Skip to content

Latest commit

 

History

History
127 lines (77 loc) · 13.4 KB

README.md

File metadata and controls

127 lines (77 loc) · 13.4 KB
Raw

Eyetracking

Useful functions for setting up Eyelink and plotting eye movements

Notes:

The code in this repo can be used when working with Eyelink eyetracking cameras and data files. The function eyetrackQA can only be run on Macs using MATLAB versions 2016 or later.

Contents:

  1. Functions to Setup Eyelink

  2. Functions to Plot Eye Movements

    1. The eyetrackQA function
    2. Example Data
    3. Wrapper Functions

Eyelink Setup

Under the EyelinkSetup folder, you will find functions that can be used to setup and calibrate Eyelink eye trackers.

  • setupEyelink_Projector: Initializes, setups, and calibrates Eyelink for eye tracking subjects looking at stimuli displayed on the projector screen at the CNI. Run this code before you start any functional scan requiring eyetracking.

  • setupEyelink_Bore: Initializes, setups, and calibrates Eyelink for eye tracking subjects looking at stimuli displayed on the bore monitor at the CNI. Run this code before you start any functional scan requiring eyetracking.

Plotting Eye Movements

Under the PlottingEyeMovements folder, you will find functions that can be used to plot eye movements, especially for subjects asked to fixate on a central dot. The most important of these functions is eyetrackQA which is described in more detail below.

The eyetrackQA Function

Outputs

eyetrackQA will take edf files and use them to plot the eye movements of subjects asked to fixate during various experiments (for example, retinotopy). In order for the code to run successfully, edf files must be organized under a subdirectory called "edffiles" under a main data directory ("dataDir") where you would like to store your data. When you run eyetrackQA, folders will be created within dataDir that will organize data and figures in the following matter:

  • dataDir/ascfiles/...

    • samples: Will store asc text files containing raw data from Eyelink. These files were created using Eyelink's edf2asc executable, and contain only sample data from the edf file. This asc file will have 4 columns of data: time, x position (in pixels), y position (in pixels), and pupil size. Any missing data (for instance, when the subject is blinking) will be labeled as NaN.

    • events: Will store asc text files containing raw data from Eyelink. These files were created using Eyelink's edf2asc executable, and contain only Eyelink messages and event data. The messages include information about calibration, validation and drift correction as well as trial start times. The events recorded by Eyelink include fixation, saccade, and blink times. For more information about how these data are organized, consult the Eyelink Manual.

  • dataDir/matfiles/...

    • raw: Will store a mat file containing data variable "raw". Raw is a 4 dimensional array where raw(:,1) stores time in miliseconds, raw(:,2) stores x position in degrees of visual angle from the center of the screen (dva), raw(:,3) stores y position in dva, and raw(:,4) stores the eye's distance from the center of the screen in dva.

    • event: Will store a mat file containing data variables "blinksFromEL", "blinksFromScript", "dataQuality", "numSaccades_ELBlinksRemoved","numSaccades_removeBlinksFun","saccadeLocs_ELBlinksRemoved","saccadeLocs_removeBlinksFun".

      1. blinksFromEL: Vector containing all of the times in raw when the subject was blinking according to the asc event file. To determine whether the subject was blinking, the asc event file was parsed to find the times when the subject was blinking as well as when the subject was saccading immediately before and after their blinks.

      2. blinksFromScript: Vector containing all of the times in raw when the subject was found to blink according to the removeBlinks function. The removeBlinks function looks for times when the subject's y position changed dramatically before and after timepoints containing missing data from when their eyes were closed. These eye movements were assumed to be associated with the subject blinking and are listed in this data variable.

      3. dataQuality: Struct with the fields "calibration_quality", "validation_quality", and "drift_quality". Each of these fields contain a character array characterizing the quality of Eyelink's calibration, validation and drift correction as "GOOD", "FAIR", "POOR", "ABORTED", "SUCCESSFUL", or "NO" in the event that the calibration, validation or drift correction was not run.

