Research data supporting "Rapid, interpretable, data-driven models of neural dynamics using Recurrent Mechanistic Models"

The data is associated to the PNAS paper "Rapid, interpretable data-driven models of neural dynamics using Recurrent Mechanistic Models" (RMMs), which is freely available on: https://doi.org/10.1073/pnas.2426916122 There are three files containing data used to train a Hodgkin-Huxley RMM, an LP neuron RMM, and a PD neuron RMM. These correspond to Figures 2, 3, and 4 of the pre-print, respectively, as well as other figures in the SI appendix. Each .mat data file corresponds to one continuous recording of the membrane voltage and associated applied current of the neuron in question. For the HH dataset, V1 and I1 are the membrane voltage and applied current of the simulated HH neuron. The HH dataset additionally contains a variable "Temp" containing the trajectories of the ground truth conductances of the HH model over the course of the simulated recording. For the LP-PD experimental dataset, V1 and I1 are the membrane voltage and applied current of the LP neuron, and V2 is the membrane voltage of the PD neuron. For the PD experimental dataset, V1 and I1 are the membrane voltage and applied current of the PD neuron. The data was collected as described in the Methods section of the paper.