For several decades, the field of human neurophysiology has focused on the role played by cortical oscillations in enabling brain function underpinning behaviors. In parallel, a less visible but robust body of work on the stochastic resonance phenomenon has also theorized contributions of neural noise − hence more heterogeneous, complex and less predictable activity − in brain coding. The latter notion has received indirect causal support via improvements of visual function during non-regular or random brain stimulation patterns. Nonetheless, direct evidence demonstrating an impact of brain stimulation on direct measures of neural noise is still lacking. Here we evaluated the impact of three non frequency-specific TMS bursts, compared to a control pure high-beta TMS rhythm, delivered to the left FEF during a visual detection task, on the heterogeneity, predictability and complexity of ongoing brain activity recorded with scalp EEG. Our data showed surprisingly that the three non frequency-specific TMS patterns did not prevent a build-up of local high-beta activity. Nonetheless, they increased power across broader or in multiple frequency bands compared to control purely rhythmic high-beta bursts tested along. Importantly, non frequency-specific patterns enhanced signal entropy over multiple time-scales, suggesting higher complexity and an overall induction of higher levels of cortical noise than rhythmic TMS bursts. Our outcomes provide indirect evidence on a potential modulatory role played by sources of stochastic noise on brain oscillations and synchronization. Additionally, they pave the way towards the development of novel neurostimulation approaches to manipulate cortical sources of noise and further investigate their causal role in neural coding.