Developing cognitive tests/tasks for training

[rustyriley]

Hi, am interested in trying to develop programs for cognitive neuropsychological tests. I've heard of someone who developed a hand movements memory test using MetaCard, before it was bought by Runtime Revolution, that I'd be interested in making "current".

There's also an online form of test of working memory similar to what I want, here:
www.aboutworkingmemory.org

Forms of other tests I'd like to develop are depicted here:
http://www.cantab.com/science/executive-function.asp
(don't forget to open the box at the top of the page).
Particular tests I'd be interested in include (from visual memory, executive function, and response control) include:
Delayed Matching to Sample
Spatial Recognition Memory
One Touch Stockings of Cambridge
Spatial Span
Affective Go/No-Go
CAmbridge Gambling Task (or Iowa Gambling Task, check Google )
Stop Signal Task

The tests from CANTAB are REALLY!! expensive I've heard.

Some software I've seen talks about "variable debounce periods" not being available using ordinary keyboards, but since I know absolutely NOTHING about all this, don't know what the significance behind this is.

Be grateful for any information/guidance on all this.

Budgetary constraints seem destined to drive me to either do it myself, or go totally away from this area -- which is a pity. I'm researching risk prediction with drink drivers, and paper and pencil measures just don't seem close enough to the mark (good sensitivity, but inadequate specificity, due to common problems across a number of different psychological problems),
Again, really appreciate any guidance / sharing of experience people could offer,
Cheers,
Russell
Wanganui, New Zealand

[Randy Hengst]

I have no guidance or specific information for you, but I'd be interested in exploring possibilities. Are there public domain descriptions of what the tests you reference contain? I assume the tests you've linked to as examples are under copyright.

[Mark]

Hi rustyriley,

I did some scientific stuff in Revolution. Sometimes it is hard, sometimes you can't believe how easy it is.

Can you be a little more explicit about what you want to do? Perhaps describe the steps of one particular test?

Best,

Mark

[oliver.hardt]

I've been programming experimental tasks as these for 10 years using first Pascal on the Apple Macintosh, and then Hypercard predominantly due to developing time, and for a couple of years now in runrev.
In the area of neuropsych testing, I implemented a version of the Wisconsin Card Sorting Test. As all neurocog tests have been published in scientific journals, including the tests that you mention in your list, copyright is not an issue.
As one of the posters before me said, given the information you provided, one cannot a priori determine how long it will take to develop a task. However, all these tasks will share a lot of code, such as code for getting demographical data, displaying instructions, storing the results in a database, and, for some tests, creating pseudo-randomized stimulus lists.
It will depend on how pressed you are to get these tasks computerized and how likely it is that you will have to maintain them and to develop new ones if it is worthwhile for your research to learn runrev. If you have never programmed before, realistically speaking it may take you anywhere between 1 and 3 months to develop an application, but you should recognize that this will be a focal area of your work time, as it is not done with 2 hours investment per day. Hence, you might also consider paying someone to perhaps code a task for you in runrev, and so you have a model or an example how to do it.
Best,
Olli.