Abstract
Automated detection of lesions using artificial intelligence creates new standards in medical imaging. For people with epilepsy, automated detection of focal cortical dysplasias (FCDs) is widely used because subtle FCDs often escape conventional neuroradiological diagnosis. Accurate recognition of FCDs, however, is of outstanding importance for affected people, as surgical resection of the dysplastic cortex is associated with a high chance of postsurgical seizure freedom. Here, we make publicly available a dataset of 85 people affected by epilepsy due to FCD type II and 85 healthy control persons. We publish 3D-T1 and 3D-FLAIR, manually labeled regions of interest, and carefully selected clinical features. The open presurgery MRI dataset may be used to validate existing automated algorithms of FCD detection as well as to create new approaches. Most importantly, it will enable comparability of already existing approaches and support a more widespread use of automated lesion detection tools.
Similar content being viewed by others
Explore related subjects
Find the latest articles, discoveries, and news in related topics.Background & Summary
Focal cortical dysplasia (FCD) is a malformation of cortical development that frequently causes drug-resistant focal epilepsy. In drug-resistant individuals, epilepsy surgery often remains the most successful option for treatment. FCDs are the most commonly resected epileptogenic lesions in children1 and the third most common in adults2. Clinical manifestation of epilepsy due to FCDs usually occurs in early childhood, but later onset is not uncommon3. Seizure semiology depends on the location of FCD, which can occur in any brain lobe. The distribution across the cortical lobes is non-uniform, and in particular, FCD type II occurs most frequently in the frontal lobe2,4,5. FCDs were initially histopathologically categorized according to Palmini6. This system was first revised in 20117 and most recently in 20228 by the International League Against Epilepsy (ILAE). Several imaging features have been described. These are cortical thickness, blurring of the gray-matter/white-matter junction, gyration anomalies, and focally increased signals in the subcortical white matter, which sometimes extend to the ventricle as a so-called “transmantle sign”9. Nevertheless, FCDs often escape conventional neuroradiological diagnosis because these abnormalities can be very subtle. In such cases, finding the correct diagnosis can be challenging, and people with epilepsy must undergo extensive diagnostics (e.g., invasive EEG). So-called “MRI negative” cases of people with focal epilepsy but without visible MRI abnormalities have a lower chance of undergoing epilepsy surgery and show a significantly worse surgical outcome10,11. The success of epilepsy surgery depends on the accurate identification and complete excision of the epileptogenic zone. However, the chance of seizure freedom after epilepsy surgery is exceptionally high in people with epilepsy due to FCD12. Various automated or semi-automated postprocessing approaches have been developed in the past few years to optimize the visualization of FCDs13,14,15,16,17,18,19,20,21,22,23,24. David et al.24, used parts of this dataset (58 individuals) for the independent validation of an artificial neural network (ANN) for robust automated detection of FCDs based on morphometric feature maps generated by the Morphometric Analysis Program (MAP)14,24. Several other descriptive studies are also based on parts of this dataset3,25,26,27,28.
There are three major problems for FCD-detection-algorithms trained and validated using single-center datasets: One is that cohorts often are too small to develop a robust classification algorithm. Second, studies may recruit training and testing data from the same sample of people with epilepsy (internal validation), which is why their performance is overrated. Third, radiological diagnoses or ratings of MRI of individuals with FCD may strongly vary from site to site. The MRI volume of an individual with FCD may be described as “MR-negative” in one site and as “radiologically described FCD” in another site with corresponding effects on the evaluation of the respective detection algorithms. It, therefore, is of importance that the same well-annotated and sufficiently large dataset is being used for the validation of different algorithms and that this dataset has not been used for the training of these algorithms. Gill and colleagues overcame these problems by developing a multicenter-validated deep learning detection algorithm for FCD15. This algorithm has been trained and validated on a multicenter dataset with many people with histologically confirmed FCD. The authors note that the dataset contains information that could compromise the privacy of research participants, which is why the dataset is not publicly available15.
