If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Experimental Psychology Lab, Department of Psychology, Cluster of Excellence “Hearing4all”, European Medical School, Carl von Ossietzky University Oldenburg, Oldenburg, 26129, GermanyResearch Center Neurosensory Science, Carl von Ossietzky University Oldenburg, Oldenburg, 26129, Germany
1 Department of Psychiatry and Psychotherapy, University of Lübeck, Lübeck, 23562 Germany. 2 Biomedical Signals and Systems, Technical Medical Center, University of Twente, Enschede, 7522 The Netherlands. 3 shared last authorship.
Rhythmic visual and electrical stimulation were combined at varying phase offsets.
•
Flicker brightness percepts are modulated dependent on the flicker-tACS phase shift.
•
Perceptual effects by tACS phase are dependent on stable flicker-phase entrainment.
•
tACS can transmit temporal information to the cortex, affecting visual perception.
Abstract
Background
Visual phenomena like brightness illusions impressively demonstrate the highly constructive nature of perception. In addition to physical illumination, the subjective experience of brightness is related to temporal neural dynamics in visual cortex.
Objective
Here, we asked whether biasing the temporal pattern of neural excitability in visual cortex by transcranial alternating current stimulation (tACS) modulates brightness perception of concurrent rhythmic visual stimuli.
Methods
Participants performed a brightness discrimination task of two flickering lights, one of which was targeted by same-frequency electrical stimulation at varying phase shifts. tACS was applied with an occipital and a periorbital active control montage, based on simulations of electrical currents using finite element head models.
Results
Experimental results reveal that flicker brightness perception is modulated dependent on the phase shift between sensory and electrical stimulation, solely under occipital tACS. Phase-specific modulatory effects by tACS were dependent on flicker-evoked neural phase stability at the tACS-targeted frequency, recorded prior to electrical stimulation. Further, the optimal timing of tACS application leading to enhanced brightness perception was correlated with the neural phase delay of the cortical flicker response.
Conclusions
Our results corroborate the role of temporally coordinated neural activity in visual cortex for brightness perception of rhythmic visual input in humans. Phase-specific behavioral modulations by tACS emphasize its efficacy to transfer perceptually relevant temporal information to the cortex. These findings provide an important step towards understanding the basis of visual perception and further confirm electrical stimulation as a tool for advancing controlled modulations of neural activity and related behavior.
Dissociations between human brightness estimation and objective viewing conditions reveal important insight into the constructive mechanisms of perceptual processing. Depending on constituents of the visual scene, feedforward projections along the visual pathway are subject to complex neural computations that shape perception in addition to the physical level of illumination [
]. A popular approach to examine the neural substrate of perception builds on frequency tagging of neural activity by sinusoidally modulated luminance stimuli that evoke phase-locked responses in visual cortex [
]. This way, single-cell firing patterns were shown to follow the frequency of flickering visual stimuli, reflecting actual as well as illusory brightness percepts [
]. Correlations between firing pattern modulation and the subjective experience of brightness were found in primary visual cortex but not earlier in the visual pathway. Further, invasive recordings in cats revealed a relation between brightness perception and enhanced discharge rates or synchronization of neural firing at the tagging-frequency in visual cortex [
Neural synchronization has long been discussed as a candidate mechanism in the cerebral cortex that increases the efficacy of neural responses in driving target neurons, thereby enhancing response saliency [
]. Noninvasive recordings in humans of flicker entrained activity, i.e., visually evoked steady-state responses (SSRs), converge with these findings. Perceptual competition between two concurrently presented flickering stimuli of equal luminance was biased in favor of the stimulus with greater inter-trial phase coherence of evoked SSRs [
Modification of brain oscillations via rhythmic light stimulation provides evidence for entrainment but not for superposition of event-related responses.
]. Thus, temporal neural dynamics at early stages of cortical processing are assumed not merely to reflect physical stimulation properties but rather to code the subjective quality of perception. Given this correlative evidence, does a manipulation of flicker-evoked rhythms affect the subjective experience of brightness under otherwise constant viewing conditions?
Transcranial alternating current stimulation (tACS) is used with the aim to modulate neural dynamics by phase-specific excitability modulation in targeted brain regions [
]. Invasive recordings in monkeys yielded key evidence that neuronal spike timing follows the phase of the externally applied electric current, thereby inducing higher synchrony and related power increases at the stimulation frequency [
]. Only recently, we have shown that pairing tACS with same-frequency flicker results in systematic enhancement and suppression of flicker-evoked SSR amplitudes dependent on the phase shift between visual and electrical stimulation [
]. Yet, crucially, whether this tACS-induced modulation of neural activity to flickering light also affects the subjective experience of flicker brightness has not been established. Evidence for perceptual deterioration or improvement by interfering with flicker-evoked responses would provide a strong claim about the functional relevance of the underlying neural response.
