Validity and Homology

Predictive validity

The ability of a neurocognitive paradigm to predict the efficacy of a treatment on a cognitive deficit in schizophrenia. This is the ultimate aim of the use of a task in experimental therapeutics research.

Face validity

The cross-species (or cross-situational) similarity of behavioral patterns in terms of their topography and apparent biological/ethological significance.

Construct validity

The ability of a paradigm to isolate and measure the neurocognitive process of interest as evidenced by 1) systematic modulation of performance by cognitive demands hypothesized to recruit the cognitive process and 2) recruitment of or dependence on neural systems hypothesized to mediate the neurocognitive process.

Neurocognitive homology

The conservation across species of the mediation of a cognitive process by homologous neural circuits and neurobiological processes. Neurocognitive homology is evidenced by similar modulation of the behavior by the same environmental demands and mediation of the behavior homologous neural circuits across the species. Construct validity in an animal model assumes neurocognitive homology.

Lower-level (inter-test) forms of validity that help establish construct validity

Criterion validity

Usually refers to two tests applied to the same species, the ability of performance in one task to predict performance on another, more ecologically valid test. (e.g. the ability of a neuropsychological pencil-and-paper test of working memory to predict the number of items from a grocery list one can remember while shopping).

Convergent validity

Evidenced by high correlation between performance patterns across cognitive tasks designed to measure the same neurocognitive process

Discriminant validity

Evidenced by low correlation between outcomes across tasks designed to measure distinct neurocognitive constructs

Parametric Properties and Reliability

Parametric properties

This term describes a mathematically modelable relationship between two or more variables across known values (the parameters) of at least one of the variables. In the context of construct validity, it is a reliable and quantifiable relationship between a cognitive demand (applied at multiple levels) and the outcome measure (behavioral performance or biomarker). Parametric manipulation of the demand on a cognitive construct is needed to support an interpretation that a change in performance reflects change in the cognitive construct. In the context of treatment assessment, this is the relationship between drug dose - and its interaction with cognitive demand - on cognitive performance.

Test-Retest and Alternative Forms Reliability

Consistency in the outcome across repeated administrations of the same task, or across different, presumably equivalent forms of the same test.