      4. numSaccades_ELBlinksRemoved: The number of saccades detected after excluding timepoints corresponding to blinksFromEL. The number of saccades are calculated by the function countSaccades, which counts the number of times the subject's eyes cross a thresholded distance from the center of the screen (default distance is 2 dva).

      5. numSaccades_removeBlinksFun: The number of saccades detected after excluding timepoints corresponding to blinksFromScript. The number of saccades are calculated by the function countSaccades, which counts the number of times the subject's eyes cross a thresholded distance from the center of the screen (default distance is 2 dva).

      6. saccadeLocs_ELBlinksRemoved: Vector containing all of the times when the subject's eyes crossed crossed some thresholded distance from the center of the screen set by the countSaccade function. These timepoints were calculated after blinksFromEL were excluded from the data.

      7. saccadeLocs_removeBlinksFun: Vector containing all of the times when the subject's eyes crossed crossed some thresholded distance from the center of the screen set by the countSaccade function. These timepoints were calculated after blinksFromScript were excluded from the data.

    • ELBlinksRemoved: Will store a mat file containing the data variable "processed_ELBlinksRemoved". This data variable takes the same shape as raw and will have all timepoints corresponding to blinksFromEL labeled as NaNs.

    • FuncBlinksRemoved: Will store a mat file containing the data variable "processed_removeBlinksFun". This data variable takes the same shape as raw and will have all timepoints corresponding to blinksFromScript labeled as NaNs.

  • dataDir/figures: Stores jpg images of the subject's eye position across the course of the experiment. These figures will have a colormap that displays how the subject's eye position corresponds to the time in the experiment. Additionally, the title will contain useful information regarding the quality of Eyelink's calibration, validation, and drift correction before the start of the experiment, as well as how many saccades were counted using the countSaccade function. These figures were all generated using the makeQAFig function.

    • raw: Plot of raw eyetracking data before any blinks were removed.
    • ELBlinksRemoved: Plot of eyetracking data after timepoints corresponding to blinksFromEL were labeled as NaNs.
    • FuncBlinksRemoved: Plot of eyetracking data after timepoints corresponding to blinksFromScript were labeled as NaNs.

Inputs

The eyetrackQA function only has 2 required inputs: fName and dataDir. fName is the name of your edf file, which must be located under dataDir/edfffiles in order for eyetrackQA to run properly. dataDir is the path to your data directory. However, eyetrackQA also has 7 optional Name, Value pair arguments that you will probably need to use. They are listed below:

  • eyetrackQA( fName, dataDir, 'RemoveBlinksFunction', ____ ): Logical input denoting whether you would like to use the removeBlinks function to label blinks in addition to using Eyelink's blink labeling. If false, you will not produce figures under dataDir/figures/FuncBlinksRemoved and will not store the data variables "numSaccades_removeBlinksFun" or "saccadeLocs_removeBlinksFun" under dataDir/matfiles/events. Default is true.

  • eyetrackQA( fName, dataDir, 'plotRaw', ____ ): Logical input denoting whether you would like to plot raw eyetracking data and save a jpg image under dataDir/figures/raw. The default is true.

  • eyetrackQA( fName, dataDir, 'experiment_screenshot', ____ ): File path to a screenshot of your experiment that you would like to appear behind the plot of eye movements. The default is an image from a checkerboard bar retinotopy stimulus, which you can find in this repo under Eyetracking/PlottingEyeMovements/ExperimentScreenShots/RecMemFixationAndBar.png.

  • eyetrackQA( fName, dataDir, 'pxlScrnDim', ____ ): Vector of length 2 containing the experiment screen width and height in pixels. Default is from the Recognition Memory Experiment ([1024 768]). This measurement is very important in order to accurately convert the subject's eye position from pixels to degrees of visual angle.

