Tethered data analysis¶

This page explains the conventions used for the tethered saccade data analysis and gives instructions on how to run the annotation and analysis.

The whole process goes like this:

Download the data files.
In Matlab, create annotations for QA. See section Data annotation.
In Matlab, run saccade extraction (command process_data). See section Saccade detection.
In Matlab, create the report figures (command create_report). See section create_report.
Use a python script to create the LaTeX report (command create_report) and compile it. See section Create the PDF report.

Data annotation¶

Download the data. Due to the size, it is not committed to the repository.:
```
$ cd data/
$ ./download_datafiles.sh
```
Fire up Matlab. Add all the directories to the path.:
```
>> addpath(genpath(pwd))
```
Start the annotation tool. The main function is annotate_species Give it as the only argument the species directory.
```
>> annotate_species('data/Dmelanogaster')
```

You are now inside the tool.

You are presented with a random slice of data. Click twice to mark a saccade: at the beginning and the end. When you have marked all saccades in the current screen, press enter to go to the next screen.

When you want to exit, press “x”+enter without clicking any point.
If you want to undo the previous selection, press “u” + enter.

Saccade detection¶

The main method to run the analysis is

process_data('data', conf)

Where conf is a struct whose fields are described next. Reference values are also shown.

conf.id: Configuration name. The processed data is created in data/<species>/processed/<conf.id>/
conf.saccade_detection_method = 'linear': linear is the only method that works well.
conf.smooth_steps = 3: to write
conf.filtered_velocity_significant_threshold = 50: to write
conf.filtered_velocity_zero_threshold = 15: to write
conf.min_significant_amplitude = 5: to write
``conf.filtered_velocity_significant_threshold = ... ``: to write

This command tries different thresholds.

>> test_different_thresholds

Create a symbolic link ‘use_for_report’ for the configuration we are going to use in the

Create the reort¶

Running the analysis:

>> create_report_data('data', 'threshold1')

How to recreate some pictures¶

Try this

$ rm data/D*/report/*hist_log*
>> create_report('data', 'use_for_report')
$ make -C data/report all.pdf

Create the PDF report¶

Run this python script:

$ python code/report/report_generation_utils/create_report_data.py data

And then compile the report:

$ make -C data/report all.pdf

Saccade analysis tools v1.0 documentation