The aperiodic, 1/f-like component of electrophysiological activity is increasingly recognized as a meaningful feature of neural function rather than background noise. Meanwhile, many EEG studies interpret transient post-stimulus changes in oscillatory power as signatures of attention, salience, or cognitive control. Such interpretations typically depend on baseline normalization, which assumes aperiodic activity remains stable from pre-to post-stimulus periods. Using high-density EEG from typically developing children, we tested this assumption across two paradigms: an audiovisual simple reaction-time task (n = 36, 19 males and 17 females) and a visual oddball task (n = 38, 20 males and 18 females). In each, conventional spectral analyses were compared with analyses that explicitly modeled and removed aperiodic activity in both windows. Across tasks, stimulus onset produced robust increases in aperiodic exponent and offset, indicating systematic 1/f changes. In the audiovisual task these were modality-specific, with central, parieto-occipital, or combined topographies by stimulus type. They diminished but persisted after ERP removal, suggesting they were not fully explained by phase-locked activity. Critically, once aperiodic activity was accounted for, the apparent theta increase was largely abolished in both tasks, including the canonical fronto-central enhancement to infrequent targets. The conventional method also overestimated beta desynchronization, especially in the induced signal, indicating a spurious consequence of spectral slope steepening. By contrast, alpha desynchronization remained robust and was enhanced, reflecting genuine suppression. Together, these findings suggest many conventional time-frequency effects, particularly apparent theta synchronization, may reflect stimulus-driven aperiodic changes rather than robust narrowband changes, challenging core assumptions of these analyses.
Significance statement Frequency-domain neural activity comprises both narrowband oscillations and broadband aperiodic activity. Conventional time–frequency analyses use baseline normalization to isolate oscillatory responses, implicitly assuming that aperiodic activity remains stable following stimulus onset. Across two EEG paradigms, we show that stimulus presentation systematically alters aperiodic activity in a sensory-and attention-dependent manner. These changes can produce apparent increases in theta-band power that do not reflect genuine narrowband oscillations and can exaggerate beta desynchronization. In contrast, alpha desynchronization remained robust after accounting for aperiodic activity. These findings demonstrate that stimulus-related aperiodic changes can substantially influence conventional time–frequency measures and should be explicitly considered when interpreting post-stimulus oscillatory responses.
