Licking-analysis

install ffmpeg

https://cyber-tenchou.com/tips/20200904/
https://eng-entrance.com/linux-vi-save
vi .zshrc
Write: export PATH=/Applications:$PATH
To save and exit: :wq

Viveo data handling

Making masks

FFMPEG

Check the number of total frames

ffprobe -v error -select_streams v:0 -count_packets -show_entries stream=nb_read_packets -of csv=p=0 /home/wanglab/Desktop/build-CameraViewer-Qt_Static-Debug/Phox2b_#14_teLC_Choco_20220908_1_350fps0.mp4

Extract frames from the video (original or tracked/labeled video).

VLC can extract frames but it drops frames. ffmpeg does a better job.
Change directory to were you want to output frame images and run this (bmp is better than png)
ffmpeg -r 1 -i /home/wanglab/Desktop/build-CameraViewer-Qt_Static-Debug/Phox2b_#14_teLC_Choco_20220908_3_350fps0.mp4 -r 1 scene%04d.bmp

Export desired frames

Frame by frame
ffmpeg -r 1 -i /home/wanglab/Programs/Hourglass/build/Phox2b_#14_teLC_Choco_20220908_3_350fps1_labeled.mp4 -vf trim=start_frame=40000:end_frame=45000 -r 1 scene%04d.bmp

Frame by frame + increase brightness
fmpeg -r 1 -i /home/wanglab/Programs/Hourglass/build/Phox2B#8_20220810_3_1_labeled.mp4 -vf eq=brightness=0.3:contrast=1.5 -r 1 scene%04d.png

Every 2 frames
ffmpeg -i /home/wanglab/Desktop/build-CameraViewer-Qt_Static-Debug/Phox2b_#14_teLC_Sprinkles_20220919_1_500fps1.mp4 -vf "select=not(mod(n,2))" -vsync vfr scene%04d.png

bmp format produces a better image quality.

Create a movie from frames

ffmpeg -r 35 -f image2 -s 640X480 -start_number 51300 -i scene%04d.bmp -vframes 4000 -vcodec libx264 -crf 10 -pix_fmt yuv420p test.mp4
or
ffmpeg -r 35 -f image2 -s 640X480 -i scene%04d.bmp -vcodec libx264 -crf 10 -pix_fmt yuv420p test.mp4
See: https://hamelot.io/visualization/using-ffmpeg-to-convert-a-set-of-images-into-a-video/

-r is the framerate (fps).
-crf is the quality, lower means better quality, 15-25 is usually good.
-s is the resolution.
-pix_fmt yuv420p specifies the pixel format, change this as needed.

Concatenate videos

make a text file 'video.txt' List the files to concatenate like this: file 'Phox2b_#19_teLC_20221205_1_350fps1.mp4' file 'Phox2b_#19_teLC_20221205_2_350fps1.mp4'

Then run this command from the directory that contains video files and video.txt: ffmpeg -f concat -safe 0 -i video.txt -c copy output.mp4

Slowing down an exciting video

ffmpeg -y -i /home/wanglab/Programs/Hourglass/build/Phox2B#8_20220720_1_1_labeled.mp4 -r 35 -s 640x480 -c:v libx264 -b:v 3M -strict -2 -movflags faststart /home/wanglab/Programs/Hourglass/build/Phox2B#8_20220720_1_1_labeled_slow.mp4

How to track the tongue and jaw using the Hourglass program

1. Open the licking_jaw_config.json (/home/wanglab/Programs/Hourglass/build/licking_jaw_config.json)
2. Specify the video file to be tracked.
   "videos": "/media/wanglab/jun/Licking/Phox2b_videos/Phox2b#8_IRt_TeLC/Phox2b_#8_teLC_20220808_1_350fps1.mp4",
3. Open Terminal
4. Change directory: /home/wanglab/Programs/Hourglass/build/
5. Run: ./Hourglass -d licking_jaw_config.json --train=false
   Check -h help to see options
6. Two .5 and one labeled video will be generated.

Video data analysis

1. extract the tongue area (areas) and jaw height (Jaw_heights) information using extract_h5.m. You can do this either by running the extract_h5.m function in the NumberOfLicksPerBout.m or extract_h5_batch.m.
2. Check the tracking quality.
3. Run: [peaks,locs] = findpeaks(smoothdata(areas,'gaussian', 12),'MinPeakDistance',17.5, 'MinPeakHeight',500); to get "peaks"
4. Save peaks.
5. Perform statistical analysis.

extract_h5.m

For video tracking data
Matlab function. Extract information from h5 files. Source file: h5 files generated by the tongue-jaw tracker that Paul generated.
Tongue: 0.h5
Jaw: 1.h5

NumberOfLicksPerBout.m

For video tracking data
For analyzing licking data from tracked video in Matlab.

Density_plot_Python_ver2.ipynb

For data produced in cell_count
For plotting scatter and density plots of neurons. It was originally written in Julia. It has been converted to Python.

LickParameters_Python_20230322.ipynb

For lickometer data
Python code: This code analyzes licking parameters from mat files generated by "Licking_Optogenetics_LEDCue.m" and save them in a csv file. The parameters analyzed by this code are: "Licks/Bout", "MultiLicks/Bout", "Bout Duration", "Frequency"

Convert_SessionData.m

For lickometer data
This code converts SessionData files generated by the behavior paradigm "LickToGetReward_Traing" and "LickToGetReward_5" into adjusted_SessionData files. In the adjusted files, the Cue start is adjusted to 0. This is for processing in Python. Updated: 2023-04-18

Plot_LickToGetReward_2023-04-19.ipynb & Plot_LickToGetReward_batch_2023-04-19.ipynb

For lickometer data
Plot data from LickToGetReward. Before running these codes, you must process SessionData by Convert_SessionData.m.

extract_lick_parameters_from_LickToGetReward_Training_2023-05-13.ipynb

For lickometer data
Extract lick_parameters from LickToGetReward_Training This script extracts licking parameters, including "mean_num_licksperbout", "mean_num_multi_licks", "mean_bout_duration", "mean_frequency", "Avg. Licks in Drinking period ", "LickThreshold". This script contains "Count_licks_in_Drinking.ipynb"