Neural Noise- A multimodal intracranial dataset for mesoscale insights in epileptic networks was reported in Nature (2025).1
- The dataset supports mesoscale network analysis and iEEG methodology and has been released to support replication.1 1
Weekly enrichment (2026-07-20)
- The dataset comprises intracranial EEG (iEEG), local field potentials (LFP), and multiunit activity (MUA) recorded from 96-channel Utah microelectrode arrays (Blackrock), released as a Scientific Data descriptor (Sci Data 12, 774, 2025).2
- Recordings capture 12 seizures across five drug-resistant focal epilepsy patients (ages 23, 60, 26, 31, and 57; three men, two women) evaluated presurgically at University Hospitals Leuven (KU Leuven).2
- Utah arrays were implanted for 14–16 days alongside clinical depth and ECoG electrodes; three sat in occipitotemporal cortex (patients 2, 4, 5) and two in dorsolateral prefrontal cortex (patients 1, 3), at varying distances from the seizure onset zone.2
- iEEG was sampled at 1024 Hz (256 Hz for two seizures in patient 1), LFP at 1000 Hz, and MUA acquired at 30 kHz then downsampled to 1000 Hz, enabling cross-scale analysis.2
- Ictal high-frequency oscillations were detected in the 80–250 Hz band with a Hilbert detector using a 5-SD envelope threshold and a ≥10 ms minimum duration, followed by visual inspection to remove false positives.2
- Epochs generally span 10 minutes before to 5 minutes after seizure onset, with a photodiode synchronizing the microelectrode array and iEEG neural processors during both task and overnight recordings.2
- Files are distributed as .mat and EDF per-seizure iEEG/LFP/MUA structures plus preoperative T1 MRI, postoperative CT, MNI electrode coordinates, and impedance logs; localization used SPM, FreeSurfer, and the iElectrodes toolbox.2
- The dataset is openly hosted on EBRAINS (v2, DOI 10.25493/ZC9M-2ES) to support replication of mesoscale epileptic-network and seizure-prediction analyses.3