(A) Trial illustration. The target/distractor cue indicates the orientation of the corresponding grating. For half of the participants, target cues were in solid lines and the distractor cues were in dashed lines, while the rules were reversed for the other half participants. Gratings were presented to each eye using a stereoscope, with the target and distractor gratings flickering at 24 Hz and 20 Hz, respectively. During binocular rivalry, participants were instructed to prioritize the target grating while suppressing the distractor grating, aiming to reproduce the color of the target grating as precisely as possible. (B) Behavioral results. Response deviation represents how many degrees a response was off from the target color. From a probabilistic mixture model, we extracted the probability of reporting target color (pT), distractor color (pD) and random guess (pU) (Figure S1). Two-way ANOVA was performed on behavioral measurements. Results are shown for target cueing (TgtCue) and distractor cueing (DistCue) conditions, separately for when target was at the dominant eye versus when distractor was at the dominant eye. *, p < 0.05, **, p < 0.01, ***, p < 0.001. Probability of reporting target color, main effect of cueing: F(1, 140) = 9.939, p = 0.002, η2 = 0.133; main effect of stimuli dominance: F(1, 140) = 3.452, p = 0.065; cueing x stimuli dominance, F(1,140) = 4.612, p = 0.033, η2 = 0.098). Probability of reporting distractor color: main effect of cueing: F(1, 140) = 10.859, p = 0.001, η2 = 0.154; main effect of stimuli dominance: F(1, 140) = 3.297, p = 0.072; cueing × stimuli dominance, F(1,140) = 4.370, p = 0.038, η2 = 0.090).

(A) Signal-to-noise ratio (SNR) of SSVER. The top panel shows normalized SNR using all channels. The topography highlights the most responsive channels to SSVER. (B) Mahalanobis-based distance decoding of target and distractor grating orientations in different cueing conditions. Colored lines above the x-axis represent time intervals with above-chance decoding accuracy, identified using cluster-based permutation testing. Black lines indicate the decoding accuracy difference between target and distractor gratings. Inserted panels show averaged decoding slopes calculated between 0 and 2 s during the rivalry phase. (C) Orientation decoding of target and distractor gratings based on stimulus dominance. Black asterisks mark significant differences between cueing conditions, while colored asterisks indicate significant decoding above chance level 0. For target dominant trials, target grating in TgtCue, t(35) = 5.916, p < 0.001, Cohen’s d = 0.986; distractor grating in TgtCue, t(35) = 1.852, p = 0.072, Cohen’s d = 0.309; target grating in DistCue, t(35) = 2.721, p = 0.010, Cohen’s d = 0.454; distractor grating in DistCue, t(35) = 3.887, p < 0.001, Cohen’s d = 0.648; For distractor-dominant trials, target grating in TgtCue, t(35) = 4.778, p < 0.001, Cohen’s d = 0.796; distractor grating in TgtCue, t(35) = 1.677, p = 0.103, Cohen’s d = 0.279; target grating in DistCue, t(35) = 4.611, p < 0.001, Cohen’s d = 0.769; distractor grating in DistCue, t(35) = 5.559, p < 0.001, Cohen’s d = 0.926.

(A) Parietal alpha activity. The time-frequency map shows the power difference between target and distractor cueing conditions at electrode P2 (marked white). The right lateralized topographical distribution is consistent with previous findings showing the right posterior intraparietal sulcus in anticipatory alpha modulation (Capotosto et al. 2011). Source reconstruction revealed brain regions showed stronger alpha power (8-12 Hz) in distractor cueing condition compared to target cueing condition. Colorbar indicates t-value from cluster-based permutation test. The raincloud plot shows individual alpha power difference between cueing conditions. The inserted topography displays the averaged alpha power difference between cueing conditions during 0 to 1.2 s after cue onset. (B) Frontal theta activity. The time-frequency map shows the power difference between distractor and target cueing conditions, with the highlighted cluster indicating time-frequency points that passed cluster-based permutation testing. These results were derived from frontal channels highlighted in the topography. Source reconstruction revealed enhanced theta power in the distractor cueing condition during the rivalry phase. (C) Orientation decoding of cued orientations using posterior channels. Left, orientation decoding of each cueing condition separately; Right, robust cross-condition generalization (T2D: train target cueing data, test distractor cueing data; D2T: reverse) indicates that target and distractor templates utilize shared sensory-level representations. (D) Distinct task states. Multivariate classification of cueing conditions (TgtCue vs. DistCue) reveals distinct neural profiles throughout the delay, suggesting that the functional role of the template is maintained. The classifier was trained on data from all EEG.

(A) Target SSVER and distractor SSVER jointly affect behavioral performance. Left, heatmap of response deviation (z-scored) as a function of target and distractor SSVERs (binned into 50 units) in target-dominant trials. The color of the scatter dots represents the average z-scores of behavioral response deviation within each bin. Dark regions indicate bins with insufficient trials. The top and right histograms show the probability density of target SSVER and distractor SSVER respectively (scale: 0 to 0.1). Right, linear regression coefficients for distractor SSVER in predicting deviation, median-split by target SSVER. Interference is significantly higher when target signals are weak (blue line). The intercept is not shown; the y-axis units are arbitrary reflecting relative values. (B-C) Frontal theta oscillations facilitate reactive suppression of distractors in target-dominant trials. (B)Increased frontal theta power (z-scored) correlates with reduced behavioral response deviation. (C)Stronger theta power is specifically associated with attenuated distractor SSVER, while target SSVER remains unaffected, suggesting a distractor-specific inhibitory mechanism. For visualization purpose, trials were median split based on frontal theta power of each participant to facilitate graphical interpretation of the effects. All statistical analyses were conducted on continuous single-trial data, and the categorical grouping was used solely for graphical presentation. (D-E) Parietal alpha power indexes indirect gating without altering sensory gain in distractor-dominant trials. (D) Under conditions of weak sensory gain (low SSVER), higher preparatory parietal alpha power is associated with smaller response deviations and an increased probability of target reporting (pT). (E) Preparatory alpha power does not directly modulate the SSVER of either target or distractor stimuli, supporting a role in higher-level gating rather than direct gain control.

Reactive inhibition and proactive gating in resolving visual competition.

Reactive inhibition via frontal theta (left panel): When the target is presented to the dominant eye, it establishes initial perceptual dominance. In this state, reactive inhibition, indexed by increased frontal theta activity, is recruited to suppress the competing distractor by directly reducing its sensory gain (red sinusoidal curve). This suppression stabilizes the target’s dominance and slows down perceptual switching, thereby optimizing behavioral performance. Proactive attentional gating via parietal alpha (right panel): When the distractor occupies the dominant eye, high-uncertainty competition is resolved through proactive gating. Elevated preparatory alpha activity signals the instantiation of a negative attentional template based on prior distractor knowledge. This template serves to segregate and differentially route competing inputs without altering early sensory gain. This gating mechanism is most advantageous under conditions of high perceptual uncertainty, where it stabilizes weak target signals and facilitates the high-level processing. DE, dominant eye; NDE, nondominant eye; VC, visual cortex; IPL, inferior parietal lobe; FC, frontal cortex.