et_webcam_precision_experiment

Data Analysis for the Eye tracking precision experiment using webcam

More info on our bioRxiv: Web-based eye-tracking for remote cognitive assessments: The anti-saccade task as a case study

Reproducing the figures

Requirements: Python 3.10+ with pyxations, pandas, numpy, scipy, seaborn, statsmodels, pyarrow.

Figure 2 (accuracy assessment)

Run precision_experiment/ notebooks. Raw data is loaded from the BIDS structure built by pyxations.

Figure 3 (anti-saccade task)

From antisaccade_experiment/, run antisaccades_pyxations_blocked.ipynb end-to-end. The notebook:

1. Builds the BIDS dataset from raw_data/ and computes derivatives.
2. Preprocesses each subject per block (interpolation to 30 Hz, baseline subtraction, min-max normalization, mirroring, rejection of trials with |x| > 1.5).
3. Aggregates error rates and RTs across subjects, both for all blocks (panels B, C) and for the first half excluding block 1 (panels D, E).
4. Picks a representative subject and plots the pixel, degrees, and normalized views (panel A).
5. Saves the composite figure to result_plots/figure3.png and prints the Wilcoxon rank-sum tests reported in the paper.