Perception-Based Face Similarity Rating Task

Overview

Participants are presented with pairs of faces and are asked to rate the visual similarity between them using a 7-point scale (1 means very different and 7 means very similar).

The stimuli consist of 57 unfamiliar face images, leading to a total of 1,596 pairs of faces in the task. Each face pair is presented once in a random order and is divided equally over 14 blocks.

If you use this task, please cite our work:

• Chang, C. H., Nemrodov, D., Lee, A. C., & Nestor, A. (2017). Memory and perception-based facial image reconstruction. Scientific reports, 7(1), 1-9. https://doi.org/10.1038/s41598-017-06585-2

• Chang, C.-H., Nemrodov, D., Drobotenko, N., Sorkhou, M., Nestor, A., & Lee, A. C. H. (2021). Image reconstruction reveals the impact of aging on face perception. Journal of Experimental Psychology: Human Perception and Performance, 47(7), 977–991. https://doi.org/10.1037/xhp0000920

For more details of the task, see the articles above (Experimental procedures in Chang et al. 2017 or Experimental Procedures—Main Task in Chang et al., 2021).

System Requirements

Operating system

• This task should work on Microsoft Windows, MacOS and Linux. However, it has only been tested on Windows 10 and MacOS 10.14 (Mojave; intel chip)

Software

• 64-bit version Matlab (the task has been tested on Matlab 2019a)

• Psychtoolbox (the task has been tested on Psychtoolbox 3.14)

How to Run the Task

Preparation

• Put the stimuli (57 unfamiliar face images) in the stims folder and run the create_imgMat.m script to create a mat file (e.g., ims_new.mat) that stores all face images.

  • Due to copyright restriction, face images are not provided here, but you can find information of the stimuli (unfam1 to unfam57) used in our work (Chang et al., 2017, 2021) in the excel file stims_names_match.csv.

• To create all pairs of faces and randomize trial order in advance for a given participant, type the following codes in the Matlab Command Window

subID = 'ips99'; % subject's ID (including group label and number)
studyID = 'ips'; % the ID of the study
% You can change subID and/or studyID

% Generate pairs of faces and randomize trial order for a given participant.
tpOrder(subID, studyID) 
% the mat file is stored in the folder order_tp_xxx, depending on the studyID you enter

% You can also create a for loop to generate trial order files for multiple participants. 

Run the face similarity rating task

• To run the face similarity rating task, type FaceSimPercep in the Matlab Command Window

  • If you want to run the task in debugging mode (i.e., the background will be transparent rather than black, which allows you to see Matlab messages/errors during the task), open the FaceSimPercep.m script and change the variable debugMode (line 20) from 0 to 1.

• A Matlab dialog box will appear (see the attached screenshot below) and you can enter information of the study, participant, current session, and determine whether or not including a practice block.

  

Results

• The results of each participant are stored in the folder results_ips ("ips" is the study ID you enter in the dialog box)