  • eyetrackQA( fName, dataDir, 'mmScrnDim', ____ ): Vector of length 2 containing the experiment screen width and height in millimeters. Default is from the Recognition Memory Experiment ([385.28 288.96]). This measurement is very important in order to accurately convert the subject's eye position from pixels to degrees of visual angle.

  • eyetrackQA( fName, dataDir, 'scrnDstnce', ____ ): Distance from the subject's eye to the experiment screen in milimeters. Default is from the Recognition Memory Experiment (540 mm). This measurement is very important in order to accurately convert the subject's eye position from pixels to degrees of visual angle.

  • eyetrackQA( fName, dataDir, 'SACCADE_THRESH', ____ ): The threshold used to count saccades in dva. Saccades will be counted using the countSaccade function anytime the subject's eye position crosses this threshold. The default is 2 dva.

  • eyetrackQA( fName, dataDir, 'length', ____ ): The length of your experiment in miliseconds. If no value is provided, will assume that the length of Eyelink's recording matches your experiment. If the length provided is less than the length of Eyelink's recording, extra recorded timepoints will be cropped.

Experiment Parameters

Below are a list of experiment setups for which the screen dimensions and distances have already been measured by AR and MN in 2019.

  • Eyetracking Room 454: This room is used for behavioral experiments outside of the scanner, such as the Recognition Memory experiment.

    • Experiment Screenshot - A screenshot of the meridian mapping portion of the Recognition Memory experiment can be found in this repo under Eyetracking/PlottingEyeMovements/ExperimentScreenShots/RecMemFixationAndBar.png
    • Pixel Screen Dimensions (Recognition Memory) - [1024 768]
    • Milimeter Screen Dimensions (Recognition Memory) - [385.28 288.96]
    • Screen Distance (milimeters) - 540

  • Scanner Bore Monitor: This is the television screen behind the scanner. It is used in most experiments in the lab.

    • Experiment Screenshot - A screenshot of a retinotopy experiment with a checkered bar can be found in this repo under Eyetracking/PlottingEyeMovements/ExperimentScreenShots/RetFixationAndBar.png. A screenshot from our kidLoc localizer experiment with an image showing a pseudoword can be found in this repo under Eyetracking/PlottingEyeMovements/ExperimentScreenShots/kidLocWord.png. A screenshot from our kidLoc localizer experiment of the fixation point with no image can be found in this repo under Eyetracking/PlottingEyeMovements/ExperimentScreenShots/kidLocFixation.png.
    • Pixel Screen Dimensions (Retinotopy) - [1920 1080]
    • Milimeter Screen Dimensions (Retinotopy) - [1040 585]
    • Screen Distance (milimeters) - 2620
    • Length (miliseconds, just for the retinotopy experiment) - 180000 (because it's not clear when the recording started and stopped in relation to the retinotopy experiment, we crop more timepoints than needed). You do not need this argument to analyze eyetracking data from the kidLoc experiment.

  • Projector Screen: This is the screen that is attached to the back of the 16 channel coil for experiments such as toonotopy

    • Experiment Screenshot - A screenshot from the toonotopy experiment is in this repo under Eyetracking/PlottingEyeMovements/ExperimentScreenShots/ToonFixationAndBar.png
    • Pixel Screen Dimensions (Toonotopy) - [1280 960]
    • Milimeter Screen Dimensions (Toonotopy) - [360 270]
    • Screen Distance (milimeters) - 300

Example Data

You can find an example of what dataDir will look like after running eyetrackQA under Eyetracking/PlottingEyeMovements/ExampleData.

Wrapper Functions

Most likely you will have many edf files that you would like to analyze. In this case, place all of these edf files in a folder named edfffiles that is a subdirectory of some data directory where you would like to store all of your eye tracking data (dataDir). By running eyetrackQAWrap( dataDir, 'Name', Value ) using the same Name, Value pair arguments as for eyetrackQA (see above), all edf files will be analyzed using the eyetrackQA function.