Here, we asked whether biasing cortical processing of rhythmic visual stimulation by tACS under constant luminance conditions translates into perceptual changes. By using oscillatory entrainment via tACS, we aimed to modulate flicker-evoked responses by concurrent electrically-induced temporal changes in neuronal excitability within the visual cortex. Participants performed a two-flicker brightness discrimination task, while tACS was simultaneously applied at one of the two flicker frequencies with systematic phase shift. We hypothesized that the modulation of neural response dynamics in visual cortex by phase-shifted visually and electrically driven neuromodulation leads to perceptual changes of flicker brightness.
2. Materials and methods
2.1 Participants
38 right-handed participants (mean age 25.74 ± standard deviation 4.06 years, 22 female, 16 male) volunteered for this study. All participants had normal or corrected-to-normal vision and no history of psychiatric or neurological illness. Handedness was assessed via the short version of the Edinburgh Handedness Inventory. Participants gave written informed consent prior to participation and were monetarily compensated. The study was approved by the ethics committee of the Hamburg Medical Association (PV4908) and conducted in accordance with the declaration of Helsinki.
2.2 Experimental design
The experimental protocol was conducted on two separate days for occipital and periorbital tACS in counterbalanced order. In the beginning of each testing session, EEG was recorded in order to assess the physiological response to visual flicker stimulation and the spectral properties of brain activity during resting-state. After psychophysical estimation of brightness discrimination thresholds for individualization of physical luminance intensity (see Supplementary Material A), participants performed the main tACS experiment.
During the tACS experiment, participants were presented with a two-flicker brightness discrimination task between an 8 Hz tACS-targeted flicker (LED8Hz) and a 12 Hz reference flicker (LED12Hz). Multi-electrode tACS at 8 Hz was simultaneously applied at six different phase shifts (0°, 60°, 120°, 180°, 240°, 300°) relative to the 8 Hz flicker cycle (Fig. 1A and B). The choice of stimulation frequencies was based on the previous literature, showing greatest flicker-evoked responses in the alpha and low beta range [
]. Thus, we decided for flicker frequencies at 8 and 12 Hz as the borders of the alpha frequency range where flicker-evoked neural responses were expected to be comparable in strength but distinct enough with regard to frequency characteristics such that tACS phase-specifically interacts with only one of the two rhythms. Yet, in principle our experimental approach is not limited to the chosen frequency range. Single flicker trials were presented at five luminance ratios at [35, 42.5, 50, 57.5, 65] % of the individual discrimination performance. The flicker phase angle at light onset of the LED8Hz was set to 0°, while the LED12Hz started randomly at four different phase angles (0°, 90°, 180°, 270°). Single flicker trials had a duration of 2 s, followed by a break of 1.75–2 s. After each trial, participants indicated the side of flicker presentation appearing as brighter per button press (or darker, depending on the assigned counterbalanced condition). 19 trials were presented for each of the six tACS-LED8Hz phase shift conditions and five flicker luminance ratios, resulting in 570 trials in total. Trial presentation was subdivided in three blocks à 13.3 min, interrupted by 10 min breaks. Side of flicker presentation and phase offsets were counterbalanced across tACS-LED8Hz phase shift conditions.
Fig. 1Experimental setup. (A) During continuous transcranial alternating current stimulation (tACS) at 8 Hz, participants were asked to judge per button press which of the two simultaneously presented flickers appeared brighter. Luminance of the LED8Hz was varied while the LED12Hz remained at constant luminance. The lag between LED8Hz and tACS onset was systematically varied across six different phase shifts. (B) Schematic of the experimental hypothesis that phase-specific electrical stimulation of same-frequency flicker-evoked responses systematically modulates brightness perception. (C) The two flickers were presented simultaneously in the lower visual field while participants fixated the upper fixation dot. (D) During occipital tACS at 4 mA peak-to-peak, estimated maximum electric field strength was about 0.6 V/m, reaching 0.3 V/m in the target regions within the upper bank of the calcarine sulcus (MNI-coordinates: -10, -90, 2 and 10, -90, 2). (E) During the periorbital control condition, estimated electric field strength in the eyeballs for tACS intensities at 80% phosphene threshold was comparable to the field magnitude induced by occipital tACS due to current shunting across the scalp (periorbital: bootstrapped mean .01, 95%-CI).
