Abstract: The assessment of cognitive functions in rodents represents a critical experimental variable in many research fields, ranging from the basic cognitive neurosciences to psychopharmacology and neurotoxicology. The increasing use of animal behavioral tests as 'assays' for the assessment of effects on learning and memory has resulted in a considerable heterogeneity of data, particularly in the field of behavioral and psycho pharmacology. The limited predictive validity of changes in behavioral performance observed in standard animal tests of learning and memory indicates that a renewed effort to scrutinize the validity of these tests is warranted. In humans, levels of processing (effortful vs. automatic) and categories of information (procedural vs. episodic/declarative) are important variables of cognitive operations. The design of tasks that assess the recall of 'episodic' or 'declarative' information appears to represent a particular challenge for research using laboratory rodents. For example, the hypothesis that changes in inspection time for a previously encountered place or object are based on the recall of declarative/episodic information requires substantiation. In order to generalize findings on the effects of neuronal or pharmacological manipulations on learning and memory, obtained from one species and one task, to other species and other tasks, the mediating role of important sets of variables which influence learning and memory (e.g. attentional, affective) needs to be determined. Similar to the view that a neuronal manipulation (e.g. a lesion) represents a theory of the condition modeled (e.g. a degenerative disorder), an animal behavioral task represents a theory of the behavioral/cognitive process of interest. Therefore, the test of hypotheses regarding the validity of procedures used to assess cognitive functions in animals is an inherent part of the research process.