With the publication of the open presurgery MRI dataset of people with epilepsy due to focal cortical dysplasia and of control persons, we aim to overcome the abovementioned challenges. We hope to enable fellow researchers to measure the generalizability of automated detection methods, so-called external validation. Approaches are best validated and compared on such an external benchmark dataset. Additionally, this dataset allows the improvement of existing approaches by training them on more data. The longstanding goal of this publication is to optimize existing or new approaches for automated lesion detection until they become part of routine clinical practice and are not exclusively reserved for specialized clinics. Improved focus localization in imaging with a widely used tool may reduce the need for invasive diagnostics and the associated health risks and healthcare system costs.
Methods
Study approval
The ethics committee of the University of Bonn (Lfd.-Nr. 346/21) approved this study including anonymized publication of the MRI datasets as well as selected clinical/demographical characteristics of people with epilepsy and control persons. It adheres to the General Data Protection Regulation of the European Parliament and the Council. It has been conducted in close collaboration with the Data Protection Officer of the University of Bonn Medical Center (AF) and it includes only data of people who provided written and informed consent for the publication of their data as adults.
Dataset selection
We first identified all people with epilepsy treated at the Department of Epileptology at the University Hospital Bonn from 2006 to 2021 due to histologically verified FCD type II or radiologically suspected FCD type II. We selected all who were over 18 years of age at the time of the conduction of our study. Individuals who were minor at the time the MRI conducted, were only included if they were adults at the time their data were prepared for publication. In parallel, we ascertained data from healthy control persons over 18 years of age. These people were contacted again and informed about the planned publication of anonymized MRI data in conjunction with selected demographical characteristics. People willing to participate in the study then received an information letter. We were able to contact 92 of 137 individuals with epilepsy and FCD by telephone. Of these, 85 (92.4%) gave informed consent for their MRI data and associated clinical characteristics to be made available anonymously to the scientific public. Accordingly, we included 85 control persons who consented to publication of their data in this study. Clinical characteristics of people with epilepsy were retrieved from the clinical records of the department of epileptology. The anatomical location of the FCD was indicated by two neuroradiologists. Complicated cases were discussed in a joint conference of experienced neuroradiologists and epileptologists, taking into account the results of all diagnostic modalities. Only in one case, the FCD could only be found after MAP14 was performed, which is why the initial MRI assessment by the neuroradiologist did not include it.
After the lesion location was determined, the definition of the region of interest (ROI) was conducted by the collaboration of two board-certified neurologists (see Region of Interest below). We summarized MRI diagnoses as “suspected FCD II”, “no abnormalities” ( = “MRI-negative”), and “other” (meaning that abnormalities were described as part of the clinical routine, but no FCD-suspect lesion was present; in these cases, the diagnosis was made only by histology). For histologically confirmed FCDs, we included only cases of FCD II (FCD IIa or IIb). FCD I or III were not included. Histological classification of FCDs was performed at the Department of Neuropathology of the University Hospital Bonn according to the ILAE classification7. For FCD type II, the results of the ILAE classification7,8 are identical to those of the Palmini classification6.
For those people who underwent epilepsy surgery, postoperative outcome was assessed 12 months after epilepsy surgery and at the latest clinical follow-up according to the Engel classification29. We also recorded the age at the MRI scan and the sex for all subjects. In people with epilepsy, age at first epileptic seizure was also ascertained.
Whether additional diagnostics (PET-CT, SPECT, MAP analysis14 or an invasive EEG) were performed in people with epilepsy can be found in the participants.tsv file30.
Demographics and clinical characteristics of the published data
Of the 85 people with epilepsy who participated in the study, 35 (41.2%) are female and 50 (58.8%) are male. The mean age at onset of the first epileptic seizure was 10 years (min.: 0.5 years; max.: 41 years; median: 8 years). At the time of MRI Scan, the mean age was 28.9 years (min.: 12 years; max.: 65 years; median: 25 years), and the mean age at epilepsy surgery was 29.2 years (min.: 12; max.: 59 years; median: 27 years). The average age of the control persons at the time of the MRI scan was matched with the age of people with epilepsy (min.: 22 years, max.: 62 years; median: 28 years).
