Screening Prep GUI Installation #

The Screening Prep GUI is found here at https://github.com/FriedCNL/ScreeningPrepGUI. Clone the repository or download the current batch. In order for the GUI to function correctly, it is important to have box set up on your computer as a local drive. Open the config.json files and add the box file paths respective to your pc.

Example config.json file:

{
“patientFolder”: “581/”,
“subjectsPath”: “/Users/addatta/ScreeningPrepGUI/Output/”,
“imgDBPath”: “/Users/addatta/Library/CloudStorage/Box-Box/ScreeningStimuli/images/”,
“imgDBCSVPath”: “/Users/addatta/Library/CloudStorage/Box-Box/ScreeningStimuli/CSVDatabases/imagesDB.csvv”,
“audioDBCSVPath”: “/Users/addatta/Library/CloudStorage/Box-Box/ScreeningStimuli/CSVDatabases/audiosDB.csvv”,
“audioDBPath”: “/Users/addatta/Library/CloudStorage/Box-Box/ScreeningStimuli/audios/”,
“videoDBCSVPath”: “/Users/addatta/Library/CloudStorage/Box-Box/ScreeningStimuli/CSVDatabases/videosDB.csvv”,
“videoDBPath”: “/Users/addatta/Library/CloudStorage/Box-Box/ScreeningStimuli/videos/”
}

 

As a check, open the internalconfig.json file and ensure both the gAPI Key is present. In case it isn’t, you can generate your own at https://pypi.org/project/Google-Images-Search/. The GUI also has a home-brew web scraper that can be used in case the API calls max out for the day.

Next we can set up a conda environment for the GUI:

$ conda create -n "ScreeningPrepGUI" python=3.10.0

$ conda activate ScreeningPrepGUI

Next, we need to install all the packages required to run it. Almost all have been included in the requirements file except for the yt-dlp function. this must be separately installed, renamed as yt-dlp, and placed into the bin folder within the ScreeningPrepGUI folder.

$ python -m pip install -r requirements.txt

Yt-dlp found here:

https://github.com/yt-dlp/yt-dlp/releases/download/2025.01.26/yt-dlp_macos.

 

Once all the packages are installed and the config file is filled out appropriately, we can run the program:

`$ python flask_backend.py`

Prior to starting stimuli selection, enter the subject name and update it based on which screening is being created. Also check that the paths look correct prior to starting collection.

Example:

 

Screening Stimuli Selection #

 

When creating stimulus for Screening 1, each category should have 25 images. Typically, images can represent more than one category, so you can end up having 25 images for each category and still not reach 100 images. In that case definitely add remaining stimuli based on the screening interview.

If you are in a position where you see the patient often, it is very valuable to continue conversing and gleaning more information out of the patient regarding their interests, as most of the time, patients do not give too much in the interview. A natural conversation is more likely to reveal underlying interests that may lead to finding concept neurons. If you do find new stimulus not mentioned before, please add it into the screening.

While searching for images, it is important to make sure that the image is only representing the concept you are testing for (unless it is a group). Ideally the image should have the concept in focus, if not it being the only thing on screen. Having too much in the image is not as valuable, as we are not sure what the patient may be responding to. It is also very helpful to avoid images which may be difficult for the patient to classify, as “wrong” answers can build frustration with the patient. In the case there is a stimulus which doesn’t definitively have a category, it is best practice to classify it as both. Typically patients will accept if they missed an existing category rather than us not including it. An important part of the instructions is to emphasize that even if there is a small bit of the category in the background, or in the corner of an image, it will count. 

 

Example Stimulus: Janine Teague (Abbott Elementary):

 

Unclear focus:

Clear focus:

 

Example Stimulus:

Here, there are plants in the background. Although it is not the focus of this picture, this picture would still be classified as having male, female, and plants in order to ensure uniformity among stimuli. Different people will make stimulus over the years, and they will have different classifications of images in their minds, which is why even if there is a small part of the image which has another category, it must be included.