During the tACS experiment, multi-electrode tACS was applied at 8 Hz via Ag/AgCl electrodes (12 mm diameter) using neuroConn stimulators (DC-Stimulator plus, neuroConn, Illmenau, Germany). One hour before tACS application, the skin beneath the stimulation electrodes was prepared with EMLA cream for local anesthesia to reduce transcutaneous stimulation effects by tACS (2.5% lidocaine, 2.5% prilocaine). Electrode impedances were kept below 10 kΩ and comparable between electrodes to ensure an evenly distributed electric field. During the tACS experiment of both sessions, stimulation was applied for 40 min in total, divided in three blocks of 13.3 min duration including a linear ramp-up time of 6 s and another 6 s at the final intensity before the start of visual stimulation.
For occipital tACS, current was applied via a 4 x 1 montage over the parieto-occipital cortex at 4 mA peak-to-peak (Fig. 1D). The stimulation montage was tailored to target the upper bank of the calcarine sulcus, as described in detail in Supplementary Material A. Estimated electric field strengths by simulations based on finite element head models were comparable to electric field strengths shown to successfully modulate neural activity in nonhuman primates and humans [
]. The periorbital tACS montage served as an active control condition for sub-threshold stimulation effects of the retinae due to current shunting across the scalp (Fig. 1E). Current was applied via four electrodes placed superior and inferior to the eyes with current flow in the vertical direction [
Noninvasive transorbital alternating current stimulation improves subjective visual functioning and vision-related quality of life in optic neuropathy.
]. Mean stimulation intensity was 0.071 ± 0.037 mA peak-to-peak. Peripheral tACS side-effects were systematically quantified at the end of each session via a questionnaire, assessing the perceived strength and temporal course of skin sensations, fatigue and phosphene perception (Supplementary Material B, Fig. B1). Out of 38 participants, two reported supra-threshold phosphene perception under occipital tACS and were therefore excluded from data analysis.
2.4 Electrophysiological recording and data analysis
EEG data were recorded from 18 Ag/AgCl electrodes (12 mm diameter) mounted in parieto-occipital regions in an elastic cap for 64 electrodes (Easycap, Herrsching, Germany). We assessed 60 s resting state EEG with eyes open and the electrophysiological response to flicker stimulation while participants were presented with an LED flickering at either 8 Hz or 12 Hz for 60 s in the right lower visual field (Fig. 1C). Electrode impedances were kept below 20 kΩ. The electrooculogram was recorded with two electrodes placed below the eyes. EEG signals were referenced to the nose tip and recorded using BrainAmp MR plus amplifiers (Brain Products GmbH, Gilching, Germany) and the corresponding software (Brain Products GmbH, Recorder 1.20). Data were recorded with an online passband of 0.016–250 Hz and digitized with a sampling rate of 5 kHz.
EEG data were preprocessed and analyzed in Matlab (The MathWorks Inc.) using the analysis toolbox FieldTrip [
] and custom-made scripts. Data were bandpass-filtered between 1 and 20 Hz using a Hamming-windowed sinc FIR (finite impulse response) filter (order 8250) and segmented in 2 s time epochs. We quantified the extent of visual entrainment by the phase locking value (PLV) between EEG and visual flicker [
]. A Hilbert transform was computed on the flicker and the EEG signal of each epoch and at each electrode. The time-dependent differences of the flicker and EEG signal phase () for each sample t served to compute the PLV
(1)
with time steps per epoch, and epochs. Comparability of flicker entrainment strength between the two tACS sessions and flicker frequencies was computed by repeated measures analysis of variance (ANOVA).
