of 75 prediction models and 43 external validations. The follow-up period was 1 month to 15 years. For predicting in-hospital mortality, the external validation set had a pooled C-index of 0.727 (95% CI 0.677-0.781, 95% PI 0.521-1.000), with sensitivity and specificity of 0.64 (95% CI 0.57-0.70) and 0.74 (95% CI 0.70-0.77), respectively. For predicting out-of-hospital mortality, the pooled C-index was 0.847 (95% CI 0.808-0.887, 95% PI 0.750-0.956) in the external validation set, with sensitivity and specificity of 0.71 (95% CI 0.55-0.82) and 0.76 (95% CI 0.74-0.78), respectively. Comparatively, the overall pooled C-indexes were 0.788 (95% CI 0.766-0.810, 95% PI 0.621-0.999) and 0.812 (95% CI 0.798-0.826, 95% PI 0.693-0.952), respectively. The metaregression revealed a gradual decline in the predictive performance of the overall model and logistic regression model alone, whereas a random forest model maintained sustained performance. Age, National Institutes of Health Stroke Scale score, and stroke-related complications were the most frequently used variables for modeling. ConclusionsThis is the first meta-analysis to demonstrate that ML-based prediction of stroke mortality is feasible. The performance of ML supports its role as an auxiliary tool for identifying high-risk populations, thereby optimizing clinical monitoring and resource allocation. However, due to substantial heterogeneity and a relatively high risk of bias in available studies, caution is warranted in real-world application. The effectiveness of ML may vary across settings, and external validation is recommended before broader implementation. Trial RegistrationPROSPERO CRD420251086321; https://www.crd.york.ac.uk/PROSPERO/view/CRD420251086321

Objective Antiseizure medications (ASMs) are the cornerstone of epilepsy treatment. However, evidence on direct comparison of ASMs is lacking. This network meta-analysis evaluated the comparative efficacy and safety of approved and investigational add-on third-generation ASMs for focal epilepsy in adolescents and adults. Methods Data were retrieved through an extensive literature search of PubMed, Embase, Cochrane Library, and ClinicalTrial.gov databases from inception through August 2025. Findings were reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline (CRD420251180027). Primary efficacy outcomes were ≥50% and 100% responder rates at 12-weeks maintenance duration. Secondary outcomes were corresponding responder rates at 8-weeks maintenance duration. Tolerability was assessed as retention rate. Treatment-emergent adverse events (TEAEs) and TEAEs leading to treatment discontinuation were the safety outcomes. Results The literature search retrieved 345 studies, of which 35 studies were included. All ASMs showed significantly higher responder rates compared with placebo. Significantly higher 100% responder rate was observed with cenobamate (CNB; 400mg/d: Risk ratio [RR] 15; 95% CI, 7.0-39; 200mg/d: RR 8.7; 95% CI, 3.9-22) at a maintenance duration of 12 weeks and 8 weeks (400mg/d: RR 15; 95% CI, 7.0-41; 200mg/d: RR 8.6; 95% CI, 4.0-24). All ASMs showed a patient retention rate comparable with placebo. For overall TEAEs, brivaracetam (BRV; 50mg/d) and BRV ranked the lowest for individual and pooled doses, respectively; placebo ranked the highest in both cases. For TEAEs leading to treatment discontinuation, CNB ranked lower than the placebo. Significance All approved and investigational ASMs were effective add-on treatments for focal epilepsy, with CNB demonstrating the greatest likelihood of achieving seizure freedom.

