Abstract
Cognitive neurogenetics is a genetics approach to the cellular and molecular neurobiology of cognitive mechanisms. It is predicated on the discovery (forward genetics)or production (reverse genetics) of mutant strains with a genetically based abnormality in a basic mechanism of cognition. The key to a successful prosecution of this approach is an effective behavioral screen for specific functional effects from genetic manipulations. An effective screen should have the following properties:
●● It should focus on specified mechanisms of cognition (e.g., odometry, interval timing, circadian timing), not on categories of phenomena (e.g., spatial learning, temporal learning, fear learning) or on experimental procedures (e.g., fear conditioning).
●● It should deliver functionally specific measurements, measurements that reflect the functioning of one particular mechanism, not the many different mechanisms that contribute to the measured strength or duration of a behavior.
●● It should give a clear phenotype for each animal screened, not a group average.
●● It should give physiologically meaningful measurements (e.g., the period of a circadian clock, the precision of interval timing), not quantities that have no physiological meaning (e.g., responses per minute).
●● It should make it possible to screen large numbers of animals in reasonable periods of time with a reasonable investment of investigator/technician time and cost, and with minimal handling of the mice.
●● It should archive the data in such a way that others can explore it in the light of future discoveries that pose new questions using old data. Put another way, it should contribute to a publically available and useable behavioral phenotype database.
●● It should produce an intact data trail that enables others to trace the connection back from published summary statistics and graphs to the raw data.
We have developed and successfully used such a system. A major part of the development was the creation of an open source Matlab™ toolbox that facilitates the analysis of the rich timestamped event records that modern computerized behavioral testing generates. The toolbox puts in the hands of investigators with limited programming experience the ability to extract from such records sophisticated statistics using only a few high level commands. The design of the toolbox ensures an intact data trail. Before presenting the system and the toolbox and some of the results obtained using them, we elaborate on the properties of an effective cognitive phenotype.