Transcranial alternating current stimulation at 10 Hz modulates response bias in the Somatic Signal Detection Task

Background Ongoing, pre-stimulus oscillatory activity in the 8-13 Hz alpha range has been shown to correlate with both true and false reports of peri-threshold somatosensory stimuli. However, to directly test the role of such oscillatory activity in behaviour, it is necessary to manipulate it. Transcranial alternating current stimulation (tACS) offers a method of directly manipulating oscillatory brain activity using a sinusoidal current passed to the scalp. Objective We tested whether alpha tACS would change somatosensory sensitivity or response bias in a signal detection task in order to test whether alpha oscillations have a causal role in behaviour. Methods Active 10 Hz tACS or sham stimulation was applied using electrodes placed bilaterally at positions CP3 and CP4 of the 10-20 electrode placement system. Participants performed the Somatic Signal Detection Task (SSDT), in which they must detect brief somatosensory targets delivered at their detection threshold. These targets are sometimes accompanied by a light flash, which could also occur alone. Results Active tACS did not modulate sensitivity to targets but did modulate response criterion. Specifically, we found that active stimulation generally increased touch reporting rates, but particularly increased responding on light trials. Stimulation did not interact with the presence of touch, and thus increased both hits and false alarms. Conclusions tACS stimulation increased reports of touch in a manner consistent with our observational reports, changing response bias, and consistent with a role for alpha activity in somatosensory detection.

stimulation (tDCS), transcranial alternating current stimulation [7], and transcranial random 48 noise stimulation (tRNS). Of these, tACS is particularly promising as a method by which to 49 interact with endogenous rhythms, since it allows application of a sinusoidal current at a 50 desired frequency. Indeed, there are several reports that tACS stimulation at or around 10 51 Hz modulates alpha power, increasing it even after stimulation has ended [8][9][10]. 52 Furthermore, modulation of alpha oscillations using tACS also influences detection of visual 53 targets phasically [8], consistent with the pattern found previously in the absence of tACS 54 stimulation [11][12][13]. 55 Effects of tACS on other sensory modalities, including audition [5] and pain [14], have 56 been reported. Most relevant here, however, is how tACS stimulation may influence 57 somatosensation. As in vision, tactile detection can be vary with the power of alpha 58 oscillations recorded over somatosensory regions [3]. We found that detection of 59 peri-threshold tactile stimuli was predicted from alpha power in a period shortly before 60 stimulus onset [3]. In that study, participants performed the Somatic Signal Detection Task 61 [15], in which they were asked to detect brief somatosensory stimuli delivered to their left 62 index finger at detection threshold. Brain activity was simultaneously recorded using EEG. 63 ALPHA TACS AND SOMATOSENSATION 4 We found that power in the alpha frequency band influenced both true and false reports of 64 somatosensory perception. As pre-stimulus alpha power increased, the probability of 65 reporting touch decreased, both in the presence and absence of target stimuli. Given that 66 alpha plays a similar role in both visual and tactile detection, and that alpha tACS 67 modulates visual detection, it follows that manipulation of alpha using tACS may also 68 modulate somatosensory detection. 69 A study by Gundlach,Müller,Nierhaus,Villringer,and Sem [16] found evidence 70 consistent with this suggestion. They had participants perform a somatosensory detection 71 task before, during, and after active alpha or sham tACS stimulation delivered over bilateral 72 somatosensory cortices. Tactile stimuli were delivered to the participants' right index finger.

73
The intensity of the stimuli was continuously varied, but maintained at detection threshold 74 using a staircase procedure. Detection thresholds for the stimuli in the periods before, 75 during, and after the stimulation period did not differ on average. However, during active 76 stimulation, detection thresholds varied in a phasic manner. Detection thresholds at opposite 77 phases of the driving oscillations differed from baseline (pre-stimulation) performance in 78 opposing fashion: some phases were associated with decreased thresholds while others were 79 associated with increased thresholds.     All participants took part in two experimental sessions separated by at least two days.

130
Before beginning the experiment, the tACS montage was set up as above. The experiment 131 itself was split into two parts. In the first part, each participant's sensory threshold (i.e., 132 50% detection rate) was established using a two-alternative forced choice adaptive staircase 133 procedure. Participants were given a series of trials consisting of two consecutive 1420 ms 134 time periods. Each time period began with a green arrow presented for 400ms on the left 135 side of the monitor and pointing down towards the participant's finger. The numbers "1" 136 and "2" were written on arrows marking the start of the first and second periods respectively.

137
After the offset of each arrow, the screen remained blank for 1020 ms. On each trial, a 20 ms 138 tactile pulse was delivered 500 ms after the offset of either the first or second arrow. After   Data analysis 165 We first performed three analyses using a standard ANOVA framework. These analyses 166 were performed primarily for comparison with previous studies using the SSDT, which used 167 ALPHA TACS AND SOMATOSENSATION 8 standard ANOVA analyses of touch reporting rates and of the signal detection measures 168 sensitivity (d ) and response criterion (c). For all analyses, we combined "Definitely yes" and 169 "Maybe yes" into "yes" reports and "Definitely no" and "Maybe no" into "no" reports.  In addition to our standard ANOVA analyses, we also fitted a Bayesian generalized 181 linear mixed effects model with a logistic link function using the brms package (see below).

182
A key advantage of using a logistic link function is that it appropriately models the change 183 in variance over the response scale: as mean reporting rates approach 100% or 0%, the 184 variance decreases. ANOVA conducted on percentages does not account for such changes in 185 variance and can lead to misleading conclusions [20]. We coded "yes" responses as 1 and "no" 186 responses as 0, combining "Definitely yes" and "Maybe yes" into "yes" reports and  In a Bayesian framework, the MCMC sampler produces a posterior distribution of 207 likely parameter values, which we summarise using 95% credible intervals. Where necessary, 208 we also calculated the ratio of posterior samples below zero relative to those above zero.

216
We first examined performance in a classical SDT framework. We found no significant    Table 1 and Figure 2). Individual dots indicate mean response rates for individual participants.

Bayesian multilevel model 241
The Bayesian GLMM proved notably different from the repeated measures ANOVA on 242 reporting rates (see Table 2 and Figure 3). The strong effect of Touch on reporting rates was 243 consistent with the ANOVA, but the model also suggests that there was a small increase in  Critically, there was little evidence of an interaction between Stimulation and Touch.

261
The posterior density spanned zero, with only a low probability of the parameter being

302
Our results do come with some caveats. First, our comparison of active versus sham 303 stimulation would not allow us to make concrete statements about the specificity of 304 stimulation at a particular frequency, since we stimulated only at a single frequency. Second, 305 since we did not record EEG before and after stimulation, we cannot be sure that we directly 306 influenced visual alpha or somatosensory alpha rhythms. Finally, since we used only a single 307 pair of stimulation locations, we cannot necessarily distinguish between non-specific effects of 308 tACS stimulation and direct effects of stimulation on the specific rhythms of interest.

309
Overall, however, our results are consistent with tACS stimulation at 10 Hz over