2.5 Statistical data analysis of tACS effects
To assess whether tACS modulates perceived flicker brightness in a phase-dependent manner, we computed the proportion of brighter ratings of the LED8Hz relative to the LED12Hz for each of the six LED8Hz-tACS phase shift conditions. As the optimal tACS phase leading to highest brightness perception varied across participants (see Supplementary Material B, Fig. B2), we quantified the degree of tACS phase-dependence by applying a parametric alignment-based method proposed by Riecke et al. [
]. Herein, the maximum brightness rating condition was aligned to the center bin and the remaining conditions were phase-wrapped. Aligning data to one phase bin naturally skews variability which could produce false positive results if not corrected for. Thus, the average brightness rating of the two bins adjacent to the bin opposite to the center bin (ADJ) was subtracted from the average brightness rating of the two bins adjacent to the center bin (MAX, Fig. 4A). In case of more than one optimal phase bin per participant, we repeatedly computed MAX-ADJ values per optimal data bin alignment and conservatively took the average of MAX-ADJ values. By excluding the center bin and the bin opposite to the center bin, analytical bias that might account for reported tACS effects was prevented [
Across the whole sample, we first analyzed the overall effect of tACS on brightness perception by t-tests assessing the statistical significance of MAX-ADJ values within and between tACS conditions. Effect sizes were determined by calculating Cohen's d and dz for one-sample and paired samples t-tests, respectively. The Bonferroni corrected alpha level of test statistics was set to 0.016 (= 0.05/3). Yet, crucially, in our experimental setup stable phase entrainment by visual flicker was the key prerequisite for observing phase-specific modulations of flicker-evoked rhythms by tACS [
]. According to this rationale, we examined the dependency of tACS-induced perceptual effects quantified by MAX-ADJ on the degree of SSR-flicker phase locking for each tACS montage and flicker frequency. The absolute correlation coefficient difference between the occipital and periorbital tACS montage was computed per electrode and tested for statistical significance by cluster permutation statistics. To generate the permutation distribution, correlation coefficients were repeatedly computed after pairwise shuffling of MAX-ADJ and SSR-flicker phase locking values between tACS montages per participant across 10,000 permutations. Electrodes were considered significant when the observed correlation difference exceeded 97.5% of the permutation distribution. For an overall test of the tACS efficacy on the final electrode cluster, we used a linear mixed model to analyze the effect of the fixed factors “tACS montage” (periorbital vs. occipital), “SSR-flicker phase locking at 8 Hz”, “SSR-flicker phase locking at 12 Hz” as well as the interaction effects between SSR-flicker phase locking and tACS montage on brightness modulation. PLVs were averaged across the previously defined cluster electrodes. The subject was set as random effect to account for differences in overall perceptual ratings independent of experimental manipulation. Bonferroni correction was used to adjust for multiple comparisons during post hoc testing. Model fitting was implemented using SPSS Statistics 27 (IBM Corp.).
To investigate the predictability of the optimal timing of tACS application, we computed circular correlation coefficients between the individual SSR-flicker phase delay and the optimal flicker-tACS phase shift leading to greatest brightness ratings. Therefore, we fit one-cycle sine waves to brightness ratings and extracted the optimal phase corresponding to the peak of the sine. Statistical significance of the correlation coefficient per electrode was tested by computing permutation distributions of correlation coefficients between individual SSR phase delays and randomly assigned optimal phase values across 10,000 iterations. The observed circular correlation was considered statistically significant when exceeding the upper 5% of the permuted distribution.
3. Results
3.1 Individualization of physical flicker luminance intensity
In order to account for interindividual differences in flicker frequency-dependent brightness enhancement, we first determined the discrimination threshold of perceptual differences between the LED8Hz and the LED12Hz without tACS. Therefore, we fit pychometric functions to the LED brightness discrimination performance at varying LED luminance ratios per participant (Fig. 2A). The mean luminance ratio between LED8Hz and LED12Hz at the 50% discrimination threshold was 0.96 with a mean standard error of 0.036 (Fig. 2B). Thus, on average participants perceived the LED8Hz as slightly brighter compared to the LED12Hz. The psychometric function was estimated on both testing sessions with high retest-reliability, as shown by a significant correlation between the mean luminance ratios within the 35% and 65% discrimination thresholds estimated per session (r = 0.84, p < .001). Yet, to ensure identical visual stimulation for the two tACS montages, participants were presented with luminance ratios determined on the first day on both testing sessions. Median peak luminance values for the LED8Hz ranged between 78 and 102 cd/m2 across brightness discrimination thresholds, while the LED12Hz had a constant peak luminance of 95 cd/m2.
Fig. 2Individualization of physical luminance intensities around flicker brightness discrimination thresholds. (A) Psychometric function for one exemplary participant. Individual luminance ratios between LED8Hz and LED12Hz were determined between the 35 and 65% brightness discrimination thresholds. (B) Distribution of the estimated luminance ratios between LED8Hz and LED12Hz across participants. (C) Bar diagram shows the proportion of trials on which the LED8Hz was judged as brighter, averaged across LED8Hz-tACS phase shifts during the main tACS experiment. As intended, depending on the luminance ratio between LED8Hz and LED12Hz, performance ranged between 35 and 65% brighter judgements. There was no statistical difference between average judgements during the occipital and periorbital tACS testing session. Error bars represent the standard error.
