A sensor-based gait measure for identifying invalid performance in simulated mTBI: an experimental proof of concept.

Abstract

Introduction: Psychologists are increasingly called upon to make complex decisions affected by an individuals' effort, relying at least in part on performance validity testing. Because most measures of performance validity are cognitively-based, there is utility in exploring alternative assessment modalities. Recently, an accelerometer-based gait assessment has been validated for use with sports concussion outcomes. The current, preregistered study serves as an experimental proof of concept utilizing the same measure of gait to differentiate individuals instructed to simulate a mild Traumatic Brain Injury (mTBI) from those with a sports-related mTBI (concussion).

Methods: 106 uninjured simulators were compared to 75 individuals with sports-related clinical mTBI and 75 normal controls. All participants completed an accelerometer-based measure of gait known as the BioKinetoGraph (BKG) as a part of a larger battery designed to assess for concussion in athletes.

Results: Unique abnormalities in the gait domains of Stride (timing), Power, Balance, and Symmetry were observed among simulating participants compared to mTBI participants. Classification analysis yielded several variables with fair to acceptable discriminability, and a multiple variable model with discriminability at the high end of the acceptable range (AUC = 0.78).

Conclusions: Findings offer preliminary support for the ability of an automated gait-based measure to capture invalid performance and inform investigations regarding effort. These experimental findings represent a first step in exploring the potential clinical utility of the BKG as a measure of performance validity.