Five (5.9%) of the 85 people with epilepsy were initially radiologically diagnosed as “no abnormalities” ( = “MRI-negative”). In four of these, the diagnosis was established histologically. Most FCDs (62.4%) were located in the frontal lobe. The remainder were distributed among the parietal (18.8%), temporal (5.9%) and occipital lobe (2.4%), as well as the insular region (3.5%). Six (7.1%) of the FCDs were spread over two brain lobes.
Out of 85 people with epilepsy, 61 people (71.8%) were drug-resistant according to the definition of the International League Against Epilepsy (ILAE)31 and 50 people (58.8%) underwent epilepsy surgery. 34 people with epilepsy (68%) were classified as FCD IIb and 16 (32%) as FCD IIa, according to the Palmini classification6. Regarding postoperative outcomes, five (10%) were missing follow-up data. Of the remaining 45 people, 80% were seizure-free (Engel class IA) at 12-month follow-up, and only 4.4% had no worthwhile improvement (Engel class IV).
Table 1 provides a summary overview of the demographic and clinical characteristics.
The clinical characteristics of all people with epilepsy and control persons are listed in the participants.tsv file, accessible on OpenNeuro along with the dataset30.
Imaging data
MRI was performed at the Life & Brain Center in Bonn using a 3 Tesla MRI-Scanner (Magnetom Trio, Siemens Healthineers, Erlangen, Germany). As part of the comprehensive MRI protocol, a fluid-attenuated inversion recovery (FLAIR) sequence and a T1 sequence, were recorded. Due to a scanner update in early 2014, two different acquisition protocols were used. Before the update, an eight channel headcoil was used, after the update, a 32 channel headcoil was used. T1-weighted images were acquired using MPRAGE sequences. Scanning parameter before the update (“t1_iso1”) were TR = 1570 ms, TE = 3.42 ms, TI = 800 ms, flip angle 15°, matrix 256 × 256 pixel, voxel size 1.0 mm × 1.0 mm × 1.0 mm. Parameters after the update (“t1_iso0.8”) were TR = 1660 ms, TE = 2.54 ms, TI = 850 ms, flip angle 9°, matrix 320 × 320 pixel, voxel size 0.8 mm × 0.8 mm × 0.8 mm. We have used three different FLAIR protocols also resulting from system upgrades and software updates. All protocols are based on a turbo spin echo (TSE) sequence and provide 1 mm isotropic images with T2-weighted FLAIR contrast. The protocols differ in the imaging acceleration techniques (Partial Fourier, GRAPPA, or both), the image filtering applied, and the repetition times TR (5 s, 6 s, 7 s) set. Furthermore, inversion pulse application is selective only in the most recent protocol from which the most data are available.
T2-weighted FLAIR parameters before the update (“flair_tr7”) were TR = 7000 ms, TE = 372 ms, TI = 2220 ms, flip angle 150°, matrix 256 × 256 pixel, voxel size 1.0 mm × 1.0 mm × 1.0 mm. Parameters after the update (“flair_tr5”) were TR = 5000 ms, TE = 388 ms, TI = 1800 ms, flip angle 120°, matrix 256 × 256 pixel, voxel size 1.0 mm × 1.0 mm × 1.0 mm. One person with epilepsy also received a discretely different isotropic flair sequence (“flair_t6”: TR = 6000 ms, TE = 388 ms, TI = 2100 ms, flip angle 120°, matrix 256 × 256 pixel, voxel size 1.0 mm × 1.0 mm × 1.0 mm. Sequence parameters are published for all participants.
Imaging data of people with FCD
High-resolution 3D T1-weighted MRIs were acquired for all people with FCD. The resolution of isotropic T1 sequences differed in their voxel size (mm3): Of the 85 people with FCD, 43 (50.6%) received T1-weighted sequences with a voxel size of 1 mm × 1 mm × 1 mm and 42 (49.4%) received T1-weighted sequences with a voxel size of 0.8 mm × 0.8 mm × 0.8 mm. Seventy-eight people with FCD (91.8%) received an isotropic FLAIR sequence. The remaining seven persons (8.2%) received a 2D FLAIR (“flair_2D”) sequence.
