Giulio Contemori¹ (), Marianna Musa¹, Martina Passaggi¹, Carolina Maria Oletto¹, Stefano Vicentin², Luca Battaglini¹, Giorgia Cona^1,3, Marco Bertamini¹; ¹University of Padova, ²University of Osnabrück, ³Padua Neuroscience Center

Metacontrast masking reduces target visibility when a spatially adjacent, non-overlapping mask follows the target after a brief delay, yet its functional locus remains debated: does masking mainly disrupt early feedforward encoding, or later feedback-mediated integration? Aydın et al. (2021) isolated nonlinear target–mask interactions using a synthetic-minus-natural ERP measure. In contour discrimination with same-polarity target and mask, they reported a largely monotonic SOA dependence of the interaction signal in an early posterior cluster (160–300 ms), whereas a later centro-parietal cluster (300–550 ms) showed a U-shaped SOA profile paralleling Type-B behavioral masking, suggesting that behaviorally relevant suppression emerges in later, recurrent stages. We asked whether these signatures generalize to stimuli requiring global integration: visual symmetry. First, we confirmed that brief (20 ms) octagonal shapes elicit a canonical Sustained Posterior Negativity (SPN), indexing regularity processing. The SPN is a well-established posterior ERP marker, typically observed as a sustained negative deflection for regular relative to irregular patterns. Next, we presented the same stimuli followed by a same-polarity, spatially non-overlapping annular mask at SOAs of 10, 50, or 200 ms (plus an unmasked baseline). Behaviorally, symmetry discrimination showed a Type-B (U-shaped) masking function with maximal impairment at 50 ms. We quantified synthetic-minus-natural interaction waveforms within Aydın’s predefined windows (160–300 ms; 300–550 ms) and within an SPN-like cluster (276–424 ms) functionally defined from unmasked symmetry–asymmetry differences via cluster-based permutation. The predefined windows showed only modest modulation and did not capture the behavioral trough. In contrast, SPN-like attenuation varied systematically across SOAs, including a robust quadratic (U-shaped) component tracking performance; exploratory brain–behavior analyses linked stronger SPN suppression to lower accuracy. These findings suggest that the electrophysiological “window” of metacontrast suppression is stimulus- and computation-dependent, motivating task-sensitive, functionally defined neural markers when adjudicating feedforward versus feedback accounts of masking.

Acknowledgements: PRIN Grant No. 2022HEKCWH; PRID initiative (CoPeSy), Department of General Psychology, University of Padova.