The triple-flash illusion reveals a driving role of alpha-band reverberations in visual perception

Rasa Gulbinaite, Barkın İlhan, Rufin VanRullen

Abstract: The modulatory role of spontaneous brain oscillations on perception of threshold-level stimuli is well established. Here, we provide evidence that alpha-band (7-14 Hz) oscillations not only modulate but also can drive perception. We used the “triple-flash” illusion: Occasional perception of three flashes when only two spatially-coincident veridical ones are presented, separated by ~100 ms. The illusion was proposed to result from superposition of two hypothetical oscillatory impulse response functions (IRF) generated in response to each flash (Bowen, 1989). In Experiment 1, we varied stimulus onset asynchrony (SOA) and validated Bowen's theory: the optimal SOA for illusion to occur was correlated, across subjects, with the subject-specific IRF period. Experiment 2 revealed that pre-stimulus parietal alpha EEG phase and power, as well as post-stimulus alpha phase-locking, together determine the occurrence of the illusion on a trial-by-trial basis. Thus, oscillatory reverberations create something out of nothing – a third flash where there are only two.

Introduction

The human brain is constantly active, with spontaneous oscillations in various frequency bands. Among these, the alpha rhythm (7-14 Hz) is particularly prominent in the visual system. Traditionally, these oscillations have been viewed as having a modulatory role—acting as a gatekeeper that biases the likelihood of perceiving a stimulus based on its timing relative to the phase of the wave. However, a more provocative hypothesis suggests that these oscillations might play a "driving" role, actively generating perceptual content. This study investigates this possibility using the "triple-flash" illusion.

The Triple-Flash Illusion and the Oscillatory Model

The triple-flash illusion is a phenomenon where two brief, spatially coincident flashes are perceived as three when separated by a specific temporal interval, typically around 100 ms. This illusion provides a unique window into the brain's temporal processing.

Experimental paradigm and the oscillatory superposition model for the triple-flash illusion.

As illustrated in Figure 1, the illusion is thought to arise from the superposition of oscillatory impulse response functions (IRFs). When the brain receives a visual stimulus, it doesn't just respond with a single spike of activity; instead, it "rings" like a bell at an alpha frequency. If a second flash occurs at the right moment (e.g., SOA = 100 ms), the oscillatory responses to both flashes sum together. Under the right conditions, this constructive interference produces a third peak that exceeds the threshold for conscious perception, resulting in the perception of a third, illusory flash. Conversely, if the flashes are too close (SOA = 50 ms) or too far apart (SOA = 150 ms), the peaks do not align to create this additional suprathreshold event.

Methods and Experimental Design

The researchers conducted two experiments to validate this oscillatory driving hypothesis. In Experiment 1, they sought to link individual differences in brain "ringing" to the illusion. They used EEG to record responses to white-noise luminance flicker, allowing them to extract a subject-specific IRF. They then varied the Stimulus Onset Asynchrony (SOA) between two flashes to find the optimal timing for each participant to experience the illusion.

Experiment 2 focused on trial-by-trial dynamics. By fixing the flash interval to the optimal SOA (~87.5 ms), the team could compare the brain's state on trials where the illusion was perceived versus trials where it was not. They employed Independent Component Analysis (ICA) to isolate signals from the occipital and parietal cortices, providing better spatial localization of the underlying rhythms.

Key Findings

The study yielded several critical insights into the nature of visual perception:

  • Temporal Matching: A significant correlation ($r = 0.51$) was found between a participant's internal IRF period and the specific SOA that most frequently triggered the illusion. This directly supports the idea that the illusion is driven by the brain's inherent oscillatory frequency.
  • Parietal Dominance: Interestingly, the "driving" force was more strongly associated with alpha rhythms in the parietal cortex than the primary visual (occipital) cortex. This suggests that the "filling in" of the illusory flash is a higher-order integrative process.
  • Pre-stimulus Influence: The likelihood of seeing the illusion was predicted by the brain's state before the flashes even appeared. Lower alpha power and a specific alpha phase in the parietal region significantly increased the probability of perceiving the third flash.
  • Phase Alignment: On trials where the illusion occurred, there was significantly stronger phase-locking to the stimuli, indicating that the illusion depends on the precise alignment of the external stimuli with internal rhythms.

Discussion and Implications

These results demonstrate that alpha-band oscillations do more than just modulate sensitivity; they can actively drive perception by creating "something out of nothing." The finding that participants perceived the illusory flash on approximately 41-45% of trials at optimal timing highlights the robustness of this effect.

The study also has potential clinical relevance. Previous research has noted that patients with schizophrenia often have a "slower" threshold for the triple-flash illusion (130-150 ms). The current findings suggest this may be a direct consequence of slowed internal alpha rhythms, potentially offering a biomarker for temporal processing deficits in clinical populations.

Future research using techniques like repetitive Transcranial Magnetic Stimulation (rTMS) could attempt to causally manipulate these rhythms, entraining the brain at specific frequencies to see if the illusion rate can be artificially controlled. This would further solidify the "driving" role of alpha oscillations in constructing our visual reality.

Summary updated

Abstract

ABSTRACT The modulatory role of spontaneous brain oscillations on perception of threshold-level stimuli is well established. Here, we provide evidence that alpha-band (7-14 Hz) oscillations not only modulate but also can drive perception. We used the “triple-flash” illusion: Occasional perception of three flashes when only two spatially-coincident veridical ones are presented, separated by ~100 ms. The illusion was proposed to result from superposition of two hypothetical oscillatory impulse response functions (IRF) generated in response to each flash (Bowen, 1989). In Experiment 1, we varied stimulus onset asynchrony (SOA) and validated Bowen's theory: the optimal SOA for illusion to occur was correlated, across subjects, with the subject-specific IRF period. Experiment 2 revealed that pre-stimulus parietal alpha EEG phase and power, as well as post-stimulus alpha phase-locking, together determine the occurrence of the illusion on a trial-by-trial basis. Thus, oscillatory reverberations create something out of nothing – a third flash where there are only two.
BibTeX
@article{gulbinaite2017triple,
  author = {Rasa Gulbinaite and Barkın İlhan and Rufin VanRullen},
  title = {{The triple-flash illusion reveals a driving role of alpha-band reverberations in visual perception}},
  journal = {Journal of Neuroscience},
  year = {2017},
  doi = {10.1523/jneurosci.3929-16.2017},
}
Figure