Focal to bilateral tonic-clonic seizures (FBTCS) is the most severe form of epileptic seizures, posing a major challenge in both management and research. This study aimed to characterize microstructural abnormalities in the normal-appearing cortex of patients with FBTCS and assess their potential for individual-level identification. We retrospectively included 135 unilateral drug-resistant focal epilepsy patients with FBTCS (FBTCS+), 78 without FBTCS (FBTCS-), and 95 healthy controls (HCs). Surface-based morphometry analyses on T1-weighted images were conducted to compare cortical thickness, gyrification index, and sulcal depth among three groups. Significant brain regions between FBTCS + and FBTCS- groups were further segmented for sub-regional analysis. Using these morphological metrics, support vector machine (SVM) models were constructed and validated with cross-validation to differentiate FBTCS+ from FBTCS- patients and left-sided from right-sided epilepsy patients at the individual level. Directly compared to the FBTCS- group, the FBTCS+ group presented significant cortical thinning and gyrification index increase in insula. Subregional analysis on insula revealed cortical thinning and gyrification index increase in the ipsilateral short gyri, and cortical thinning in the ipsilateral long gyrus and central sulcus. Using morphological metrics of insular subregions, SVM models were optimized to differentiate FBTCS+ from FBTCS- patients with an area under the curve (AUC) of 0.735; and to distinguish left-sided from right-sided epilepsy patients with an AUC of 0.871. Our results supported brain network abnormalities related to FBTCS and highlighted the value of insula, particularly its subregions, in identifying FBTCS+ patients and predicting lateralization at the individual level.

Vagus nerve stimulation (VNS) has been proven as an effective and safe adjunct therapy for epilepsy, but real-world evidence is limited. This study aimed to evaluate outcomes of VNS and its cumulative effect through a prospective, multicenter, real-world survey in China, with dynamic follow-up. A total of 83 sites in China participated and 124 epilepsy patients enrolled. Visits were scheduled at 1, 3, 6, 9, and 12 months after VNS. Primary outcomes included seizure response rate (≥ 50

Epilepsy is a common neurological disorder affecting about 50 million people worldwide. Disparities in healthcare resources lead to geographical variation in its diagnosis, treatment, and management. In particular, the gap between plains and plateau regions in western China remains understudied. Assessing healthcare professionals’ expertise and delivering targeted training represent effective strategies to narrow this treatment gap. We compare epilepsy knowledge between physicians in more developed plains regions and less developed plateau regions of western China in order to inform targeted interventions to address regional disparities. In this cross-sectional study, physicians attending epilepsy training sessions in three cities representative of the western plains region of China (Chengdu) and of the western plateau region (Lhasa and Xichang) completed a questionnaire assessing knowledge in several modules of the disease, including its diagnosis, preoperative evaluation, drug and surgical treatments, and management during pregnancy. Of the 349 participants given questionnaires, 325 (93.1

BackgroundPiloerection, a physiological response to cold or emotional stimuli, is a rare autonomic manifestation of epileptic seizures. The anatomical correlates and electrophysiological mechanisms of pilomotor seizures remain poorly understood.MethodsWe conducted a retrospective analysis of 13 patients with pilomotor seizures identified from 8482 individuals monitored at the Epilepsy Center of West China Hospital. Demographics, seizure characteristics, neuroimaging, and neurophysiological findings were analyzed to determine the epileptogenic zones and associated etiologies.ResultsMost (11/13) of our cases showed temporal lobe origin of seizures, with distinctive ictal encephalogram patterns including rhythmic delta activity. Piloerection was consistently accompanied by other autonomic or psychic symptoms. Most patients responded to anti-seizure medications.ConclusionsPilomotor seizures are associated with temporal lobe epilepsy and are frequently associated with specific etiologies such as autoimmune encephalitis.

PURPOSE:To provide consensus-based recommendations for the use of sodium channel blockers (SCBs) in the management of focal epilepsy. METHODS:A three-round modified Delphi procedure was conducted among a Delphi panel of 24 Chinese experts to build a consensus. A steering committee developed 9 statements related to SCBs for the treatment of focal epilepsy, and these statements were evaluated and voted upon by the expert panel. RESULTS:The expert panel achieved consensus on nine statements regarding the treatment recommendations for oxcarbazepine, lamotrigine, lacosamide, eslicarbazepine, topiramate, zonisamide and cenobamate in focal epilepsy patients and treatment adjustments for SCBs. CONCLUSION:This is a Chinese expert consensus on the use of SCBs in focal epilepsy developed using the modified Delphi method. These recommendations can help clinicians in their practice and guide future research.