Imaging data of healthy control persons
We recorded high-resolution isotropic T1-weighted and isotropic FLAIR sequences for all 85 control persons. All control persons underwent the same isotropic FLAIR sequence. Regarding T1 sequences, as for people with FCD, T1-weighted isotropic images were performed with a voxel size of 0.8 mm × 0.8 mm × 0.8 mm (78 controls, 91.8%) as well as with a voxel size of 1 mm × 1 mm × 1 mm (seven control persons, 8.2%).
An overview of the MRI sequences performed in people with epilepsy as well as the healthy control persons is given in Table 2.
For the visualization of epileptogenic lesions, the FLAIR sequence is the most appropriate32. However, morphometric analysis of 3D T1-weighted sequences is often performed as part of the preoperative workflow to detect subtle lesions27,33. Hence, we publish the FLAIR and the T1 sequence of all participants. Our MRI protocol has been adjusted according to changing clinical standards, resulting in different sequences being performed across the sample (see above). Details of the acquisition parameters of each scan can be found in the associated JSON files on OpenNeuro30.
We provided all images in NIfTI format and structured the dataset according to the BIDS specification34. NIfTI format is used by the neuroimaging research community35,36. This data format can be processed with common neuroscientific tools such as FreeSurfer or FSL, allowing easy usability of the dataset.
Data anonymization
First, all personal identifiable information (such as name, date of birth, date of scan) were removed from the datasets. Secondly, all published MRI sequences were processed using the pydeface software37. This software removes all facial characteristics (so-called “defacing”). Images were visually inspected to ensure the anonymization process was successful and the neurocranium remained unaffected. Age was further categorized into age ranges of five years at onset of epilepsy, at MRI examination and, if performed, at surgery.
Region of interest (ROI)
Ground-truth lesion masks of the dysplastic cortical regions were manually labeled based on 3D-FLAIR-weighted sequences. This was done by the collaboration of two neurologists (F.S. and T.R.), both with many years of experience in the field of epilepsy imaging, resulting in only one lesion per subject. Each lesion mask was created by one rater and reviewed by the other one. As outlined above, the general location of the lesion was determined by two neuroradiologists. Complicated cases were discussed in a joint conference of experienced neuroradiologists and epileptologists, taking into account the results of all diagnostic modalities. The definition of the ROIs was initiated in the plane where the FCD was found to be most visible and reviewer in all three plans. In case of disagreement about the extension of the ROI, all available information (MAP, SPECT, PET, invasive EEG diagnostics, postoperative MRI) was used until both reviewers agreed on the mask. Since both raters had access to all the above-mentioned information as well as the postoperative MRI images, if applicable, it was possible to create lesion masks for MRI images that were initially classified as “MRI negative” or as “no abnormalities.”
Data processing
Preprocessing of the MRI data included conversion of the DICOM data to NIfTI format and reorganizing the NIfTI files into the Brain Imaging Data Structure (BIDS) using Dcm2Bids (https://github.com/UNFmontreal/Dcm2Bids).
Furthermore, the anonymization of the MRI-sequences was ensured by removing voxels with facial features using the pydeface software (https://github.com/1-w/pydeface). Based on the FLAIR-weighted sequences, manual labeling of the region of interest was performed using the FMRIB Software Library (FSL) editing tool to create lesion masks38. Detailed information about the code for preprocessing can be found in the Code Availability section.
Data collection and preprocessing workflow is shown in Fig. 1.
