MATLAB code, demo, and sample data to analyze data for trajectory divergence as in [Russo et al Neuron 2020].
Author: Abigail Russo, May 2019
The data included in this demo were recorded in the Churchland lab at Columbia University. Single-unit neural data were recorded sequentially from a macaque monkey during the performance of a hand-pedaling task. Monkeys grasped a hand crank and cycled through a virtual environment for a number of prescribed cycles as indicated by a visual cue. The data included here correspond to two seven-cycle conditions (forward bottom-start and backward bottom-start for monkey C). For more information about data processing and the task, see Russo et al Neuron 2020.
Data should be formatted as for jPCA (Churchland et al 2012). Each data structure (e.g. D_m1, located in M1_sampleData.mat) contains C elements corresponding to the number of conditions (here, 2). The t x n matrix 'A' contains the trial-averaged firing rates as a function of time for each neuron. All other fields are optional. The 'times' field indicates the time course of each sample in A and the 'analyzeTimes' field indicates which of these times should be analyzed for tangling. For more information on formatting, see the divergeAnalysis code comments and jPCA documentation.
- SMA_sampleData.mat: trial-avergaged firing rates from 77 neurons (monkey C)
- M1_sampleData.mat: trial-avergaged firing rates from 116 neurons (monkey C)
- kinematic_sampleData.mat: trial-avergaged kinematic data corresponding to a variety of behavioral parameters such as hand position and velocity (monkey C)
- divergeAnalysis.m: compute trajectory divergence on data formatted as described above
- divergence_demo.m: visualize task kinematics, compute divergence on sample datasets, plot results as in Russo et al Neuron 2020, Figure 7.
- formattedScatter.m: makes pretty scatter plots
Supporting functions: AxisMMC.m (makes pretty axes, used by formattedScatter)