Sodium valproate (VPA) is widely recognized as the first-line treatment for patients with epilepsy (PWE). However, current studies lack evidence to determine the best add-on medication following VPA monotherapy failure. Lamotrigine (LTG), levetiracetam (LEV), oxcarbazepine (OXC), topiramate (TPM), and carbamazepine (CBZ) also exhibit broad-spectrum activity for seizures. This study aims to compare the therapeutic efficacy of different anti-seizure medication combinations in PWE following valproate monotherapy failure. Individuals were categorized into five groups: VPA + LTG, VPA + LEV, VPA + TPM, VPA + OXC and VPA + CBZ. Each group was further subdivided based on seizure type: generalized onset, focal onset, or unknown onset. The effectiveness of these five groups was compared using variance, χ2 test and Kaplan–Meier survival analysis. A total of 2656 PWEs were included in this study. The ≥ 50

OBJECTIVE:To identify the Electroencephalogram (EEG) microstate characteristics that can distinguish between patients with first unprovoked seizure (FUS) and newly diagnosed epilepsy (NDE), providing insight into predicting the progress of FUS to NDE, and to find the predictive biomarkers for the responsiveness to initial antiseizure medication (ASM) therapy. METHODS:Fifty-six NDE patients in a drug naïve state, 26 FUS patients, and 31 healthy controls (HCs) were compared on microstates features of 21-channel resting-state EEG without artifact. Four classic EEG microstates (A, B, C and D) were derived. The global explained variance (GEV), mean duration (MD), time coverage (TC), and frequency of occurrence (FO) of each microstate were calculated. RESULTS:NDE and FUS patients exhibited decreased MD, TC, and FO in microstate C compared to the HCs. The FUS patients showed decreased MD, TC, and FO in microstate A compared to the NDE patients. Non-seizure free (NSF) patients showed longer MD, higher TC, and FO in microstate B compared to the seizure free (SF) patients. SIGNIFICANCE:EEG microstate serves as electrophysiological markers that can distinguish between patients with FUS and NDE. Additionally, EEG microstate parameters can also serve as the predictive biomarkers for the responsiveness to initial ASM therapy.

Vagus nerve stimulation (VNS) has been widely used in the clinical treatment of epilepsy, while its effects on comorbidities in epilepsy remain incompletely elucidated. This study aimed to evaluate the impact of VNS on comorbidities and quality of life in adult patients with epilepsy. A longitudinal, multicenter cohort study was conducted from 2021 to 2024 among adult patients with epilepsy who underwent VNS implantation. We enrolled 128 participants from 83 hospitals. The inclusion criteria were patients over 18 years old, diagnosed with epilepsy according to the 2014 International League Against Epilepsy guidelines, and having complete data from at least two follow-up visits. Standard assessment tools, including diagnosis according to International Classification of Diseases, 10th Edition (ICD-10), Neurological Disorders Depression Inventory for Epilepsy (NDDI-E), Generalized Anxiexy Disorde-7 (GAD-7), Pittsburgh Sleep Quality Index (PSQI), and Quality of Life in Epilepsy-31 (QOLIE-31) were used to evaluate comorbidities and quality of life. Statistical analysis was performed using SPSS 26.0. The major clinical measurements were changes in the scales above before and after VNS implantation during follow-up. Generalized estimation model was applied to illustrate the effect over time an its relation to seizure control. A total of 113 participants met the inclusion criteria. Baseline characteristics were comparable between the comorbidity and non-comorbidity groups in terms of gender, seizure onset, age at VNS implantation, seizure types, or the number of antiseizure medications used. Significant improvements were observed from the implantation to the end of follow-up. The PSQI score decreased from 5.43 ± 3.60 to 4.44 ± 3.14 (P < 0.01), indicating better sleep quality. Depressive symptoms (NDDI-E) and anxiety symptoms (GAD-7) decreased significantly, with scores dropping from 6.49 ± 4.67 to 4.83 ± 4.37 (P < 0.01) and from 7.15 ± 5.06 to 4.95 ± 3.69 (P < 0.01), respectively. The QOLIE-31 score increased from 54.40 ± 15.70 to 61.33 ± 16.19 (P < 0.01), suggesting improved quality of life. Further analysis indicated that in the early second postoperative follow-up (1 month after implantation), the scales had already improved significantly (P < 0.001 for PSQI and QOLIE-31, P = 0.006 for NDDI-E and GAD-7). We did not find any statistically significant difference between patients with comorbidity and those without on the efficacy of any scales in this study. The efficacy of VNS on the four scales above was related to follow-up time, with a slightly rebound at the last two follow-ups. The NDDI-E as well as the GAD-7 scores were related to better seizure control according to the GEE model. Higher stimulation currents over 1 mA did not improve the efficacy of VNS on the comorbid conditions. VNS implantation significantly improved sleep quality, mental health, and overall quality of life in adult patients with epilepsy. Such effects could be observed shortly after the implantation and were mostly long-lasting. Further research is needed to validate its long-term effects.