Data Records
The unprocessed magnetic resonance imaging data with their associated metadata and manually labeled ROIs are publicly available on OpenNeuro30 (https://openneuro.org/datasets/ds004199/). The data and metadata were organized according to the machine-readable Brain Imaging Data Structure (BIDS)34. BIDS is an organizing and naming convention for neuroimaging data and associated metadata designed to facilitate data sharing and reuse. All neuroimaging data are available in the recommended compressed NIfTI files (.nii.gz). The tabular data file of clinical metadata is in text file format with tab-delimited values and all data descriptor files are in JavaScript Object Notation format. All MRI data files are accompanied by JSON files containing MRI acquisition parameters. Identifying metadata (e.g. name, date of birth, date of MRI scan) were removed. The top-level BIDS directory contains a dataset_description.json file with a description of the dataset, a descriptive README file, and the participants.tsv file (with the accompanying participants.json file) containing clinical and demographical metadata of subjects. The name and values of each column of the participants.tsv file is further described in the accompanying participants.json file. Missing information were indicated by “n/a” in accordance with the BIDS convention.
The imaging dataset is organized into 170 main folders. Each folder name represents an individual ID starting with “sub-” followed by a number. Each folder contains the subfolder “anat”, where the FLAIR and T1-weighted sequences in NIfTI format and the associated JSON files can be found. More detailed information on the sequences can be viewed in the associated JSON file. For people with epilepsy, the corresponding ROI in NIfTI format can also be found in the “anat” folder.
Technical Validation
Overall, the data quality was rated as acceptable after visual inspection (by F.S.). It was first assured that the field of view comprised the complete brain, that there were no strong motion artifacts, that there were no obvious field inhomogeneities and no metal artifacts affecting the neurocranium (e.g., orthodontic retainers). The quality review was intentionally designed to resemble a visual quality review by a clinical neuroradiologist to ensure realism of the dataset. To create a benchmark dataset that was as close to reality as possible, MRI sequences that contained, e.g., slight motion artifacts were also included. In addition, the images of people with epilepsy had previously been part of the clinical routine in epileptological diagnostics and thus corresponded to a qualitative standard in everyday clinical practice. MRI data from control persons were also visually inspected and reviewed for neurological abnormalities or previously unknown diseases. After removal of the viscerocranium (i.e., “defacing”), the MRI sequences were again visually inspected to verify that the anonymization process was successful. The regions of interest were labeled and reviewed independently by two experienced raters. The dataset conforms to the Brain Imaging Data Structure (BIDS)34, version 1.6.0, and its validity was checked using the BIDS Validator39, version 1.8.9 (http://bids-standard.github.io/bids-validator/) .
Usage Notes
All data are publicly available and are hosted on the OpenNeuro platform (https://openneuro.org/)30. They are organized according to the BIDS standard34.
Code availability
Preprocessing of MRI data was performed with freely available neuroimaging tools (DICOM to NIfTI36 conversion and defacing37). The manually labeled ROIs were generated with the free software “FSL”38. In a final step, the correctness of the BIDS structure of our data was checked using the BIDS validator39. Source code and associated web link can be found in Table 3.
References
Guerrini, R. & Barba, C. Focal cortical dysplasia: an update on diagnosis and treatment. Expert Rev. Neurother. 21, 1213–1224 (2021).
Blumcke, I. et al. Histopathological findings in brain tissue obtained during epilepsy surgery. N. Engl. J. Med. 377, 1648–1656 (2017).
Rácz, A. et al. Age at epilepsy onset in patients with focal cortical dysplasias, gangliogliomas and dysembryoplastic neuroepithelial tumours. Seizure 58, 82–89 (2018).
Wagstyl, K. et al. Atlas of lesion locations and postsurgical seizure freedom in focal cortical dysplasia: a MELD study. Epilepsia 63, 61–74 (2022).
Lamberink, H. J. et al. Seizure outcome and use of antiepileptic drugs after epilepsy surgery according to histopathological diagnosis: a retrospective multicentre cohort study. Lancet Neurol. 19, 748–757 (2020).
Palmini, A. et al. Terminology and classification of the cortical dysplasias. Neurology 62 (2004).
Blümcke, I. et al. The clinicopathologic spectrum of focal cortical dysplasias: a consensus classification proposed by an ad hoc task force of the ILAE diagnostic methods commission. Epilepsia 52, 158–174 (2011).
