VALIDATION DATASET: Opening the Black Box: Microsecond-Accurate Display Timing in Behavioral Research (APTEN Trace)

This deposit contains the full validation package for the manuscript "Opening the Black Box: Microsecond-Accurate Display Timing in Behavioral Research" (submitted to Behavior Research Methods). The paper introduces APTEN Trace, a kernel-level Event Tracing for Windows (ETW) monitor that instruments the DxgKrnl display pipeline for trial-level stimulus timing validation in psychology and neuroscience experiments. The validation reported in the manuscript is based on three oscilloscope-validated studies across three displays (60 Hz, 120 Hz, and 240 Hz; 8,000 trials total), with cross-vendor confirmation on NVIDIA, AMD, and Intel GPUs. This archive contains every raw capture, derived event table, and analysis script required to reproduce all reported statistics and figures independently of the APTEN Trace measurement engine. Contents data/raw/rigol/ — raw oscilloscope captures (Rigol DS1054Z, 4 channels at 1 GSa/s) data/raw/etw/ — raw ETW traces from APTEN Trace sessions data/derived/events/ &mdashduced by the oscilloscope/ETW analyzer (Weibull-fit photodiode onsets, T10/T50/T90 threshold times, hardware/software timestamp data/derived/figures/ — figure assets used in the manuscript scripts/ — analysis scode> (main statistical analysis), etw_oscilloscope_analyzer.py (events extraction), and standalone validato vs. logistic comparison, threshold-sensitivity analysis, frame-slip concordance, and study-3 dual-mode cross-val config/studies_brm.yaml — curated study definitions mapping each reported study to its raw and derived f paper/ — submitted manuscript PDF, Supplementary S4 (Weibull CDF vs. logistic sigmoid head-to-head), table apten_helper.py — int to install dependencies, verify the toolchain, knit the R analysis, and re-run any included Python validator from stu README.md — installation and reproduction guide Quick start From the package root, run: python apten_helper.py and follow the menu. The helper installs Python and R dependencies, runs a quick verification, and exposes the v a single interface. Reviewers without R installed can still execute all Python-based validators. Reproducibility scope All numerical results, tables, and figures reported in the manuscript are reproducible from the include The APTEN Trace measurement engine itself is not open-source at this time; however, the manuscript provides specification (ETW provider GUIDs, payload structures, ring-buffer architecture, three-stage matching pipeline, and auto-calibration algorithm) at a level of detail sufficient for independent re-implementation. Citation If you use this dataset, please cite the accompanying manuscript (citation will be updated upon publication) López Córdoba, D., et al. (2026). APTEN Trace: Validation package for kernel-level display timing inst Windows. Zenodo. https://doi.org/10.5281/zenodo.20090700 License Code released under the MIT License. Data CC-BY 4.0. See LICENSE files inside the archive.