The field of epilepsy neural regulation represented by VNS is rapidly developing. Our aim was to investigate the safety and effectiveness of vagus nerve stimulation (VNS) as an adjunct therapy for pediatric epilepsy in a multi-center study across China. Children with epilepsy undergoing VNS as supplementary treatment were consecutively enrolled in this study. Eligibility was limited to children aged 1–16 years with a confirmed epilepsy diagnosis, a stable antiseizure medication regimen, and a minimum of two seizures per 28-day cycle during the 8-week retrospective baseline. Eighty-seven children (54 males; mean age 8.21 ± 3.88 years, range 0–16) were included, with seizures beginning at an average age of 3.03 ± 2.90 years. A ≥ 50

The treatment landscape for focal seizures in China is distinct from those in other regions, with oxcarbazepine and sodium valproate being more commonly used than newer antiseizure medications (ASMs). Cenobamate, a novel ASM, has demonstrated significant efficacy in reducing seizure frequency. However, its efficacy and safety have not been assessed within the Chinese population. The current study analyzed the 24-week double-blind period data of Chinese participants from a randomized, double-blind, placebo-controlled clinical trial (NCT04557085). Efficacy was assessed by determining seizure frequency reduction and responder rates across the cenobamate dose groups (100, 200, and 400 mg) and concomitant ASM groups. Safety was assessed by treatment-emergent adverse events (TEAEs) and treatment-related adverse events (TRAEs) during the double-blind treatment period. This post hoc analysis included 227 participants with a median age of 32 (interquartile range 24–40) years with focal seizures across 24 sites in China. Cenobamate demonstrated a greater median percentage reduction in seizure frequency (100 mg, 39.5

Mesial temporal lobe epilepsy (mTLE) is the most common form of focal epilepsy, often associated with hippocampal sclerosis. Increasing evidence suggests the pivotal role of neuroinflammation in mTLE onset and progression. We used morphometric similarity network (MSN) analysis and the Allen Human Brain Atlas (AHBA) database to investigate structural changes between mTLE and healthy controls, as well as correlation with inflammation-related gene expression. We identified widespread alterations across the frontal and parietal lobes and cingulate cortex linked to neuroinflammatory genes such as PRR5, SMAD3, and IRF3. This correlation was even more pronounced in mTLE patients with hippocampal sclerosis compared to those without. Enrichment analysis highlighted pathways related to neurodevelopment and neurodegeneration, supporting a bidirectional link between mTLE and neurodegenerative diseases. These findings suggest that brain-wide macroscopic morphometric alternations in mTLE are correlated to the neuroinflammation process. It provides circumstantial evidence from a new perspective to support the bidirectional link between mTLE and neurodegenerative diseases.

