Sunday, February 28, 2010

Intellectual Property and Norms in PEBL

One of the most frequent questions that comes up is about whether PEBL violates intellectual property of others 'standard' tests.  There are really four closely-related issues here, which I wanted to lay out here.  They include (1) copyright (2) trademark (3) patents and (4) norm relevancy.

Copyright covers things like written material, artwork, source code, possibly instructions, and so on.  For experiments distributed with PEBL, these are all implemented afresh, either de novo  or based on descriptions from published articles.  There is no copyright here, because we don't use other test's source code or images.

However, published norms might be copyrighted.  Certainly, if published in a journal, the publisher owns copyright to the article, and possibly the data tables (although these might be subject to fair use or fact-based exclusions to copyright law).  This can potentially create some issues because even when norms for a particular test are copyrighted, PEBL may not be able to incorporate these within software code to give immediate feedback about performance percentile.  This gives one good reason to have some high-quality norms for PEBL's tests.

Things like trade names and logos are covered by trademarks and service marks.  A number of companies have implemented versions of tests they either designed themselves or which arose in the research community.  Their mark protects them from others using the name in ways for commerce that might confuse the public give them a bad name.  For example, somebody owns the name 'Wisconsin Card Sorting Test', as it applies to a set of paper materials used in neuropsych evaluation.  That mark doesn't apply to computerized versions, and I suspect it may not even be proper, because the original test, reported by Berg in the late 40s on primates, was probably in common use terminology for decades before the mark was registered.  Nevertheless, In all material I distribute, I am careful to state that my tests are workalikes of other tests when they are, and use different names (such as "Berg's card sorting test") in cases where there is a prominent commercial version, whether or not they have a registered trade name.  I see no issue related to trade names/ trade marks/ etc.

One thing I do see sometimes is researchers will publish a screenshot from PEBL in a report, but not cite PEBL as the source of the test.  In these cases, either people are using just the PEBL screenshots because they are easier to find and free to use (while using another version to test subjects), or they are not giving proper credit to PEBL.  

A third intellectual property issue is in regard to patents. Processes and apparatus can be patented, which would protect the inventor from re-implementations for a limited amount of time (I think up to 14 years).  Luckily, I don't think there are many tests out there that have been patented, and I think the research and clinical communities should shun those types of tests. Although there are many fine uses for patents, I view such patenting as anti-scientific in this context.  For a patent to be valid, a test needs to be patented within, I believe, 1 year of its first public dissemination, a restriction which prevents most new psych tests from being patented as they often arise in a research lab and they get discussed at conferences long before the university research offices know about them.  So, for example, even if the Wisconsin Card Sorting test had been patented, its patent would have expired in the 1960s.  Bechara & Damasio published details of the 'Iowa gambling task' in 1994, so even if it had been patented, that patent would have expired several years ago.   To my knowledge, I implement and distribute no tests that have any patents applied to them. 

The norms issue is a problem. There are currently no published norms for any of PEBL's tests.  My target audience is primarily the research community,  who can use the tests without caring about how a population performed.  Whether other norms could apply is up to the professional judgment of the user, and I would say that it depends on the test, and the context of use.  From reading neuropsych handbooks, I get the impression that even the 'standard' versions of tests often use norms from alternate versions, or are decades old (which I would bet certainly matters for things like attention tests).  Of course, some of these norms are themselves copyrighted, and so it might be difficult to use them even if they did apply.

I am interested in developing norms for some subset of the tests, at least for limited populations.  If you are a researcher with access to subject populations and resources to contribute, please sign up to the pebl-norms email list at
Post a Comment