For the main tACS experiment, trials were presented at five luminance ratios around discrimination threshold, at each of the six tACS-LED8Hz phase shifts. To verify that participants accurately performed the brightness discrimination task, we averaged the proportion of brighter judgements of the LED8Hz across tACS-LED8Hz phase shift conditions, separately for each testing session. As shown in Fig. 2C, the proportion of trials on which the LED8Hz was rated as brighter ranged between 35% and 65% of judgements. Without taking flicker-tACS phase shifts into account, there was no significant difference between mean brighter ratings in the occipital and periorbital tACS condition across luminance ratios (t(35) = −0.78, p = .439, dz = -.13).
3.2 Interindividual differences in phase stability of visually evoked responses
A basic prerequisite for phase-specific interactions between visual and electric neuromodulation is the phase stability of the targeted flicker-evoked responses. We examined individual sensory entrainment by single flicker stimulation in the right lower visual field with spectral analysis of EEG data that was recorded prior to electrical stimulation. Amplitude spectra during visual stimulation showed peaks at the flicker stimulation frequencies and its harmonics (Fig. 3A ). To check for evoked neural phase stability, we computed the degree of SSR-flicker phase alignment for each participant (Fig. 3B). Phase locking values (PLVs) across all parieto-occipital electrodes were positively correlated between the two testing days (8 Hz: r = 0.61, p < .001; 12 Hz: r = 0.59, p < .001). Repeated measures ANOVA revealed that flicker phase entrainment was comparable between sessions and flicker stimulation frequencies (frequency: F(1,35) = 0.81, p = .373, = 0.023; session: F(1,35) = 0.64, p = .428, = 0.018; frequency∗session: F(1,35) = 0.08, p = .782, = 0.002). Thus, the two testing sessions did not differ significantly with respect to the detected brain state being targeted by tACS. Topographies of PLVs reflect the spatial specificity of electrophysiological flicker responses, with maximum SSR-flicker phase alignment in the hemisphere contralateral to the flickering stimulus presented in the right visual field (Fig. 3C).
Fig. 3Sensory entrainment by visual flicker. (A) EEG amplitude spectra averaged across participants reveal peaks at the stimulation frequencies of 8 and 12 Hz (and the first harmonic at 16 Hz). Amplitude values were averaged across all recorded parieto-occipital electrodes. (B) Steady-state response (SSR) during 8 Hz flicker stimulation shown exemplary for one participant at electrode PO7. The absolute value of the mean time-dependent differences between SSR and flicker phase reveal the strength of phase alignment by visual flicker. (C) Phase alignment to flicker frequencies averaged across all electrodes did not differ between the occipital and periorbital tACS condition. The error bars represent the standard error. Insert on the right shows the mean topography of EEG-flicker phase locking values (PLVs) averaged across both flicker frequencies, testing sessions and participants. Phase alignment is highest in visual areas of the left hemisphere contralateral to the flickering stimulus.
Fig. 4Phase-dependent modulation of flicker brightness perception by same-frequency tACS over the occipital cortex. (A) Left: Proportion of brighter judgements of the LED8Hz compared to LED12Hz for one exemplary participant under occipital tACS. The best LED8Hz-tACS phase shift associated with highest brightness ratings was aligned to the center bin (0°). For statistical analysis, phase-dependent modulation was quantified by the MAX-ADJ measure as illustrated. The insert shows data for the periorbital tACS montage. Right: No statistically significant overall effect of tACS phase on brightness perception was detectable when disregarding interindividual differences in phase locking of steady-state responses (SSR) to visual flicker. Error bars represent the standard error. (B) Only for the 8 Hz flicker, an electrode cluster showed a significant difference between tACS montages in the predictability of tACS-induced perceptual effects (MAX-ADJ) by SSR-flicker phase locking. Based on this electrode cluster, linear mixed model analysis revealed a significant interaction between tACS montage and SSR-flicker phase locking at 8 Hz. While brightness modulation under periorbital tACS did not relate to neural phase stability at 8 Hz, occipital tACS effects were dependent on successful flicker phase alignment (p = .009 < Bonferroni corrected α = 0.025). (C) The optimal flicker-tACS phase shift associated with greatest brightness judgements of the LED8Hz was correlated with the individual phase delay between LED8Hz onset and cortical SSR at 8 Hz.
