Neural oscillations can be modulated by non-invasive brain stimulation techniques, including transcranial alternating current stimulation (tACS). However, direct evidence of tACS effects at the cortical level in humans is still limited. In a tACS-electroencephalography co-registration setup, we investigated the ability of tACS to modulate cortical somatosensory information processing as assessed by somatosensory-evoked potentials (SEPs). To better elucidate the neural substrates of possible tACS effects we also recorded peripheral and spinal SEPs components, high-frequency oscillations (HFOs), and long-latency reflexes (LLRs). Finally, we studied whether changes were limited to the stimulation period or persisted thereafter. SEPs, HFOs, and LLRs were recorded during tACS applied at individual mu and beta frequencies and at the theta frequency over the primary somatosensory cortex (S1). Sham-tACS was used as a control condition. In a separate experiment, we assessed the time course of mu-tACS effects by recording SEPs before (T0), during (T1), and 1 min (T2) and 10 min (T3) after stimulation. Mu-tACS increased the amplitude of the N20 component of SEPs compared to both sham and theta-tACS. No differences were found between sham, beta-, and theta-tACS conditions. Also, peripheral and spinal SEPs, P25, HFOs, and LLRs did not change during tACS. Finally, mu-tACS-induced modulation of N20 amplitude specifically occurred during stimulation (T1) and vanished afterwards (i.e., at T2 and T3). Our findings suggest that TACS applied at the individual mu frequency is able to modulate early somatosensory information processing at the S1 level and the effect is limited to the stimulation period.