4.3 Adaptation and MMN
According to the regression analysis, both N1 initial adaptation and P2 subsequent adaptation amplitudes significantly predicted the MMN, explaining 49.98% and 10.43% of the MMN variance, respectively. While the effect of N1 initial adaptation aligns with the adaptation hypothesis of MMN, the contribution of P2 subsequent adaptation amplitudes suggests that this hypothesis alone may not suffice, as it proposes that the N1 adaptation alone explains MMN. Furthermore, a pure adaptation mechanism struggles to explain the N1 rebound effect. Thus, while the present study underscores the role of adaptation in MMN, it does not dismiss model adjustment or predictive coding accounts. Instead, our main findings support a predictive coding account of MMN, consistent with prior research (e.g., Alain et al., 1999; Herholz et al., 2009; Symonds et al., 2017; Wacongne et al., 2011).
Notably, MMN appears to comprise two distinct processes related to the RP: One involving the less negative N1 inclination during the first repetition, indicative of a purer adaptation effect, and the other involving the P2 subsequent adaptation with a positive deflection, reflecting memory trace formation required by model adjustment. In addition, the N1 initial adaptation effect possibly indicates the extent of model adjustment following the emergence of deviants, while MMN quantifies the degree of error detected by the model. Hence, the positive correlation between N1 initial adaptation and MMN suggests that participants who exhibited heightened sensitivity to the tone changes also had the most precise models. In addition, the increasing P2 amplitude across repetitions aligns with the predictive coding account, which posits that prediction error decreases as top-down predictions match bottom-up inputs (Friston, 2005). This explanation integrates both adaptation and model adjustment accounts (Garrido et al., 2008). Importantly, the present study demonstrates that the predictive coding account holds even when the expectations are controlled, as evidenced by the adaptation findings and the distinct topographies of N1 initial adaptation and P2 subsequent adaptation.
            Our regression findings underscore the importance of considering different components and time windows when examining the connection between adaptation and MMN. Specifically, while the N1 initial adaptation contributes to the MMN, the subsequent adaptation of the P2 predicts it. These results suggest that future studies investigating adaptation or its association with MMN should examine both N1 and P2 components. Focusing solely on one component may hinder the discovery of comprehensive adaptation effects. Notably, isolating the N1 and P2 processes could provide insights into the relationship between adaptation and MMN.