Tobias Bauer

Deep learning and ultra-high-field 7 Tesla (7T) MRI offer potential to improve detection of focal cortical dysplasia (FCD). However, existing models are predominantly trained on 3T MRI, and sufficiently large, annotated 7T datasets are not yet available. We evaluated a 3T-trained FCD detection model on paired 3T and 7T MRI acquired prospectively during presurgical assessment in 2023-2024. The cohort included 115 patients with epilepsy and 10 healthy controls, all undergoing T1-weighted and fluid-attenuated inversion recovery imaging at both field strengths. A publicly available 3D U-Net generated candidate lesions (>0.1 mL), which were compared with manual annotations and independently reviewed by two neuroradiologists and two epileptologists. Based on clinical MRI reports, patients were classified as MRI-suspected FCD (n=11), MRI-negative (n=64), or other lesion (n=40). In MRI-suspected FCD, sensitivity was 45% (5/11) at 7T versus 55% (6/11) at 3T. Among MRI-negative patients, diagnostic yield was 2% (1/64) at 7T versus 11% (7/64) at 3T. Across MRI-negative and other-lesion groups, the model generated eight false-positive candidate lesions at 7T versus 24 at 3T. Cross-field prediction concordance was 71% (5/7 locations) in MRI-suspected FCD, but only 26% (8/31) combined MRI-negative and other-lesion subgroups. Thus, a 3T-trained model can identify FCD on 7T MRI without retraining, but 7T MRI did not improve diagnostic yield. Domain adaptation and improved discrimination between FCD and non-FCD findings are needed before clinical deployment at 7T.

Objective: Epilepsy is characterized by widespread cortical and subcortical structural alterations beyond the epileptic zone. Importantly, the clinical manifestation of epilepsy, including seizure types, psychiatric comorbidities, and treatment responses, has been shown to differ between sexes. However, sex differences in structural alterations in epilepsy are seldomly reported in neuroimaging studies, partly due to limited sample sizes and single‐center designs. We therefore tested whether epilepsy‐related structural alterations differ between males and females, and whether the associations between structural measures and clinical characteristics vary by sex. Methods: We analyzed clinician‐assigned biological sex differences in common epilepsies and their related clinical variables using structural neuroimaging biomarkers in an international multicenter cohort of 1253 epilepsy patients and 1077 healthy controls. We studied cortical thickness and subcortical volume in temporal lobe epilepsy (TLE) and genetic generalized epilepsy (GGE). Results: Both male and female patients with TLE showed widespread cortical and subcortical alterations in sex‐stratified analyses, whereas GGE showed more extensive alterations in female patients and only subtle alterations in male patients. However, no significant sex‐by‐diagnosis interactions were identified in either the TLE ( t = −2.153 to 1.705, all p _FDR >.05) or GGE ( t = −1.797 to 2.774, all p _FDR >.05) cohort. Likewise, no significant interactions were found between sex and age at onset ( t _TLE = −1.835 to 1.092, t _GGE = −2.394 to 1.392, all p _FDR >.05) or disease duration ( t _TLE = −2.643 to 2.048, t _GGE = −1.234 to 2.017, all p _FDR >.05) in either patient group. Significance: Overall, although sex differences in behavioral and clinical outcomes of epilepsy have been consistently reported in previous studies, our results indicate that sex‐related clinical differences in epilepsy are not strongly captured by macroscale structural magnetic resonance imaging measures, and may instead arise from hormonal, neurochemical, or functional–network mechanisms.

Background: Software tools can support MRI diagnosis of focal cortical dysplasias (FCDs) – subtle, surgically-treatable lesions that underlie drug-resistant focal epilepsy (DRFE). MELD-graph, a widely used open-source tool combining surface-based morphometry with a convolutional neural network, was developed using 1.5T and 3T MRI data. 7T MRI is increasingly used for lesion detection in DRFE. This study investigated whether MELD-graph is robust to domain shift at 7T. Methods: We collated paired 3T and 7T T1-weighted MRI, and FLAIR where available, from 23 patients with FCD (6 histologically-verified, 17 identified radiologically) and 39 healthy controls at four epilepsy centres. Manual lesion labels were used as ground truth. MELD-graph was run on 3T images, on 7T images with the default pipeline, and on 7T images with adaptations for 7T MRI (SPM bias-field correction, SynthStrip skull-stripping, sub-millimetric FreeSurfer reconstruction), with and without ComBat harmonisation. FCD detection performance and surface morphometry features were compared. Results: Sensitivity for FCD detection was 65% (15/23) at 3T, 48% (11/23) for 7T default and 52% (12/23) for 7T adapted, with Combat harmonisation. Among radiologically 3T-negative patients, 2/8 were detected at 3T and 3/8 at 7T; and among histologically confirmed FCD, 3/6 at 3T and 1/6 at 7T. Specificity was significantly lower at 7T (default 39%, adapted 45%) than 3T (76%). False-positives predominantly occurred in the basal temporal lobes. ComBat harmonisation improved specificity in all conditions and largely compensated for field-strength dependent differences in surface morphometry features. Adding FLAIR at 7T increased false-positive predictions. Conclusion: Sensitivity is comparable but specificity of MELD-graph is lower at 7T compared to 3T. Adaptations for 7T MRI can reduce segmentation errors but do not significantly improve FCD detection, while harmonisation improves performance. The adapted pipeline is released open-source. 7T MELD-graph outputs may inform radiologist review in challenging cases but require cautious interpretation.

