Fork the consciousness, or download the project and create your own.

Rufin VanRullen

CNRS / Centre de Recherche Cerveau et Cognition (CerCo), Toulouse

Integrated information Consciousness is intrinsic causal power, measured as integrated information

Rufin VanRullen is a Research Director at the French National Centre for Scientific Research (CNRS) and head of the “Perceptual and Cognitive Dynamics” team at the Centre de Recherche Cerveau et Cognition (CerCo) in Toulouse. His research sits at the intersection of visual psychophysics, electrophysiology, and artificial neural networks.

VanRullen has extensively researched the debate over whether conscious perception is a continuous flow or a sequence of discrete temporal frames. Through psychophysical experiments (such as the continuous wagon-wheel illusion) and EEG phase analysis, he demonstrated that attention and visual consciousness operate periodically, synchronized to alpha and theta brain rhythms. Sensory information is sampled in discrete rhythmic bursts that are subsequently interpolated into an apparent continuous stream.

In recent work combining neuroscience with AI, VanRullen applies oscillatory and recurrence mechanisms to deep neural networks, demonstrating that periodic temporal gating and feedback loops are necessary to achieve human-like perceptual stability and conscious scene parsing.

The discrete-sampling position rests on several converging lines of evidence from his laboratory. The continuous wagon-wheel illusion, where rotating wheels appear to slow or reverse under continuous illumination, indicates that some stage of visual processing samples the input in snapshots rather than continuously. His group reported 10 Hz perceptual echoes in human EEG, where a repeated visual stimulus reappears in the brain’s response at the alpha rhythm, and showed that the phase of ongoing alpha oscillations at stimulus onset predicts whether and how strongly a target is perceived. He consolidated these findings in the perceptual cycles account, which proposes that attention sweeps the visual scene periodically, around 8 to 12 times per second, and that each sweep constitutes one sample of the conscious moment.

The account matters for consciousness science because it imposes a temporal resolution limit. If conscious perception advances in discrete samples, then any theory of consciousness, biological or artificial, must specify the sampling rate, the carrier rhythm, and the interpolation mechanism that produces continuity. His deep-learning work draws the engineering consequence. A feedforward network running continuously has no natural sampling frame, so his team equips convolutional and recurrent architectures with oscillatory gating and feedback, testing whether rhythm-driven sampling reproduces the stability and segmentation that biological vision shows. The site’s analysis of his oscillatory binding program examines that research program in detail.

Known for. Perceptual rhythms, Brain oscillations, Continuous vs. discrete temporal frames in visual consciousness, deep learning models of neural perception

Coverage on this site

Related researchers