3.3 Phase-dependent modulation of flicker brightness perception by occipital tACS
To investigate the degree of perceived flicker brightness modulation by the phase shift between visual flicker and tACS, we examined the proportion of brighter ratings of the LED8Hz relative to the LED12Hz for each of the six LED8Hz-tACS phase shift conditions. tACS phase-dependent perceptual modulation was quantified by a parametric alignment-based method (MAX-ADJ), with values greater than zero reflecting a cyclic modulation of brightness ratings across flicker-tACS phase shifts (Fig. 4A, left; see Supplementary Material B, Fig. B3 for average brightness ratings across the whole sample). Without taking interindividual differences in phase stability at the tACS-targeted rhythm into account, brightness modulations by tACS could neither be distinguished from zero (periorbital: t(35) = −0.86, p = .398, d = -.14; occipital: t(35) = 0.92, p = .363, d = 0.15) nor between tACS conditions with the statistically necessary certainty (t(35) = 1.41, p = .168, dz = 0.23) (Fig. 4A, right). However, this overall analysis crucially does not consider the mediating role of successful flicker-evoked phase stability for the detection of tACS effects. As stable flicker phase entrainment is the prerequisite for phase-specific flicker-tACS interactions, we investigated the dependency of the tACS effect on SSR-flicker phase locking per flicker frequency. Only for the 8 Hz flicker, permutation statistics revealed an electrode cluster showing a significant difference between tACS montages in the predictability of perceptual modulations by SSR-flicker phase locking (mean cluster correlation and p-value: r = 0.26, p = .009; topography in Fig. 4B, left). The cluster covered left-posterior scalp regions, contralateral to the flickering stimulus. The finding that differences in the tACS efficacy could only be predicted by neural phase stability at 8 Hz but not at 12 Hz supports a frequency-dependency of the individual flicker responsiveness [
]. Accordingly, phase locking values at 8 and 12 Hz averaged across cluster electrodes were not significantly correlated (sessionperiorbital: r = 0.21, p = .226; sessionoccipital: r = 0.29, p = .086), even though spanning a very similar range of values. Fig. 4B shows the relation between tACS-induced brightness modulation and SSR-flicker phase locking averaged across cluster electrodes for both frequencies. For this data, the frequency- and montage-specificity of tACS effects was further examined in a linear mixed model analysis. The analysis revealed a significant interaction effect between fixed factors “tACS montage” and “SSR-flicker phase locking at 8 Hz” (F(1, 40.89) = 6.19; p = .017). All other main and interaction effects were non-significant (see Supplementary Material B, Table B1). The statistically significant interaction term was broken down by conducting separate models for the two “tACS montage” groups on the fixed factor “SSR-flicker phase locking at 8 Hz”. Analysis showed that under occipital electrical stimulation, flicker entrainment strength significantly predicted the degree of perceptual modulation (F(1, 34) = 7.59; p = .009 < Bonferroni corrected α = 0.025). However, under periorbital tACS, SSR-flicker phase locking had no predictive value for the tACS efficacy (F(1, 34) = 0.81; p = .376) (Fig. 4B). Thus, the dependency of the tACS effect on SSR-flicker phase locking was specific for the occipital stimulation montage, the tACS-modulated flicker frequency at 8 Hz and showed a spatially specific cluster-distribution contralateral to the visual flicker.
3.4 Cortical phase delay of the SSR predicts the optimal timing of tACS application
While tACS is assumed to influence neuronal excitability with near zero phase lag, cortical SSRs showed variable phase lags relative to flicker onset. Accordingly, the optimal tACS-flicker phase shift that enhances brightness perception was expected to vary across subjects (see Supplementary Material B, Fig. B2). Circular correlation analysis showed that the SSR-flicker phase delay was predictive of the optimal tACS phase, only under occipital tACS (mean electrode correlation and p-value: r = 0.31, p = .028) (Fig. 4C).
4. Discussion
In this multimodal study, we showed that biasing the temporal pattern of cortical excitability by tACS can influence the subjective brightness perception of upcoming rhythmic visual input. Critically, this phase-specific modulatory effect of tACS was only found when electrically stimulating the visual cortex, but not for the retinal control montage. The degree of modulation depended on stable neural phase locking to the targeted visual flicker and the delay between visual presentation and cortical SSR. Taken together, our data are best explained by a phase-specific cortical interaction between flicker- and tACS-entrained rhythms, causally modulating neural temporal dynamics and related perception.
In this study, the combination of rhythmic visual and electrical stimulation allowed the investigation of tACS effects with control over the phase of the targeted neural rhythm. Previous studies assessed the influence of tACS on ongoing or task-related activity in visual cortex [
] with only low neural frequency- and phase-consistency of the targeted process, possibly hampering rigorous investigations of neural entrainment effects. In contrast, visual flicker stimulation enables the setting of the oscillatory phase in the visual cortex with high signal-to-noise ratio. Irrespective of whether SSRs reflect entrained intrinsic oscillations or repetitive evoked neural responses [
Modification of brain oscillations via rhythmic light stimulation provides evidence for entrainment but not for superposition of event-related responses.
