A systematic evaluation of the oMST's alternate stimulus sets, languages, and contaminant detection methods

The optimized Mnemonic Similarity Task (oMST) is a brief self-administered cognitive screening assessment derived from the research-grade MST. The present study examined its psychometric properties, including the comparability of difficulty across stimulus sets, cross-linguistic validity, and methods for identifying contaminant response behavior.

Two experiments were conducted using the Prolific global online testing platform. Experiment 1 assessed practice effects and the reliability of repeated oMST administration, as well as stimulus difficulty across 18 independent stimulus sets. Experiment 2 compared performance on the English, Spanish, and Chinese versions of the oMST to evaluate cross-linguistic validity. Finally, simulated data and data from Experiment 1 were used to reexamine the commonly used recognition memory (REC) exclusion threshold of 0.50 and to compare it with alternative screening metrics for contaminant response behavior.

oMST performance was reliable across repeated administrations, with no evidence of practice-related performance gains. In addition, performance was broadly consistent across the 18 stimulus sets, indicating largely comparable difficulty. The English and Spanish versions of the oMST had comparable performance, whereas the Chinese version showed slightly lower performance. Based on analyses of both simulated and actual participants, we demonstrate that contaminant filtering using the traditional REC < 0.50 threshold identifies both guessing and low-effort behaviors, while using a threshold based purely on Foil Accuracy can selectively identify guessing behavior.

These findings support the oMST as a reliable and accessible behavioral tool for assessing mnemonic discrimination, a memory ability associated with hippocampal function, across repeated administrations and across a broad range of stimulus sets.