Background: Hippocampal sclerosis (HS) is a common cause of drug-resistant focal epilepsy (DRFE) and amenable to neurosurgical treatment. Detection relies on MRI but can be challenging. 7 Tesla (T) ultra-high field MRI and automated MRI post-processing tools have independently been shown to improve radiological diagnosis of HS. However, combining these approaches remains underexplored. This study evaluated whether AID-HS, a tool for HS detection developed using 3T MRI, generalises to 7T MRI data. Methods: We collated a dataset of paired 3T and 7T T1-weighted MRI from four epilepsy centres, including 23 patients with HS, 39 healthy controls, and 23 individuals with focal cortical dysplasia as disease controls. Histopathology served as the gold standard for defining HS where available (n=7), otherwise radiological findings (n=16). AID-HS was applied to images acquired at both field strengths, and sensitivity and specificity for detection and lateralisation of HS were compared. Additionally, agreement of hippocampal features across 3T and 7T was evaluated. Results: We found no evidence of a difference in performance of AID-HS between 3T and 7T. Sensitivity for detection of unilateral HS was 63% (12/19) at 3T and 68% (13/19) at 7T (McNemar’s exact test p=1.0). Specificity in controls was 97% (60/62) at 3T and 100% (62/62) at 7T (p=0.5). Bilateral HS was correctly flagged in 3 of 4 cases using feature-based criteria, with high specificity in controls. Quantitative hippocampal features showed moderate to good agreement across field strengths (ICC 0.70 to 0.98), with small differences observed for volume and thickness estimates. Conclusion: AID-HS provides robust detection and lateralisation of HS across multiple 7T MRI centres, highlighting its potential to enhance lesion detection. Future work is needed to investigate whether models trained on 7T data can leverage the improved image quality for further gains in HS detection performance.

Diagnostic MRI evaluation of temporal lobe epilepsy (TLE) depends on the subjective visual interpretation of MRI images. These interpretations could be enhanced by quantitative artificial intelligence (AI) support tools. Humans often make sequential and conditional decisions during their radiological interpretations, such as whether an abnormality is present and if present, characterizing the abnormality. It is not known whether it is superior to train AI to treat every decision separately in a similar step-wise manner or to train a model holistically on all decisions simultaneously. Here, we analyzed three large epilepsy MRI datasets [n=3,676, 2,320 people with epilepsy and 1,356 healthy controls (HC)] to perform two tasks: 1) establish the presence of a TLE pattern on MRI and 2) determine TLE pattern lateralization. We compared Step-wise models that independently classify TLE versus HC and lateralize patients as left TLE (L-TLE) or right TLE (R-TLE), against a Simultaneous model trained to distinguish all three classes in a single step. To do this, 3D volumetric T1-weighted images were input into an EfficientNetV2 model multiple times to ensure reproducibility of results. Class prediction, model classification confidence, and saliency maps were output for interpretability. Step-wise models outperformed the Simultaneous model on both tasks (both ps<.001), with an average ∼2.8% accuracy increase for discriminating HC from TLE and an average 12.7% accuracy increase for distinguishing L-TLE from R-TLE. For both the Step-wise and Simultaneous models, important features discriminating TLE from HC included the known TLE limbic pattern involving the hippocampus, para-hippocampal cortical regions, cingulate cortex, and lateral temporal regions. However, there was less concordance between the Step-wise and Simultaneous models for the L-TLE versus R-TLE task (all Fisher’s Zs>10.5, ps<.001); the Step-wise model focused less on subcortical regions such as the thalamus and hippocampus and focused more on distributed cortical pathology. Across the two Step-wise models, 95.1% of TLE patients had accurate classifications in either HC versus TLE and/or L-TLE versus R-TLE tasks. These results included 69.6% of patients being both correctly labeled as TLE and lateralized, 13.9% being correctly labeled TLE but lateralized incorrectly, and 11.6% being lateralized correctly but not detected as TLE. These findings provide evidence that diagnostic tasks with simpler, Step-wise AI models may enhance diagnostic performance and interpretability in clinical workflows. Future AI clinical support tools can leverage this step-wise approach in the early identification of TLE-related structural patterns, supporting timely diagnosis and treatment decisions.