BACKGROUND Anxiety is a common comorbidity in patients with Crohn’s disease (CD). Data on the imaging characteristics of brain microstructure and cerebral perfusion in CD with anxiety are limited. AIM To compare the imaging characteristics of brain microstructure and cerebral perfusion among CD patients with or without anxiety and healthy individuals. METHODS This prospective comparative study enrolled consecutive patients with active CD and healthy individuals who visited the study hospital between January 2022 and January 2023. Anxiety was measured using the Hospital Anxiety and Depression Scale-Anxiety. The imaging characteristics of brain microstructure and cerebral perfusion were measured by diffusion kurtosis imaging and intravoxel incoherent motion. RESULTS A total of 57 participants were enrolled. Among the patients with active CD, 16 had anxiety. Compared with healthy individuals, patients with active CD demonstrated significantly lower radial kurtosis values in the right cerebellar region 6, lower axial kurtosis (AK) values in the right insula, left superior temporal gyrus, and right thalamus, and higher slow and fast apparent diffusion coefficients (ADCslow and ADCfast) in the bilateral frontal lobe, bilateral temporal lobe, and bilateral insular lobe (all P < 0.05). Compared with patients with CD without anxiety, patients with CD and anxiety exhibited significantly higher ADCslow values in the left insular lobe and lower AK values in the right insula and right anterior cuneus (all P < 0.05). CONCLUSION There are variations in brain microstructure and perfusion among CD patients with/without anxiety and healthy individuals, suggesting potential use in assessing anxiety-related changes in active CD.

Background:The insula plays a crucial role in the pathophysiology of patients with migraine without aura (MWoA), but the exact neurometabolic mechanisms are still unclear. This study aimed to explore possible neurometabolic mechanisms in the insula during the interictal period in MWoA patients, and neurometabolic differences between high frequency (HF) and law frequency (LF) headache patients via proton magnetic resonance spectroscopy (1H-MRS). Methods:A total of 22 MWoA patients and 22 age-, gender-, and education-matched healthy controls (HCs) were included in this prospective cross-sectional study. The subjects underwent routine T1-weighted imaging (T1WI) and single-voxel 1H-MRS scans, with the region of interest fixed in the left insula. Metabolites, including myo-inositol (Ins), N-acetyl aspartate (NAA), choline-containing compound (Cho), creatine and phosphocreatine (Cr), glutamate and glutamine (Glx), were quantified via the linear combination model (LCModel) software, and then corrected for the partial volume effect of cerebrospinal fluid (CSF). The MWoA patients were categorized into LF and HF headache groups according to their headache frequency. Metabolic differences between the groups were tested by an analysis of covariance (ANCOVA), and the clinical relevance of these metabolites was analyzed by Pearson or Spearman correlation analyses. Results:During the interictal period of headache, the Ins (MWoA vs. HCs: 5.16±1.14 vs. 6.21±1.14, P=0.017), NAA (MWoA vs. HCs: 5.70±1.23 vs. 6.59±1.12, P=0.015), and Glx (MWoA vs. HCs: 12.88±1.63 vs. 14.36±2.17, P=0.020) concentrations were significantly decreased in the insula of the MWoA patients compared to the HCs. Further, the HF headache patients had obviously higher Cr levels than the LF headache patients (HF vs. LF: 6.16±0.67 vs. 6.01±0.91, P=0.037). The headache frequency of the MWoA patients was positively correlated with the headache-attributed lost time-90 days (HALT-90) scale (r=0.560, P=0.010) and Hamilton Depression Rating Scale (HAMD) (r=0.529, P=0.017) scores. In addition, a higher HALT-90 score was associated with a higher Cho level in the MWoA patients (r=0.654, P=0.002). Conclusions:The dysfunction or loss of neurons and glial cells, and excitatory neurotransmitter conversion imbalance may be the key changes in the insula of interictal MWoA patients. HF headaches are characterized by hypometabolism, which may be caused by more serious mitochondrial dysfunction.

