The paired-click paradigm is widely used to index sensory gating via suppression of the auditory P50 component. Prior variability in P50 findings across studies, laboratories, and populations has prompted calls for methodological standardisation. While filtering choices have been empirically examined previously, other preprocessing steps such as segmentation windows and artifact-handling strategies have varied without direct comparison. The present study constructed four preprocessing pipelines combining two segmentation strategies (Short vs Long) with two artifact-handling approaches (threshold-based Rejection vs a hybrid Correction pipeline consisting of non-ocular rejection followed by Gratton and Coles ocular correction) to quantify the effects of these decisions on P50 precision.
Precision of P50 estimates was quantified using a bootstrapped standardized measurement error (bSME). The authors applied bSME to peak-to-peak P50 amplitudes at electrodes Cz and FCz for S1, S2, and the suppression score. Using bSME provides a standardised index of measurement precision that can be compared across preprocessing choices and outcomes in the paired-click context.
Fifty-six neurotypical adults completed a paired-click paradigm comprising 120 trials. Four preprocessing pipelines were defined by crossing Segmentation (Short vs Long) with Artifact Handling (Rejection vs Correction). The Rejection pipelines used threshold-based trial exclusion for artifacts. The Correction pipelines first rejected non-ocular artifacts and then applied the Gratton and Coles method for ocular correction, thereby retaining trials that would otherwise be discarded for ocular activity. Peak-to-peak P50 amplitudes were measured at Cz and FCz for stimulus 1 (S1), stimulus 2 (S2), and the suppression score (S1–S2). Trial retention for each pipeline was recorded and used in subsequent analyses.
Pipelines implementing ocular correction retained more trials than pipelines relying on threshold-based rejection. Across both Cz and FCz and for S1, S2, and suppression outcomes, Correction pipelines produced lower bSME values, indicating higher precision, relative to Rejection pipelines. These findings were consistent across the electrodes and outcome measures reported.
Segmentation and Artifact Handling did not act independently. Under the Correction approach, Segmentation (Short vs Long) had a negligible impact on bSME, meaning that when ocular artifacts were corrected rather than rejected, the choice of segmentation window had little effect on measurement precision. In contrast, under the Rejection approach, Long segmentation consistently produced higher bSME values (lower precision) compared with Short segmentation. Thus, Long segmentation exacerbated the loss of precision when combined with rejection-based artifact handling.
A regression analysis examined the relationship between trial retention and pipeline-related differences in bSME. The analysis showed that trial retention accounted for the large majority of differences in bSME across pipelines. In other words, the improved precision seen in Correction pipelines was largely attributable to higher numbers of retained trials after preprocessing.
Based on these results, the authors recommend that paired-click studies adopt ocular correction methods (specifically, the pipeline tested that combined non-ocular rejection with Gratton and Coles ocular correction) rather than rejecting trials containing ocular artifacts. Where matched S1–S2 averaging is not required, they advise using Short segmentation, because Long segmentation produced lower precision when paired with rejection-based artifact handling. Overall, the study supports the standardisation of preprocessing steps beyond filtering to improve precision and comparability of P50 suppression estimates.
This report summarises the analyses and recommendations as presented by the authors. The dataset comprised 56 neurotypical adults and 120 trials per participant; further details such as exact segmentation durations, filter settings, statistical effect sizes, and full regression coefficients were not reported in the abstract-level source used here. The article is a preprint and has not been certified by peer review. Readers should consult the full preprint for methodological specifics and complete results.
Preprocessing decisions—particularly the strategy for handling ocular artifacts and the choice of segmentation window—meaningfully influence the precision of P50 estimates in the paired-click paradigm. Correcting ocular artifacts and using Short segmentation (when feasible) improved precision by retaining more trials and lowering bSME. These findings argue for clearer standardisation of EEG preprocessing in P50 research to reduce variability across studies.