BACKGROUND: The perception-action coupling task (PACT) was designed as a more ecologically valid measure of alertness/reaction times compared to currently used measures by aerospace researchers. The purpose of this study was to assess the reliability, within-subject variability, and systematic bias associated with the PACT.METHODS: There were 16 subjects (men/women = 9 / 7; age = 27.8 ± 3.6 yr) who completed 4 identical testing sessions. The PACT requires subjects to make judgements on whether a virtual ball could fit into an aperture. For each session, subjects completed nine cycles of the PACT, with each cycle lasting 5 min. Judgement accuracy and reaction time parameters were calculated for each cycle. Systematic bias was assessed with repeated-measures ANOVA, reliability with intraclass correlation coefficients (ICC), and within-subject variability with coefficients of variation (CVTE).RESULTS: Initiation time (Mean = 0.1065 s) showed the largest systematic bias, requiring the elimination of three cycles to reduce bias, with all other variables requiring, at the most, one. All variables showed acceptable reliability (ICC > 0.70) and within-subject variability (CVTE < 20%) with only one cycle after elimination of the first three cycles.CONCLUSIONS: With a three-cycle familiarization period, the PACT was found to be reliable and stable.Connaboy C, Johnson CD, LaGoy AD, Pepping G-J, Simpson RJ, Deng Z, Ma L, Bower JL, Eagle SR, Flanagan SD, Alfano CA. Intersession reliability and within-session stability of a novel perception-action coupling task. Aerosp Med Hum Perform. 2019; 90(2):77-83.