], they allow a precise targeting by tACS. Yet, for conclusive data analysis, consideration of interindividual differences in flicker-evoked neural phase stability is crucial. Even if tACS had entrained neural activity in our whole sample, phase-specific flicker-tACS interactions are only measurable when phase control by visual flicker is sufficiently high. This emphasizes that the systematic assessment of variables that are assumed to mediate the tACS efficacy or the measurability of stimulation effects are of utmost relevance. In this highly controlled stimulation paradigm, we thereby showed that occipital tACS biases brightness perception in a frequency- and phase-specific manner.
Specifically, perceptual changes were positively correlated with the strength of neuronal phase locking to the tACS-targeted 8 Hz flicker, but not to the 12 Hz reference flicker. This finding supports neural phase stability at the tACS-targeted rhythm as the prerequisite for observing phase-specific interactions between SSR and applied electric field [
], and favors an interaction of entrained same-frequency rhythms. Importantly, mean brightness ratings across flicker-tACS phase shifts did not differ between occipital and periorbital tACS, implying a phase-dependent enhancement and reduction of perceived brightness under occipital electrical stimulation. In the same vein, our previous study demonstrated enlarged and suppressed SSR amplitudes depending on the flicker-tACS phase shift [
]. Taken together, the clear relation between tACS-flicker phase shift and brightness perception strongly points towards an interaction of entrained rhythms in visual cortex that can give rise to SSR amplitude changes and drive perceptual modulations.
The optimal timing of electrical neuromodulation relative to flicker onset leading to highest brightness judgements showed pronounced interindividual variability (see Supplementary Material B, Fig. B2). This is in line with previous studies showing that the optimal tACS phase that enhances perception varies uniformly across participants [
]. Crucially, entrainment of the temporal pattern of neural activity by tACS is expected to facilitate or hamper concurrent visual processing depending on their relative temporal alignment. While tACS can instantaneously affect neural activity, SSRs typically show varying phase delays relative to flicker onset, likely due to interindividual differences in SSR propagation along the visual pathway and anatomical differences [
]. In the present data, circular correlation analysis showed a predictive value of the cortical 8 Hz SSR phase delay for the optimal timing of tACS application, exclusively for the occipital tACS montage. As the predictive value of the cortical phase cannot be explained by an interaction between visual input and electric field in the retina, the correlation additionally speaks against a mediation of occipital tACS effects on SSRs by retinal co-stimulation. Rather, these data are in support of a neuronal interaction of visually and electrically driven neuromodulation at a cortical level.
By interfering with cortical visual processing via tACS while keeping physical luminance properties constant, our data corroborate potential mechanistic insight into the neural substrate of perception. As tACS has previously been shown to induce shifts in neuronal spike timing [
], applied electric fields in our study are expected to have biased the temporal pattern of cortical excitability. These excitability modulations may influence the net synchrony in population firing to upcoming visual input, that generates SSR amplitude changes on a neural population level. Flicker-evoked response amplitudes have been repeatedly shown to correlate with the subjective experience of brightness [
]. As electrophysiological recordings during the application of tACS are contaminated by electrical artifacts, the direct measurement of SSR amplitude changes during tACS was not possible [
]. However, our simulations of neuronal firing patterns to concurrent dual flicker and tACS input (Supplementary Material A) as well as our previous study results [
] demonstrate that this SSR amplitude modulation is feasible, and that it depends on the degree of phase alignment of neural activity to the flicker. The behavioral data observed here reproduce this relation between phase stability of tACS-targeted SSRs and perceptual modulations. Neural populations responding only weakly to visual input might be particularly susceptible to low intensity tACS [
]). Thus, extending previous correlative results on the physiological substrate of perception, data suggest a causal relation between brightness perception and the amplitude of rhythmic neural population activity in visual cortex.
A further mechanism that may additionally influence neural responsiveness to visual input builds on pulsed inhibition of evoked responses by tACS-entrained intrinsic alpha oscillations. Besides oscillatory amplitudes, the phase of ongoing alpha rhythms is assumed to exert a cyclic inhibitory influence on neuronal excitability [
]. Accordingly, the systematic flicker-tACS phase shifts may have varied the coincidence of luminance peaks with different phases of the tACS-entrained alpha cycle, generating brain-state dependent effects on evoked responses and brightness perception. To experimentally examine these potential mechanisms of tACS, invasive recordings would be necessary that allow the simultaneous assessment of local field potentials and neuronal spiking patterns. Presumably, tACS may exert its effect via multiple ways of action, including a phase-specific modulatory influence on endogenous alpha rhythms as well as on neural populations encoding sensory information.
