In everyday life, finding specific items among distractors often requires joint contribution of visual attention and memory recall. We recently showed that we can successfully disentangle overlapping neural responses during natural viewing by applying deconvolution methods to coregistered EEG and eye-tracking data. We estimated temporal response functions (TRFs) through regularized linear models for main effects and their interactions, capturing fine-grained spatial and temporal activation patterns. Starting from hypothesis-driven models, we replicated established effects, including well-known components for visual processing and target detection. Extending to interactions with a data-driven approach, TRF estimates remained consistent across increasingly complex models, with their performance evaluated via explained variance (R²), and collinearity through variance inflation factors (VIF). In particular, in the main effects, we identified a late activation consistent with the P300 component for target detection. Furthermore, we also showed an interaction with correct detections, indicating missed detections elicited a similar but weaker activation and a more nuanced role of this component. These analyses demonstrate how deconvolution methods can uncover the dynamic interplay of cognitive processes underlying real-world search behavior.