Code associated with Aberra AS, Lopez A, Grill WM, Peterchev AV. (2023). "Rapid estimation of cortical neuron activation thresholds by transcranial magnetic stimulation using convolutional neural networks". Neuroimage
- Install python dependencies (recommended to install into a virtual environment)
- Add path to python executable to
python_exec.m - Download test dataset from doi.org/10.5281/zenodo.7326394 (~5.6 GB)
- Run init.m to set up paths to dependencies (set download_test_dataset = 0 if step 2 is skipped)
- OPTIONAL: To interpolate E-field vectors at all sampling grids from test dataset, run interpEfieldSample_all.m. Note: Running interpolation for all 15 neuron models, 12 rotations, and ~5,000 positions serially on a single CPU would take several weeks and >40 GB storage, parallelizing on high-performance computing cluster strongly recommended.
interpEfieldSample.muses all available CPUs to parallelize within model across all positions. Interpolated Efields for all neuron models at a single azimuthal rotation included in the downloadable dataset (found indnn_neuron_stim/output_data/nrn_efields/layer_set_1/M1_PA_MagVenture_MC_B70_ernie/nrn_pop1_maxH)
To generate Fig. 3, Supp. Fig 1, and Supp. Fig 3, see plot_Fig3_panels.m, plot_SuppFig1.m, and plot_SuppFig3.m, respectively in plot_manuscript_figs/
To run threshold estimation on single neuron model/position, see plot_manuscript_figs/run_estimation_single_neuron.m script
Dependencies:
Python: keras/tensorflow h5py scipy numpy
MATLAB toolboxes:
Parallel Computing Toolbox
version >r2022a, specifically for:
exportgraphics.m
>2020b:
for max/mean/min/median functions with 'omitnan' flag (vs. nanmax)
FEM E-field simulation was conducted in SimNIBS v3.1 (https://simnibs.github.io/simnibs/build/html/index.html). The necessary MATLAB library functions to load/save/process output files are included in this repository (written by Andre Antunes and Axel Thielscher). Original code can be found at the SimNIBS github repository: https://github.com/simnibs/simnibs/tree/master/simnibs/matlab
NEURON model data derived from models adapted from Blue Brain Project as part of Aberra AS, Peterchev AV & Grill WM (2018). Biophysically Realistic Neuron Models for Simulation of Cortical Stimulation. Journal of Neural Engineering 15, 066023. Original model code can be found at https://senselab.med.yale.edu/ModelDB/ShowModel?model=241165#tabs-1