QUEST
July 8, 2000
If memory serves, the code supplied in the 1983 article is in BASIC. The VideoToolbox download page provides code in C. The PsychToolbox includes QUEST coded in MATLAB. Beau Watson's Psychophysica includes QUEST coded in Mathematica.
Denis Pelli
October 9, 1996
This QUEST package is historically descended from the algorithm that Beau Watson and I published some years ago.
Watson, A. B. and Pelli, D. G. (1979) The QUEST staircase procedure. Applied Vision Association Newsletter, 14, 6-7. (This is basically a long abstract.)
Watson, A. B. and Pelli, D. G. (1983) QUEST: a Bayesian adaptive psychometric method. Percept Psychophys, 33 (2), 113-20.
I would also encourage looking at Ewen's recent paper, which points out some elegant refinements that the QUEST package now supports.
King-Smith, P. E., Grigsby, S. S., Vingrys, A. J., Benes, S. C., and Supowit, A. (1994) Efficient and unbiased modifications of the QUEST threshold method: theory, simulations, experimental evaluation and practical implementation. Vision Res, 34 (7), 885-912.
Some C code was written in 1985 by A. "Beau" Watson and K. Baldwin, and revised and expanded in 1986 by Andrew Fitzhugh, under Beau's supervision. When I visited NASA for a sabbatical in 1987 I supervised a complete rewrite by EJ Chichilniski (now a well-known visual scientist) and "Phil" Felasfa Wodajo. I've also enhanced the code to do quantiles, which i've shown in an unpublished manuscript to be a more efficient algorithm than modes. However, the code is agnostic; it allows you to do a trial at any intensity whatsoever, will appropriately update the posterior pdf (probability density function) and compute its mode, quantile, or mean. I've used the code continuously since then for all my psychophysics. I've fixed all known bugs, over the years have code to a variety of tests for accuracy, etc. It's been a LONG time since any bugs have appeared, so, for most practical purposes, you may assume it's bug free.
I have given copies of this code privately to many colleagues over the years, but haven't published it, because we've never written a manual for it. It would appear that we may never get around to writing a manual, but so many people seem to find it useful even without a manual, and so many others keep requesting a copy of the code that it seems best to simply make the code publicly available, as is.
I would be very grateful to hear of any bugs that you can document, but I am not offering programming help beyond what's already here. I will keep the publicly available copy up-to-date, including any future bug fixs (though I expect them to be rare). I'll send a notice to the VideoToolbox mailing list when I post updates. So you may want to have your name added to that list. Just let me know: denis@psych.nyu.edu
Beau Watson's Mathematica version of QUEST is available and is described in an upcoming issue of Spatial Vision.