Emotionally arousing stimuli are known to rapidly draw the brain's processing resources, even when they are task-irrelevant. The steady-state visual evoked potential (SSVEP) response, a neural response to a flickering stimulus which effectively allows measurement of the processing resources devoted to that stimulus, has been used to examine this process of attentional shifting. Previous studies have used a task in which participants detected periods of coherent motion in flickering random dot kinematograms (RDKs) which generate an SSVEP, and found that task-irrelevant emotional stimuli withdraw more attentional resources from the task-relevant RDKs than task-irrelevant neutral stimuli. However, it is not clear whether the emotion-related differences in the SSVEP response are conditional on higher-level extraction of emotional cues as indexed by well-known event-related potential (ERPs) components (N170, early posterior negativity, EPN), or if affective bias in competition for visual attention resources is a consequence of a time-invariant shifting process. In the present study, we used two different types of emotional distractors – IAPS pictures and facial expressions – for which emotional cue extraction occurs at different speeds, being typically earlier for faces (at ~ 170 ms, as indexed by the N170) than for IAPS images (~ 220–280 ms, EPN). We found that emotional modulation of attentional resources as measured by the SSVEP occurred earlier for faces (around 180 ms) than for IAPS pictures (around 550 ms), after the extraction of emotional cues as indexed by visual ERP components. This is consistent with emotion related re-allocation of attentional resources occurring after emotional cue extraction rather than being linked to a time-fixed shifting process.