Main content

Home

Menu

Loading wiki pages...

View
Wiki Version:
**NASCO 3.00** This program creates dot pictures according to your specifications. You can find on this repository: a) NASCO.mlapp, which is the NASCO app (coded in MATLAB). NASCO requires MATLAB (greater than 15a). b) NASCO_OSX.zip, which contains NASCO with the MATLAB runtime (compiled and working on OSX only). c) NASCO.m, which is the code source of NASCO. d) The raw data from the Guillaume et al. (JNC, 2020). There are four functionalities in NASCO: 1) Single array creation, 2) Array pair creation, 3) Automatised generation and automatised display. 4) Generate dot arrays from a CSV file. Please note that the display mode currently needs the installation of the Psychtoolbox-3 (www.psychtoolbox.org). You can modify any of the five (related) values: Number (N), Total Area (TA), Item size (IS), Convex Hull (CH), and Mean Occupancy (MO). As they are intertwined, any modification will affect the other values; but you can 'lock' one dimension if you do not want the latter to change. All values, except for N, actually correspond to a number of pixels. Note that: a) IS depends on the radius of the circle, and the script automatically corrects for the value of IS, b) CH is computed by creating target convex hull area and then filling the area by the remaining dots, which can take additional computational time if the space to fill the remaining dots is small. In the no pixel perfect mode, CH is computed through (many) iterations, and we thus need to accept some negligible errors (less than 1%) to get results in a reasonable amount of time. c) the color codes are RGB triplets. d) there is no upper limit to the Number value (NASCO adequately works up to a thousand dots, and it can work up to very BIG quantity such as 50,000 when image size is sufficiently large, but processing time is there very long to fill the dots in target convex hull). e) NASCO can generate arrays with less than 3 dots, ignoring CH and MO specifications when N equals to 1. NASCO considers a mixed measurement of CH and Field Area as CH when N equals to 2 in the no pixel perfect mode (in order to avoid the under- or overestimation of the extensive space) f) NASCO tries its best in full automatisation and in pair generation modes, but some combinations are not physically possible. This means that the generated pictures might (in some rare cases) not correspond to the wanted properties. This is mostly the cases for the values of IS (the dot size) because not every value are possible! Please systematically verify that the generated arrays properties actually correspond to what you expected before using them in your experiment. For the most precise control of dot properties, we recommend users to use the single generation mode. g) If you want to change the picture names, please avoid using '-' to avoid uninteded glitches (and rather use '_'). This script is completely free for use in scientific, educational, and diagnostic context. Please refer to the authors (Guillaume, Schiltz & Van Rinsveld, Journal of Numerical Cognition, 2020) when using it. Finally, in order to improve the script, please contact the first author if you encounter any problem or bug. Last update: **3.00.** NASCO will now create convex hull areas - by default - at pixel level precision. Patch note: 2.0. NASCO now have three functionalities: single creation mode (as in NASCO_1), pair generation mode, and full automatisation. 2.01 Corrected some colour issues specific to Windows (thanks to Arnaud V. for feedback). 2.20. New feature: NASCO can generate pictures based on a CSV file (thanks to Joanne E. for the suggestion). 2.21. Corrected some glitches with the automatic picture name labelling. 2.30. NASCO is now less strict regarding the required ratio between Item Size and Mean Occupancy. You can now create more dense arrays! *1.02. NASCO correctly saves the picture with the intended color background. 1.03. Small orthographical correction. 1.10. NASCO goes BIG. There is not anymore any upper limit for the Number value. We were able during our tests to generate up to 50,000 dots with the largest image size, but this has the disadvantage of requiring a lot of processing time. We recommend not to go further than 500 or 1,000 dots (as a function of the image size). Please note that you can at any time interrupt MATLAB processing with the CTRL + C combination keys, but this interruption will probably lead to some graphical bugs at the following start of NASCO. It is possible to correct these visual bugs by generating any dot array and then re-start NASCO. 1.11. NASCO should now use all available space to generate dot arrays (i.e., no more blank margins). 1.12. NASCO now handles the global form of the array and now displays dots arranged within a square or within a circle. 1.13. NASCO does not consider the CH anymore when generating exactly 2 dots, as CH is too narrow and it is not convenient to control for the expansion of the stimuli space. It rather takes into consideration a mixed parameter that is the mean of the narrow CH (which underestimates the space) and the field area (the smallest circle that encompasses the two dots, which overestimates the space). You can tick the 'Show CH' mark to verify how this mixed parameter looks like.* Guillaume, M., Schiltz, C., & Van Rinsveld, A. (2020). NASCO : A new method and program to generate dot arrays for non-symbolic number comparison tasks. Journal of Numerical Cognition, 6(1), 129‑147. https://doi.org/10.5964/jnc.v6i1.231