Hemispherotomy is a surgical intervention for severe drug-resistant epilepsy that fully disconnects the lesional brain hemisphere from the rest of the nervous system. This creates a rare opportunity to explore whether consciousness can be sustained without external input. The default-mode network, linked to introspective mental activity, is considered a necessary, though not sufficient, condition for consciousness. In this prospective case-control study, resting-state functional MRI data from 26 individuals post-hemispherotomy and 24 healthy controls were analysed. Default-mode network activity was assessed using seed-to-voxel correlations based on the peak activity in the precuneus of controls, delineated by group independent component analysis. In connected hemispheres, typical default-mode network patterns were preserved and did not differ significantly from controls. Isolated hemispheres also showed preserved positive default-mode network connectivity, though negative connectivity was significantly reduced compared to controls. In two patients with both pre- and post-surgical scans, the individual default-mode network remained detectable but was divided between the isolated and connected hemispheres. These findings suggest that default-mode network activity persists in isolated hemispheres, meeting a necessary condition for consciousness. However, the presence of default-mode network activity alone does not confirm conscious awareness, encouraging further investigation into the neural correlates of consciousness in isolated hemispheres.

Introduction: Brain structural differences consistent with an older-appearing brain have been reported in people with epilepsy, but the extent to which these differences reflect clinical characteristics vs broader socioeconomic context is unclear. We investigated whether country-level socioeconomic factors are associated with neuroanatomical differences in adults with epilepsy using MRI-based age prediction, along with epilepsy subtype, sex, and clinical factors. Methods: Structural MRI and clinical data were collected from 26 epilepsy centres across 12 countries in the Americas, Australia, Europe, Asia and Africa. MRI-based age estimates were estimated using a previously developed prediction model trained on 29,175 healthy subjects. Brain predicted age difference (BrainPAD) was calculated as the difference between MRI-predicted brain age and chronological age. National gross domestic product (GDP) per capita and income inequality (Gini index) were obtained from the World Bank. Associations between BrainPAD and epilepsy subtype (temporal lobe epilepsy, extratemporal epilepsy, and genetic generalised epilepsy), national socioeconomic context (GDP per capita and Gini index), age and sex were assessed using regression models. Results: We analyzed 2,109 individuals with epilepsy and 1,041 healthy non-epilepsy controls (57% female; median age = 35; range 17-83). BrainPAD was higher in epilepsy than controls (β 4.2 years, SE 0.4; t=10.6), with increases ranging from 2.5 to 6 years across subtypes. Male sex was associated with 1 year higher BrainPAD relative to females (SE 0.33, t=3.12). There were no main effects of GDP or Gini index; however, significant interactions between were observed. The effect of epilepsy on BrainPAD was greater in countries with lower GDP per capita (t=-2.74) and higher income inequality (t=2.72). Conclusions: Clinical factors and socioeconomic context both influence brain structural ageing in epilepsy. These findings highlight the importance of geographic and economic diversity in neuroimaging research and underscore the relevance of global socioeconomic context when interpreting brain health measures.

Background: 7T MRI received FDA/CE clearance almost 7 years ago. However, until today, it has not yet been widely adopted in clinical routine. This is mainly due to field inhomogeneities that impede whole-brain coverage. Moreover, the long scan times often associated with high-resolution imaging are an additional limiting factor. Purpose: To combine calibration-free parallel transmit technology (pTx) using universal pulses (UP) with advanced imaging acceleration strategies to achieve homogenous multicontrast 7T MRI with whole-brain coverage and high spatial resolution in short scan time. Materials and Methods: Ten healthy volunteers were scanned both with conventional vendor-provided sequences and with custom sequences for anatomical whole-brain imaging [ -weighted, -weighted, FLAIR, and susceptibility-weighted]. The scan times for the 2 anatomical protocols were matched (25 minutes). In addition, a quantitative MRI protocol [multi-parametric mapping (MPM) and chemical exchange saturation transfer (CEST)] was scanned twice using custom sequences with conventional (circular polarized) and UPs, respectively, in a scan time of 2×25 minutes. Moreover, 4 patients with different neurological diseases were scanned, namely temporal lobe epilepsy, spinocerebellar ataxia, cerebral amyloid angiopathy, and glioblastoma. For the patients, only optimized custom sequences with UPs were acquired. Results: Compared with conventional implementations, the custom sequences provide strongly improved image homogeneity and quality with significantly higher SNR and CNR across the whole brain, including cerebellum and brain stem. Moreover, UPs improve the repeatability of derived quantitative parameters. The suggested protocol has additionally been successfully demonstrated in 4 patients with different neurological pathologies. Conclusions: Homogeneous whole-brain 7T MRI with high spatial resolution and high image quality is possible in clinically feasible scan times. The developed protocol can be applied without any expert knowledge and is ready for clinical use. The approach could largely extend applicability of UHF MRI in neuroradiology paving the way for increased routine use of 7T MRI.

