Analysis work based on eye data processed by Michael's software on matlab.

This repository contains the functions, analysed datasets from each mouse and .mlx live scripts managing the analysis plots.

1. Run etGUI.m to analyse the pupil from the Michael's software.
2. Use getEyeData.m/getEyeData_dev.m to acess the filtered raw eye data.
3. Save them and, if you want to work on responses to direction gratings, use main_combineEyeDb.m.

Just one if you only want the analysis plots and data: Circular Statistics Toolbox (Directional Statistics) http://www.jstatsoft.org/v31/i10

If you want to run the codes that generate the datasets, you need access to the zserver at the cortex lab and the codes should run perfectly.

'analysis and plots' folder: contains mostly the .mlx live script files which organise the plots and the codes that generate them. 'functions' folder: contains the functions I've written. 'datasets' folder: these are processed datasets for distribution anaylsis to different stimuli and trace analysis to gratings which are saved from the live scripts' codes so you don't have to do that again.

1. The analysis was created on sparse_EyeDb_live.mlx (to sparse noise stimuli), spontaneous_EyeDb_live.mlx (to spontaneous stimuli), gratings_EyeDb_live.mlx (to direction gratings), and natMovies_EyeDb_live.mlx (to natural images).
2. These live scripts are rather simple: they all use the functions main_SparseNoise, main_gratings, main_natMovies, and main_spontaneous to extract filtered raw eye data from getEyeData_dev.m only in experiments with the specific stimuli present and they are saved as 'StimuliType_MouseID_EyeBallData.mat'.
3. The saved data then are classified as running vs stationary and histograms of running vs stationary are plotted in each live script.
4. Variables in the StimulusMouseID_X_EyeBallData.mat:

• Variables as nx1 size cells (n is the number of experiments):
  • area: the real pupil sizes (in px)
  • pupil: pupil size/10th percentile
  • x: gauss filtered x coordinate minus mean
  • y: gauss filtered y coordinate minus mean
  • ts: timepoints
  • ball: running speed
• combarea, combpupil, combx, comby, combts, combball: these are combined matrice from above cell arrays.

The processes are mainly organised in the combine_EyeDb_live.mlx, plot_comparison.mlx, and plot_comparison_std.mlx live script files. To reproduce the same results, you can follow this process:

1. Open combine_EyeDb_live.mlx live script:

• Run main_combineEyeDb.m function on each FRXXX_X_eyeDb.m db cell array.
• Save the corresponding variables into one 'MouseID_X_EyeBallData.mat' file.
• This function also plots all the variables with all the raw traces in individual trials responding to different gratings. -At the bottom, there are some old plots for running speed distributions.

2. Variables in the 'MouseID_X_EyeBallData.mat' file:
   • They are first obtained from getEyeData_dev.m in the main_combineEyeDb.m function and processed with getStimulusSweepsLFR.m:
     • MouseID_X_pupil: the pupil size/10th percentile and, afterwards, is reduced from baseline.
     • MouseID_X_info: information of the experiments, from ppbox.infoPopulateTempLFR().
     • MouseID_X_eyeX: X coordinate gauss filtered minus mean (not reduced from baseline).
     • MouseID_X_eyeY: Y coordinate gauss filtered minus mean (not reduced from baseline).
     • MouseID_X_deyeX: X coordinate gauss filtered minus mean and reduced from baseline afterwards.
     • MouseID_X_deyeY: Y coordinate gauss filtered minus mean and reduced from baseline afterwards.
     • MouseID_X_ball: running speed obtained from getRunningSpeed() and processed with getStimulusSweepsLFR().
     • MouseID_X_dball: running speed same as above but reduced from baseline.
     • MouseID_X_angle: angles calculated from the complex number of eyeX and eyeY.
     • MouseID_X_dangle: angles calculated from complex number of deyeX and deyeY.
     • MouseID_X_movement: eye movement displacement also from complex number of eyeX and eyeY.
     • MouseID_X_dmovement: eye movement displacement from complex number of deyeX and deyeY.
     • MouseID_X_velocity: MouseID_X_movement differentiated.
     • MouseID_X_dvelocity: MouseID_X_dmovement differentiated.
3. Open plot_comparison.mlx:
   • This file organises all the scatter plots.
   • The .mat files are loaded and plotted with the function plotSweepRespScatter() and angles are plotted with plotSweepRespScatterAngle().
   • The sections are:
     • Pupil size scatter plots for individual mice and combined.
     • Pupil size scatter plots for running vs stationary.
     • Eye X scatter plots for individual mice and combined.
     • Eye Y scatter plots for individual mice and combined.
     • Running speed scatter plots for individual mice and combined.
     • d running speed scatter plots for individual mice and combined.
     • eye movement scatter plots for individual mice and combined.
     • Eye movement velocity scatter plots for individual mice and combined.
     • Eye movement angles scatter plots for individual mice and combined.
     • Eye movement dangles scatter plots for individual mice and combined.
     • There's also a polar subplot at the end which is created by plotSweepRespPolarAngle().
4. Open the plot_comparison_std.mlx:
   • This file organises all the trace plots across timecourse.
   • In this live script, the first half is much longer as it doesnt have a specific plotting function and the next half uses plotSweepRespRun() and plotSweepRespAxisTrace().
   • The variables are consistently called as 'responses_MouseID_X' in every section.
   • The sections can be divided into 3 parts: 1. trace across timecourse for individual mice and combined, 2. running vs stationary traces, 3. eye movement trace.
   • First part sections:
     • Pupil size trace plot (10 mice in one plot) and combined trace plot.
     • eyeX (absolute value).
     • eyeY (absolute value).
     • deyeX (absolute value).
     • deyeY (absolute value).
     • dmovement.
     • velocity.
     • angle (circ_mean() used to calculate the mean instead of nanmean()).
     • dangle (circ_mean() used to calculate the mean instead of nanmean()).
     • dball.
     • ball.
   • Second part sections (uses plotSweepRespRun()):
     • Pupil size, running vs stationary (threshold at 5 cm/s).
     • deyeX (abs).
     • deyeY (abs).
     • angle (input the last parameter as 1 and the function will use circ_mean() instead).
     • dmovement.
     • velocity.
   • Third part sections (eye movement trace averaged across trials):
     • 13 subplots of eye movement traces in one plot using eyeX and eyeY.
     • Eye movement trace plots using eyeX and eyeY, one plot for each mouse.
     • Eye movement trace plots using deyeX and deyeY, one plot for each mouse.

Check directory of videos.csv if you want the directories of some/all videos (it's also in 'videos to analyse' file in Sylvia's OneDrive folder). Or use getDirectoryAT.m and save_directories.mlx to do some edits.