BACKGROUND:Catamenial epilepsy (CE) is marked by an increase in seizure frequency during specific phases of the menstrual cycle. Advanced techniques such as serum metabolomics and 16S rRNA sequencing are being employed to investigate potential mechanisms and therapeutic approaches for managing CE. METHODS:A total of 30 epilepsy patients (15 with CE and 15 temporal epilepsy) along with 15 healthy controls, were enrolled in this study. Human fecal samples and serum were collected for metabolomics analysis and 16S rDNA sequencing, respectively. RESULTS:A total of 117 candidate metabolites and eight gut microbiota specific to CE were identified. Glycerophospholipid pathways were the most enriched among the candidate metabolites. Notably, serum lysophosphatidylinositol (LPI) 20:4 was associated with an increased seizure frequency. Additionally, differential gut metabolites were found to affect γ-aminobutyric acid degradation. A network linking sex hormones, metabolites, and gut microorganisms was constructed. CONCLUSION:Findings clarify metabolite-microbiome-brain axis interactions, aiding CE pathogenesis understanding.

OBJECTIVE:Data on seizure and pregnancy outcomes in Asian women with epilepsy are limited. We used a Chinese pregnancy registry to assess the impact of seizures and antiseizure medications (ASMs) on pregnant women with epilepsy and their children. METHODS:This is an ongoing prospective multicenter study of pregnant women with epilepsy that has been running since 2012. Eligible participants were consecutively enrolled and had multiple follow-ups up to one year after delivery. We assessed ASM use and seizure frequency during pregnancy to establish potential effects on the mothers and infants and to identify relevant correlations. Descriptive analysis was used to estimate proportions. Logistic regression was used to identify the relevant risk factors and correlations. RESULTS:Of 1907 potentially eligible pregnancies, we included 1763 in 1483 women with known outcomes from January 2012 to February 2022. There were 1278 completed pregnancies, resulting in 1270 live births. Tonic-clonic seizures occurred in fewer than one-third of pregnancies in each trimester. Compared with baseline, seizure frequency remained relatively stable throughout approximately two-thirds of the pregnancies. The majority were on ASM, with levetiracetam (39%), oxcarbazepine (19%), and lamotrigine (17.5%) being the most commonly used. In contrast, only 14.2% of pregnancies were exposed to valproate (VPA). There was a declining trend in treatment adjustments over the course of the pregnancy, with most changes occurring in the first trimester. The incidence of major congenital malformation (MCM) was 4.4%, with cardiogenic and orofacial anomalies being the most common. VPA use (p < .001), lack of folic acid use (p = .009), positive family history of MCM (p = .006), and topiramate (TPM) use (p = .04) were the most important predictors of MCM. SIGNIFICANCE:Seizure control remained stable for the majority of women with epilepsy throughout pregnancy. Family history of MCM, VPA use, TPM use, and not taking folic acid were strong predictors of MCM in infants born to women with epilepsy.

Artificial intelligence (AI) has emerged as a transformative tool in the analysis and management of epilepsy through its integration with electroencephalography (EEG) data. The adoption of AI-assisted solutions in managing epilepsy holds the potential to significantly enhance the efficiency and accuracy for diagnosing this complex condition. However, AI-assisted EEG technologies are infrequently adopted in clinical settings. In this Review, we provide an overview of AI applications in seizure prediction, detection, syndrome classification, surgical planning, and prognosis prediction. Additionally, we explore the methodological considerations and challenges that are relevant in clinical settings. Overall, AI has the potential to revolutionize epilepsy management, ultimately improving patient outcomes and advancing the field of precision medicine. Fostering interdisciplinary collaborations between AI researchers, neurologists, and ethicists will be crucial in creating integrated solutions that address both technical and clinical requirements.