A growing number of studies show that transcranial stimulation with direct current (TDCS) or alternating current (TACS) applied above the auditory cortex can alter auditory cognition (for reviews, see Refs. [1e3]). These auditory effects are often interpreted as being mediated by neural excitability changes under the temporal electrodes and as reflecting a causal role of auditory cortex in the studied auditory phenomenon. A widely neglected alternative is to attribute this causal role,

more »

... or fully, to neuronal structures belonging to earlier stages of the auditory stimulus processing hierarchy. In support of this possibility, a large proportion of the TDCS/ TACS current applied to the scalp propagates along the highly conductive skin [4] by which it might reach receptors in the peripheral sensory system [5, 6] . Moreover, the small current fraction that directly penetrates the brain may spread along corticofugal projections by which it might reach subcortical and peripheral structures (for a review on such projections, see Ref. [7]). In the present study, we tested whether TACS can modulate the function of outer hair cells (OHCs) in the human cochlea. We applied 4-Hz TACS to nineteen normally-hearing participants and recorded distortionproduct otoacoustic emissions (DPOAEs; see Supplementary Material S1) from their right ear. We predicted that a modulation of DPOAE occurs during (online) and after (offline), but not before TACS. The experimental design ( Fig. 1A ) included four independent variables: To assess online effects, we used a within-subject experimental design where participants underwent two sessions involving either both TACS and sham stimulation (main session) or neither (control session). To assess offline effects, we presented sham stimulation two times, before and after TACS (Sham pre and Sham post ). To assess electrode-ear distance effects [5,6], we applied TACS at two different distances from the test ear using the electrodes above the ipsilateral (T8 and Cz; TACS near ) or contralateral auditory-motor cortex (T7 and Cz; TACS far ), respectively. This allowed us to test whether OHCs are more vulnerable to TACS when the latter is applied at a nearby scalp region. Finally, to avoid a ceiling effect, we probed DPOAEs at four different test levels near the participant's DPOAE threshold (see Supplementary Material S1.2). Together, this yielded two sessions, each with four blocks (20 min) resembling four conditions. Each block comprised four sub-blocks (5 min) resembling four different test levels. The order of sessions, blocks and sub-blocks was counterbalanced across participants. The obtained OAE recordings were transformed into the frequency domain and measures of interest were extracted for each condition and each participant (see Supplementary Material S2). The relevant measures were the level of the DPOAE, the average level of its spectral sidebands (the DPOAE frequency ± the TACS frequency), and the noise floor (average level at neighboring frequencies, for control). The DPOAE-sideband level was taken to measure the strength of the periodic modulation of DPOAE, the level of the DPOAE itself was taken to measure more sustained (non-periodic) DPOAE changes, and the level of the noise floor was taken to measure general (frequency-unspecific) effects on the OAE recording. To test for an online TACS effect, a three-way repeated-measures ANOVA including factors session, distance (near and far), and test level was used, which revealed no main effect of session on any of the aforementioned measures (DPOAE sideband: P ¼ 0.11, F 1, 18 ¼ 2.92; DPOAE: P ¼ 0.89, F 1, 18 ¼ 0.02; noise floor: P ¼ 0.14, F 1, 18 ¼ 2.35) and no significant interaction (see Fig. 1B left) . Thus, TACS had no instantaneous periodic or sustained effect on DPOAE or OAE recording. To test for an offline TACS effect, a two-way ANOVA including factors time (pre and post) and test level was used, which also revealed no main effect of time on any measure (DPOAE sideband: P ¼ 0.28, F 1, 18 ¼ 1.26; DPOAE: P ¼ 0.46, F 1, 18 ¼ 0.56; noise floor: P ¼ 0.23, F 1, 18 ¼ 1.58) and no significant interaction (see Fig. 1B center). Thus, TACS had no periodic or sustained aftereffect on DPOAE or OAE recording. To test for an electrode-ear distance effect, a two-way ANOVA including factors distance (near and far) and test level was used. As already suggested by the lack of a significant online TACS effect above, this revealed no main effect of distance on any measure (DPOAE sideband: P ¼ 0.31, F 1, 18 ¼ 1.11; DPOAE: P ¼ 0.79, F 1, 18 ¼ 0.08; noise floor: P ¼ 0.29, F 1, 18 ¼ 1.19) and no significant interaction (see Fig. 1B right) . In sum, these results provide no evidence for our hypothesis that TACS affects OHC function. They allow two mutually exclusive interpretations: On the one hand, it is conceivable that TACS alters OHC function as we hypothesized, but that we failed to detect it due to a lack of statistical power or current intensity. However, our sample size was comparable to that of previous studies that have Contents lists available at ScienceDirect