Autoimmune encephalitis (AIE) with anti-leucine-rich glioma-inactivated 1 (LGI1) antibodies typically manifests with subacute cognitive deficits, seizures, and psychiatric symptoms, mostly in older adults. Immunotherapy (IT) leads to the cessation of seizures in most patients, yet some develop AIE-associated epilepsy (AEAE) and persistent cognitive deficits. The aim of this large multicentric retrospective observational cohort study was to assess long-term outcomes of patients with anti-LGI1 encephalitis regarding seizures and AEAE and to identify associated factors. We included patients with anti-LGI1 encephalitis from 3 national referral centers/consortia meeting the following inclusion criteria: (I) definite LGI1 limbic encephalitis (Graus criteria); (II) occurrence of seizures; and (III) follow-up period ≥24 months. We aimed to (1) determine the risk of seizure recurrence (ROSR) on remission, (2) investigate clinical and paraclinical biomarkers for an effect on time to seizure remission using Cox proportional hazard modeling (n = 188), and (3) assess the risk of AEAE and determine associated factors (n = 236). AEAE was observed in 5.9% (16/271) of the full cohort. Both AEAE (16/16 vs 129/215, p = 0.001) and longer time to seizure remission (OR 1.36 per year, p = 0.025) were associated with persistent cognitive impairment. Patients with pilomotor seizures had a lower rate of seizure remission (hazard ratio [HR] 0.58, 95% CI 0.55-0.60, p < 0.001) while patients under IT administration had a higher rate of seizure remission over time (HR 12.4, 95% CI 9.67-16.0, p < 0.001). In addition, patients receiving second-line IT tended to achieve earlier seizure remission (log-rank test, p = 0.019). The ROSR at 12, 60, and 120 months on seizure remission was 9% (95% CI 4.5%-13%), 20% (95% CI 11%-28%), and 53% (95% CI 14%-74%), respectively. In conclusion, our results demonstrate that AEAE in anti-LGI1 encephalitis is rare and suggest that the diagnosis of epilepsy is inappropriate in patients reaching seizure remission because of a relatively low ROSR. Accordingly, on seizure remission, the diagnosis of acute symptomatic seizures would be appropriate. Moreover, we validate and quantify the importance of IT for seizure remission and identify biomarkers associated with lower rates of seizure remission. Late remission of seizures and AEAE were associated with persistent cognitive impairment.

Background: High-field (HF) MRI is a standard diagnostic tool for brain cancer, but its high cost and technical demands limit accessibility in low- and middle-income countries. Recent advancements in ultra-low field (ULF) MRI technology, including the development of portable scanners, offer a promising solution to these challenges. This study evaluates the diagnostic capabilities of ULF-MRI in detecting brain cancer and compares radiological evaluation using ULF- with HF-MRI. Methods: Consecutive patients with suspected or confirmed brain tumors undergoing routine 3T HF-MRI at the University Hospital Bonn were recruited for this study and underwent ULF-MRI. Eligible patients were at least 18 years old and had MRI-abnormalities in the HF-MRI. The 0.064 Tesla Swoop portable MR Imaging System was utilized. HF-MRI and ULF-MRI scans were independently evaluated by two experienced neuroradiologists and results were compared. Results: Thirteen patients were recruited, of whom 11 (85%) were diagnosed with brain tumors. In 11/11 (100%) patients with brain tumors, ULF-MRI identified tumor lesions corresponding to the findings of HF-MRI. In 7/11 (63.6%) identification of all tumor lesions could be achieved. Three of four further relevant imaging findings in HF-MRI (e.g. acute hydrocephalus or concomitant ischemia) were also found in in ULF-MRI. Conclusion: This single-center study demonstrates that ULF-MRI is a practical tool in neuro-oncology, which may particularly be helpful in resource-limited settings. Further research is required to define the role of ULF-MRI alongside existing imaging modalities for brain cancer diagnosis and management.