Najm, I. et al. The ILAE consensus classification of focal cortical dysplasia: an update proposed by an ad hoc task force of the ILAE diagnostic methods commission. Epilepsia 63, 1899–1919 (2022).
Widdess-Walsh, P., Diehl, B. & Najm, I. Neuroimaging of focal cortical dysplasia. J. Neuroimaging 16, 185–196 (2006).
Bien, C. G. et al. Characteristics and surgical outcomes of patients with refractory magnetic resonance imaging-negative epilepsies. Arch. Neurol. 66, 1491–1499 (2009).
Téllez-Zenteno, J. F., Ronquillo, L. H., Moien-Afshari, F. & Wiebe, S. Surgical outcomes in lesional and non-lesional epilepsy: a systematic review and meta-analysis. Epilepsy Res. 89, 310–318 (2010).
Fauser, S. et al. Long-term seizure outcome in 211 patients with focal cortical dysplasia. Epilepsia 56, 66–76 (2015).
Antel, S. B. et al. Automated detection of focal cortical dysplasia lesions using computational models of their MRI characteristics and texture analysis. Neuroimage 19, 1748–1759 (2003).
Huppertz, H. J. et al. Enhanced visualization of blurred gray-white matter junctions in focal cortical dysplasia by voxel-based 3D MRI analysis. Epilepsy Res. 67, 35–50 (2005).
Gill, R. S. et al. Multicenter validation of a deep learning detection algorithm for focal cortical dysplasia. Neurology 97, E1571–E1582 (2021).
Urbach, H. et al. “Within a minute” detection of focal cortical dysplasia. Neuroradiology 1, 1–12 (2021).
Thesen, T. et al. Detection of epileptogenic cortical malformations with surface-based MRI morphometry. PLoS One 6, 16430 (2011).
Hong, S. J. et al. Automated detection of cortical dysplasia type II in MRI-negative epilepsy. Neurology 83, 48–55 (2014).
Ahmed, B. et al. Cortical feature analysis and machine learning improves detection of ‘MRI-negative’ focal cortical dysplasia. Epilepsy Behav. 48, 21–28 (2015).
Adler, S. et al. Novel surface features for automated detection of focal cortical dysplasias in paediatric epilepsy. NeuroImage Clin. 14, 18–27 (2017).
Jin, B. et al. Automated detection of focal cortical dysplasia type II with surface-based magnetic resonance imaging postprocessing and machine learning. Epilepsia 59, 982–992 (2018).
Nöth, U. et al. Detection of cortical malformations using enhanced synthetic contrast images derived from quantitative T1 maps. NMR Biomed. 33, e4203 (2020).
Maiworm, M. et al. Improved visualization of focal cortical dysplasia with surface-based multiparametric quantitative MRI. Front. Neurosci. 14, 622 (2020).
David, B. et al. External validation of automated focal cortical dysplasia detection using morphometric analysis. Epilepsia 62, 1005–1021 (2021).
Rácz, A. et al. Post-surgical outcome and its determining factors in patients operate on with focal cortical dysplasia type II—a retrospective monocenter study. Front. Neurol. 12 (2021).
Wagner, J. et al. Focal cortical dysplasia type IIb: completeness of cortical, not subcortical, resection is necessary for seizure freedom. Epilepsia 52, 1418–1424 (2011).
Wagner, J., Weber, B., Urbach, H., Elger, C. E. & Huppertz, H. J. Morphometric MRI analysis improves detection of focal cortical dysplasia type II. Brain 134, 2844–2854 (2011).
Salemdawod, A. et al. Predictors of postoperative long-term seizure outcome in pediatric patients with focal cortical dysplasia type II at a german tertiary epilepsy center. J. Neurosurg. Pediatr. 29, 83–91 (2021).
Engel, J. J., Van Ness, P. C., Rasmussen, T. B. & Ojemann, L. M. Outcome with respect to epileptic seizures. in Surgical Treatment of the Epilepsies 609–621 (Raven Press, 1993).
Schuch, F. et al. An open presurgery MRI dataset of people with epilepsy and focal cortical dysplasia type II, OpenNeuro, https://openneuro.org/datasets/ds004199/ (2023).
