Establishing Psychometric Expectations for Neurobiological Assessments

Across psychological science, there has been an explosion of new tools and technologies over the last decade. In an upcoming symposium at the 28th APS Annual Convention in Chicago, May 26–29, 2016, experts will discuss the cross-cutting implications of emerging psychometric standards unique to neurobiological measures.

From neuroimaging to genetic analysis, these new tools ostensibly provide previously unimaginable access to underlying biological processes in the brain. However, these neurobiological measures aren’t always held to the same internal psychometric standards for temporal stability and consistency standards as traditional assessments are.

The symposium will include speakers from various subdisciplines of psychological science. Each speaker will offer expertise on specific neurobiological methods and applications and will highlight important methodological questions about developing standards and expectations for key neurobiological tests:

• APS Fellow Greg Hajcak Proudfit (Stony Brook University) will discuss EEG and psychophysiological measures of emotion.
• Greg Siegle (University of Pittsburgh) will discuss neuroimaging measures of emotional-information processing.
• Douglas Granger (Arizona State University) will discuss hormonal assessment through the collection of saliva.
• Arpana Agrawal (Washington University in St. Louis) will discuss genetic measurement in relation to substance abuse.
• APS Fellow Christopher Patrick (Florida State University) has written extensively about merging data from different sources to arrive at a better approximation of constructs such as personality or well-being. It’s currently impossible to directly measure “latent constructs” such as personality or happiness. Although it’s widely accepted that a survey can’t directly measure these latent constructs, it’s also important to note that newer tools are still indirect approximations. Patrick will discuss general ideas about the standards that various neurobiological tests should meet.

“They are widely acknowledged as really valuable tools that have the potential to inform research across all of psychology,” says symposium organizer Douglas Samuel of Purdue University. “What we’re hoping to do within this symposium is to reinforce that these tools are still measures.”

Though they may offer exciting insights, new, biologically based tests are not a direct window into personality or the brain itself. Just like other measures, they offer an indirect approximation of psychological traits and mechanisms.

For example, it’s common for fMRI studies to compare a group of approximately 20 depressed patients with a group of 20 matched controls. When each subject is asked to perform a task, fMRI may show that certain areas of the brain (for example, the prefrontal cortex) are less active in those diagnosed with depression compared with controls.

However, the conclusion that depression is related to less activity in the prefrontal cortex still leaves many important questions unanswered — including questions about how much error may be a part of the test itself. In contrast, researchers widely acknowledge the limitations of “older” methods such as self-report questionnaires.

Symposium speakers also will delve into additional methodological queries that must be answered before neurobiological testing is put on par with more established testing methods. To continue the example of fMRI studies of depressed individuals, would the same findings still emerge at a different time of day or using a different task?

“Although these are valuable tools, they are still measures,” Samuel explains. “That is to say, they are still approximations of latent constructs, and so the measurement properties really matter for determining how they are used.”