Neural Noise- Stimulation mapping combined with whole-brain modeling reveals gradients of excitability and recurrence in cortical networks (Nature).
- The approach uses intracranial EEG (iEEG) plus stimulation to inform where and how to stimulate and decode for neuromodulation and BCI.
- Findings support trial and target selection in clinical neurophysiology and BCI. 1
Weekly enrichment (2026-07-20)
- The primary study appeared in Nature Communications (s41467-025-58187-6) and drew on 36 patients with drug-resistant focal epilepsy undergoing 323 presurgical intracerebral electrical stimulation (iES) sessions with simultaneous stereo-EEG (sEEG) and scalp high-density EEG.2
- Analyses revealed an anatomical excitability gradient: iES-evoked responses were stronger in high-order association cortex than in low-order sensorimotor regions (hd-EEG low- vs high-order W = 5713, p = 0.0008; sEEG W = 4567.5, p = 0.0013).2
- The team fit a connectome-based whole-brain model using Jansen-Rit neural-mass dynamics across 200 regions (Schaefer 200 parcellation) to each patient’s hd-EEG data.2
- Responses were characterized across seven canonical resting-state networks spanning limbic, somatomotor, dorsal-attention, default-mode, and frontoparietal systems.2
- An in-silico “virtual dissection” that suppressed extrinsic connections extinguished the excitability gradient, indicating it depends causally on recurrent feedback from non-stimulated regions.2
- High-order networks showed more inter-network functional integration, whereas low-order networks exhibited more segregated, localized processing.2
- Statistical comparisons used Wilcoxon-Mann-Whitney U tests benchmarked against a null distribution of 1000 time-wise permutations at p < 0.05.2
- Clinical/BCI implication: mapping where recurrent feedback dominates could inform target and trial selection for neuromodulation therapies and stimulation-based interfaces; a preprint version is available on bioRxiv.2 3