Underlying Mechanisms of Visual Mismatch Responses – An EEG–fMRI Study
Schlossmacher et al., iScience (2025)

Underlying Mechanisms of Visual Mismatch Responses – An EEG–fMRI Study

1) What the study demonstrates clearly

This work provides robust, multimodal evidence that the mechanisms underlying visual deviance detection are not unitary, but vary systematically across time and along the cortical hierarchy:

• Early / sensory stages (posterior occipital cortex) → predominance of adaptation

• Later / hierarchical stages (anterior occipital cortex and superior parietal lobule – SPL) → predominance of prediction error (PE)

This organization emerges convergently across methods:

• EEG

• vMMN (160–210 ms) → mainly driven by adaptation

• P3 (300–600 ms) → mainly driven by prediction error

• fMRI

• Posterior occipital cortex → adaptation-related activity

• Anterior occipital cortex + SPL → prediction-error-related activity

Together, these findings align elegantly with a hierarchical predictive processing framework of the brain.

2) Conceptual contribution

The study addresses a long-standing ambiguity in oddball paradigms:

Are stronger neural responses to rare stimuli driven by genuine mismatch detection or by neural fatigue/adaptation to frequent stimuli?

The empirical answer provided here is: both, but at different times and different hierarchical levels.

• Adaptation explains early, sensory-level differences.

• Prediction error explains later, higher-order differences, reflecting model updating.

This resolves an artificial theoretical dichotomy and shows that adaptation and prediction error are complementary, hierarchically organized processes rather than competing explanations.

3) Methodological strength

• Use of simultaneous EEG–fMRI, still rare and technically demanding.

• Inclusion of an equiprobable control condition, essential for disentangling mechanisms.

• Explicit use of Bayesian statistics to assess evidence for absence of effects.

• Strong temporal–spatial coherence between ERP and BOLD findings within the same dataset.

Overall, this represents one of the cleanest experimental designs to date for studying visual mismatch mechanisms.

4) Critical considerations

• The lack of robust effects in regions such as the inferior frontal junction or anterior insula suggests that:

• the visual deviance may not have been sufficiently salient, or

• the paradigm preferentially engaged dorsal attention networks (SPL) rather than ventral salience networks.

• EEG–fMRI correlations for specific mechanisms (adaptation vs. PE) were inconclusive:

• this indicates that temporal signatures (ERPs) and spatial signatures (BOLD) do not map one-to-one at the individual level;

• it highlights the need for latent computational modeling approaches to bridge modalities more effectively.

5) Integrative interpretation (core insight)

The central message of the findings can be summarized as follows:

The brain first adapts to what is expected, and only later signals an error when the world violates its internal model.

In other words:

• Early mismatch responses are primarily sensory–physiological.

• Late mismatch responses are cognitive–predictive, involving hierarchical model updating.

6) Theoretical implications

This study repositions the visual MMN (vMMN):

• not as a pure marker of prediction error,

• but as a mixed phenomenon, strongly dominated by adaptation when stimuli are task-irrelevant.

At the same time, it reinforces the P3 as:

• a robust index of hierarchical prediction error, consistent with Bayesian brain models.

7) Conclusion

Schlossmacher et al. convincingly demonstrate that visual mismatch processing is neither instantaneous nor homogeneous.
Instead, it is temporal, hierarchical, and dependent on cortical level.

This work sets a new benchmark for future studies on mismatch responses, predictive perception, and the neural organization of sensory awareness.

First, the sensor adapts.
Then, the model is corrected.

Ref.:

Schlossmacher, I., Protmann, I., Dilly, J., Hofmann, D., Dellert, T., Peters, A., Roth-Paysen, M.-L., Moeck, R., Bruchmann, M., & Straube, T. (2025). Underlying mechanisms of visual mismatch responses – An EEG-fMRI study. IScience, 28(12), 114039. https://doi.org/10.1016/j.isci.2025.114039