Altering the primacy bias—How does a prior task affect mismatch negativity?

Mullens, Daniel; Woodley, Jessica; Whitson, Lisa; Provost, Alexander; Heathcote, Andrew; Winkler, István; Todd, Juanita · 2014 · Psychophysiology

DOI: 10.1111/psyp.12190

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the "primacy bias" in auditory deviance detection, a phenomenon where the initial role a sound plays in a sequence (as a frequent standard or rare deviant) disproportionately influences how it is processed when roles later reverse. Specifically, mismatch negativity (MMN), an event-related potential reflecting automatic change detection, is typically larger for sounds first encountered as deviants than for those first encountered as standards, regardless of subsequent probability statistics. The authors sought to determine if this bias stems from automatic assumptions about the behavioral relevance of sounds. They hypothesized that assigning explicit behavioral relevance to both tones prior to the oddball sequence would alter or abolish this bias. Thirty healthy participants completed a go/no-go detection task where short (30 ms) and long (60 ms) tones were presented with equal probability. Half the participants responded to short tones (short go-stimulus group), while the other half responded to long tones (long go-stimulus group). This was followed by an unattended oddball sequence where the roles of the tones alternated between standard (p = .875) and deviant (p = .125) at either slow (2.4 min stability) or fast (0.8 min stability) rates. The sequence order was manipulated across three blocks to test the persistence and flexibility of the bias. EEG data were recorded to measure MMN amplitudes. The results demonstrated that prior behavioral training significantly modulated the primacy bias. In the first sequence order, the typical primacy bias was abolished; MMN amplitudes for both tones were larger in slow-changing sequences compared to fast-changing ones, indicating that prior equal relevance prevented the automatic undervaluation of the initial standard. However, the bias returned in subsequent orders, dependent on the specific go-stimulus assignment. In the second order, where the short tone was the first deviant, the short go-stimulus group showed a strong primacy bias (enhanced MMN for the short tone in slow sequences), whereas the long go-stimulus group did not. This suggests that prior relevance creates persistent, stimulus-specific changes in cortical responsiveness. The bias patterns differed between groups in a way that could not be fully explained by simple latent inhibition, pointing toward a meta-learning process where the brain adjusts its monitoring strategies based on prior experience with sound relevance. These findings imply that the primacy bias is not merely a rigid statistical learning artifact but is influenced by the attribution of behavioral value to sounds. The study suggests that prior learning about the importance of specific stimuli has long-term impacts on automatic deviance detection mechanisms. Even when sounds are initially learned as equally important, the system retains a bias linked to the specific behavioral role (go vs. no-go) assigned to them. This highlights the dynamic nature of auditory processing, where automatic sensory memory and deviance detection are shaped by prior behavioral contexts and expectations, rather than solely by immediate acoustic statistics.

Key finding

Prior assignment of behavioral relevance to tones in a go/no-go task initially abolishes the primacy bias in mismatch negativity, but the bias subsequently returns in a manner dependent on the specific tone designated as the go-stimulus.

Methodology

lab_experiment

Sample size: 30

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-28.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• task switching