Reliability, Attenuation, and Order Effects of EEG Components Across Multiple Assessments




Montgomery, Patrick

Journal Title

Journal ISSN

Volume Title



How does one guarantee consistency in measurements of neural activity? Realistically, this cannot be done for any individual measure, but variation from the norm will not significantly impact a large enough sample. The task, then, is to account for non-targeted neural activity common across participants to control for influences on target measures. Often, this is easier said than done. The goal of the current research was to aid in uncovering potential sources of unexplained variability in established electroencephalography (EEG) phenomena using two common tasks. Specifically, the reward positivity and P300 event-related potentials (ERPs) were captured via the two-armed bandit task and the oddball task, and were analysed across three areas: Reliability, attenuation, and order effects. Importantly, these data were captured using a unique testing schedule involving five cognitive assessments across an approximately two-hour period. Reliability was tested for both the difference and the conditional waves for each component to see if these lined up with commonly reported values. Previous attenuation studies have established this effect across long durations, but this analysis sought to determine whether this pattern held across consecutive testing sessions. Order effects were expected to occur between the bandit and oddball tasks based on the interplay between neural regions and neurotransmitter activity. Results: Excellent reliability was found for all P300 measures and for the conditional reward positivity measures. These findings support the use of conditional waves instead of difference waves regarding the reward positivity specifically, as the difference wave appears to mask the high reliability present in each conditional wave. Attenuation results were unanticipated, showing no effect for the reward positivity and an opposite effect for the P300. The suggestion invoked is that something to do with the divergences between the current study’s task and the long-duration tasks previously used to exhibit ERP attenuation altered participants’ reactions to the oddball task. Further investigation into this unusual component behaviour is warranted. No order effects were discovered across two analyses focused on these results. Although effects were anticipated, the absence is encouraging for EEG research as this suggests that order effects need not be accounted for in tasks or experiments that elicit both the reward positivity and the P300 ERPs. Altogether, these findings reveal areas wherein reward positivity and P300 components are robust, and areas in which they require further investigation.



Electroencephalography, Attenuation, Order Effects, Reliability, P300, Reward Positivity, Event Related Potential