For tACS applications, it is crucial to consider the role of potential peripheral stimulation effects that might add to transcranial neuromodulation. Innervation of cranial and peripheral nerves in the skin by tACS may induce additional rhythmic activation in the somatosensory cortex. In one study, this co-stimulation confound was shown to mediate neural entrainment effects of motor cortex tACS, with the target region being in close proximity to the somatosensory system [
]. Yet, when using active control site stimulation or EMLA cream for blocking cutaneous afferents, recent studies convincingly supported a direct transcranial neural efficacy of tACS [
]. By applying EMLA cream in our study, participants’ ratings of skin sensation were on average low to moderate and showed no correlation with the strength of perceptual modulation by tACS (Supplementary Material B, Fig. B1). This finding conforms with recent work by Kasten and colleagues [
] demonstrating that alpha power enhancements by tACS over the parieto-occipital cortex could be predicted by electric field variability within the brain, but not by current strength in the skin. Taken together, this data supports the assumption that flicker brightness changes were to the most part mediated by direct transcranial modulation of occipital neural dynamics. Yet, to completely exclude an involvement of somatosensory co-stimulation, a control electrode montage matched for somatosensation would be needed. The dependency of transcutaneous co-stimulation effects on the area of stimulation makes the development of valid tACS control montages a challenging goal for future studies. Even more critical for visual perception research is the possible co-stimulation of the retinae due to current shunting across the scalp [
]. The direct electrical stimulation of photoreceptors or ganglion cells could enhance neural responses along the visual hierarchy, thereby influencing perceived brightness. To encounter this methodological confound, our active tACS control montage was explicitly tailored to induce subthreshold stimulation of the retinae. Estimated electric field strengths in the eyes under periorbital tACS were comparable to the expected ocular electric field strength under occipital stimulation. Importantly, as periorbital tACS had no modulatory influence on brightness perception, data strongly suggest that occipital tACS effects were not mediated by subthreshold stimulation of the retinae. The inclusion of tACS control conditions along with simulations of electric fields might, therefore, be central to reach substantive conclusions on the tACS efficacy.
In conclusion, our findings highlight the importance of temporally coordinated activity in visual cortex for subjective brightness perception. We propose that the phase-specific interaction between flicker- and tACS-related activity shaped subjective brightness perception via amplitude modulations of population activity in visual cortex. Our data corroborate the capability of tACS to transmit perceptually relevant temporal information to the human cortex, but also underlines that successful proof for its efficacy is dependent on individual functional properties, emphasizing the necessity for custom-fit stimulation protocols. Transfer of the advantageous methodological approach applied here to other sensory modalities may further help to advance knowledge on the link between cortical sensory processing and perception. Thus, by controlled modulations of brain signals, tACS can be utilized to deepen our understanding of human brain function in basic and clinical science.
CRediT authorship contribution statement
Marina Fiene: Conceptualization, Methodology, Investigation, Software, Formal analysis, Validation, Visualization, Data curation, Writing – original draft, Writing – review & editing. Jan-Ole Radecke: Conceptualization, Methodology, Writing – review & editing. Jonas Misselhorn: Conceptualization, Methodology, Writing – review & editing. Malte Sengelmann: Conceptualization, Methodology, Writing – review & editing. Christoph S. Herrmann: Conceptualization, Methodology, Writing – review & editing. Till R. Schneider: Conceptualization, Methodology, Writing – review & editing, Funding acquisition. Bettina C. Schwab: Conceptualization, Methodology, Writing – review & editing, Supervision. Andreas K. Engel: Conceptualization, Methodology, Writing – review & editing, Funding acquisition, Project administration.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was supported by the Deutsche Forschungsgemeinschaft (SFB 936/A3 awarded to A.K.E. and T.R.S.; SPP 1665/EN 533/13-1 and SFB TRR 169/B1 awarded to A.K.E.; SPP 1665/SCHN 1511/1–2 awarded to T.R.S.) and by the Studienstiftung des deutschen Volkes (awarded to M.F.). We thank Karin Deazle and Rebecca Burke for assistance in data recording, Florian Pieper for methodological support and Florian Göschl for helpful discussions on the data.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Modification of brain oscillations via rhythmic light stimulation provides evidence for entrainment but not for superposition of event-related responses.
Noninvasive transorbital alternating current stimulation improves subjective visual functioning and vision-related quality of life in optic neuropathy.
00001-8&id=gr1.jpg){kind=link}
00001-8&id=gr2.jpg){kind=link}
00001-8&id=gr3.jpg){kind=link}
00001-8&id=gr4.jpg){kind=link}