Kwan, P. et al. Definition of drug resistant epilepsy: Consensus proposal by the ad hoc Task Force of the ILAE Commission on Therapeutic Strategies. Epilepsia 51 (2010).
Urbach, H., Mast, H., Egger, K. & Mader, I. Presurgical MR imaging in epilepsy. Clin. Neuroradiol. 25, 151–155 (2015).
Wang, Z. I. et al. Voxel-based morphometric magnetic resonance imaging (MRI) postprocessing in MRI-negative epilepsies. Ann. Neurol. 77, 1060–1075 (2015).
Gorgolewski, K. J. et al. The brain imaging data structure, a format for organizing and describing outputs of neuroimaging experiments. Sci. Data 3, 1–9 (2016).
Cox, R. W. et al. A (sort of) new image data format standard: NIfTI-1. Neuroimage 22 (2004).
Li, X., Morgan, P. S., Ashburner, J., Smith, J. & Rorden, C. The first step for neuroimaging data analysis: DICOM to NIfTI conversion. J. Neurosci. Methods 264, 47–56 (2016).
Gulban, O. F. et al. poldracklab/pydeface: v2.0.0. Zenodo https://doi.org/10.5281/zenodo.3524401 (2019).
Jenkinson, M., Beckmann, C. F., Behrens, T. E. J., Woolrich, M. W. & Smith, S. M. FSL. Neuroimage 62, 782–790 (2012).
Blair, R. et al. bids-validator. Zenodo https://doi.org/10.5281/zenodo.6391626 (2022).
Funding
Open Access funding enabled and organized by Projekt DEAL.
Author information
Authors and Affiliations
Contributions
F. Schuch: data collection and data analysis, data curation, quality control, obtaining consent forms, data checking, manuscript preparation and revision. L. Walger: data preparation, data processing, data checking, manuscript preparation. M. Schmitz: data collection. B. David: data preparation, data processing. T. Bauer: data collection, figure preparation. A. Harms: data preparation. L. Fischbach: data collection, obtaining consent forms. F. Schulte: data collection, obtaining consent forms. M. Schidlowski: obtaining consent forms, manuscript preparation. J. Reiter: data collection, obtaining consent forms. F. Bitzer: data collection, obtaining consent forms. R. von Wrede: data collection. A. Rácz: data collection. T. Baumgartner: data collection. V. Borger: data collection. M. Schneider: data collection. A. Flender: study design and conceptualization. A. Becker: data collection. H. Vatter: data collection. B. Weber: data collection. L. Specht-Riemenschneider: study design and conceptualization, legal advice, manuscript revision. A. Radbruch: study design and conceptualization, data collection. R. Surges: study design and conceptualization, data collection. T. Rüber: study design and conceptualization, design of data sharing plan, data collection supervision, review and final approval of the version submitted for publication.
Corresponding author
Ethics declarations
Competing interests
RvW has received travel support, fees as speaker or for serving on the advisory board from Angelini, Apocare, Arvelle, Cerbomed, Desitin, Eisai, GW pharmaceuticals-JAZZ pharma and UCB Pharma. These activities were not related to the content of this manuscript. ARad serves on the scientific advisory boards for GE Healthcare, Bracco, Bayer, Guerbet, and AbbVie; has received speaker honoraria from Bayer, Guerbet, Siemens, and Medscape; and is a consultant for, and has received institutional study support from, Guerbet and Bayer. RS has received fees as speaker or consultant from Angelini, Arvelle, Bial, Desitin, Eisai, Janssen-Cilag GmbH, LivaNova, Novartis, Precisis, UCB Pharma, UnEEG and Zogenix; these activities were unrelated to the work of this manuscript. The other authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Schuch, F., Walger, L., Schmitz, M. et al. An open presurgery MRI dataset of people with epilepsy and focal cortical dysplasia type II. Sci Data 10, 475 (2023). https://doi.org/10.1038/s41597-023-02386-7
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41597-023-02386-7
- Springer Nature Limited