# Test 1 ## Purpose The purpose of this test was to see how accurately [the psychophysics plugin](https://jspsychophysics.hes.kyushu-u.ac.jp/) could present the stimuli. ## Methods [The test program](https://www.hes.kyushu-u.ac.jp/~kurokid/jspsychophysics/demos/accuracy_check_20240620.html) presented two white rectangles for 50 ms/3 frames and an audio file for 100 ms, and measured actual onsets and offsets of these stimuli using [the Black Box Toolkit](https://www.blackboxtoolkit.com/index.html). ![enter image description here][1] This program was coded using [jsPsych](https://www.jspsych.org/latest/) and [the psychophysics plugin](https://jspsychophysics.hes.kyushu-u.ac.jp/). The presentation time for visual stimuli was specified in terms of milliseconds (ms) or frames. The audio file was pre-created to have a length of 100 ms. The psychophysics plugin has [Pixi mode](https://jspsychophysics.hes.kyushu-u.ac.jp/pixijs/) in which visual stimuli are presented using the [PixiJS](https://pixijs.com/) library. ### Apparatus - The Black Box Toolkit (Sensor sensitivity was identical across conditions) - jsPsych V.7.3.4 - Psychophysics plugin V.3.7.0 - PixiJS V.6.5.8 ### Windows laptop PC - hp ProBook 430 G3, Intel Core i5-6200U CPU 2.3GHz 2.40GHz 64bit - 8GB RAM - Intel(R) HD Graphics 520, 1366 x 768, 60.059 Hz 8bit - Windows 10 Pro 22H2 - Chrome 126.0.6478.115 - Firefox 127.0.1 - Edge 126.0.2592.68 ### Microsoft Surface Laptop 4 (15-inch) - 11th Gen Intel(R) Core(TM) i7-1185G7 @ 3.00GHz 64bit - 16.0GB RAM - Intel(R) Iris(R) Xe Graphics, 2496 x 1664, 59.98 Hz, 8bit - Windows 11 Pro - Chrome 126.0.6478.127 - Firefox 127.0.1 - Edge 126.0.2592.68 ### iMac (Retina 5K, 27-inch, Late 2015) - 3.2 GHz quad core Intel Core i5 - 16GB 1867 MHz DDR3 - AMD Radeon R9 M390 2 GB, 2560 x 1440 - Mac OS 12.7.5 - Chrome 126.0.6478.127 - Firefox 127.0.1 - Safari 17.5 (17618.2.12.111.5, 17618) ## Results - [Results of hp Laptop](https://www.hes.kyushu-u.ac.jp/~kurokid/accuracy_check_20240626/hp_notePC.html) - [Results of Surface](https://www.hes.kyushu-u.ac.jp/~kurokid/accuracy_check_20240626/Surface.html) - [Results of iMac](https://www.hes.kyushu-u.ac.jp/~kurokid/accuracy_check_20240626/iMac.html) In all the computers, Firefox failed to present visual stimuli on several trials. Therefore, its use is not recommended for experiments with brief presentation times of less than 50 ms/3 frames. See yellow marker on the firefox sheets of each xlsx file for details. I had the same failure in Safari, but the second measurement was fine. See the iMac_20240626.xlsx for details. Never failed on Chrome and Edge. For Chrome and Edge, the accuracy was higher when the time was specified in frames than in ms. Especially, the accuracy was poor for the first rectangle when the time was specified in ms in the Pixi mode. Perhaps it took time to present the first visual stimulus in the Pixi mode. **This issue was examined in detail in Test 2 below.** It has been reported that it is difficult to present the visual and audio stimuli at intended SOAs, and the issue remains at this time. # Test 2 After Test 1, I noticed that PixiJS might take a long time to initialize and affected the stimulus presentation right after the start of the trial. (In my plugin, PixiJS is initialized for each trial.) Therefore, I predicted that if there is a delay before stimulus presentation, the stimulus will be presented accurately. ![enter image description here][2] ## Results [As I predicted, a delay (500 ms) improved the accuracy of the first visual stimulus.](https://www.hes.kyushu-u.ac.jp/~kurokid/accuracy_check_20240626/Summary_of_Test2.html) In a normal experiment, the fixation point would improve the accuracy of the first visual stimulus. [1]: https://files.osf.io/v1/resources/pj4sb/providers/osfstorage/667cd30bd6a72902270b1299?mode=render [2]: https://files.osf.io/v1/resources/pj4sb/providers/osfstorage/667cfa67b0ed59026979e55b?mode=render