Can we “see” without being consciously aware of it?

Deep states of meditation and psychedelic‑induced experiences are often described as similar: intense, transformative, and difficult to put into words. But what truly shapes these experiences and the way they are narrated? To what extent do they depend on the substance, the meditative practice, or the context in which they are lived and interpreted?

Recognising a face seems immediate and automatic. But what happens in the brain when an image is so faint that we barely notice it? And can brain activity distinguish a face that is consciously perceived from a visual stimulus that goes unnoticed?

To answer these questions, Adélaïde de Heering and colleagues investigated whether a neurophysiological technique known as frequency‑tagging could serve as an objective marker of the conscious perception of faces. This technique is based on recording the brain’s electrical activity (EEG) while images are presented rapidly and rhythmically, allowing highly specific neural responses to particular visual categories, such as faces, to be identified.

In the study, participants viewed long sequences of images presented at a constant rate of six images per second. Periodically, a face appeared among these images at a fixed interval, making it possible to “tag” the brain’s electrical response specifically associated with faces.

Participants viewed these sequences at two very low levels of contrast: 1%, at which faces were practically invisible, and 1.5%, slightly above the threshold required for conscious perception. At the end of each sequence, participants were asked to categorise the gender of the face, rate how visible it had been, and report their level of confidence in their response. Throughout the task, attention was carefully monitored to ensure that any differences in the results could not be explained simply by fluctuations in attention.

The results were clear. At 1% contrast, participants were unable to identify face gender above chance level, reported little or no visibility, and the brain showed no face‑specific response. Under these conditions, there were no neural signatures indicating conscious perception. At 1.5% contrast, however, participants were able to identify faces correctly, reported greater visibility, and expressed confidence in their decisions. In these sequences, a robust and selective brain signal associated with faces also emerged, recorded in brain regions known to be involved in face processing. Moreover, the strength of this signal increased gradually with both subjective visibility and reported confidence.

This study shows that, within this paradigm, the brain response associated with faces does not merely reflect stimulus presentation or visual attention but closely tracks conscious experience. The findings therefore reinforce the idea that conscious perception, even when subtle and close to the threshold of visibility, leaves measurable traces in brain activity. This study was published in the scientific journal Cortex, in the article Frequency-tagging as a measure of conscious face perception, as a part of research project 50/22 - The steady-state visual evoked potential (SSVEP) tool as a marker of subjective visibility, supported by the Bial Foundation.

Abstract

Steady-state visual evoked potentials (SSVEPs) from frequency-tagging (FT) paradigms are widely used to investigate visual processing. Yet their association to conscious perception remains unclear. To test whether SSVEP occurs during conscious perception or without consciousness, 32 participants saw sequences of different images presented at 6 Hz (6 images per second) and containing faces every fifth image (1.2 Hz). All images were presented at either 1% contrast or 1.5% contrast. After each sequence, participants had to categorise the face gender (objective perception) and rate their confidence in this categorisation and the visibility of the faces (subjective perception). During the sequence presentation, participants' attention was monitored via an orthogonal fixation-cross task. Results showed that, at 1.5% contrast, the face signal was higher during correct than during incorrect gender categorisation and increased linearly with both visibility and confidence ratings. In line with participants’ performance on the fixation-cross task, the signal collected at 6 Hz also indicated that attention related more closely to confidence than to visibility. At 1% contrast however, no face perception occurred behaviourally, which was confirmed by the absence of brain signal recorded in response to face instances. Overall, these findings show that SSVEPs can track both the objective and subjective perception of faces at a conscious contrast (1.5%). These findings bring new evidence that SSVEPs can be used as a marker of conscious perception.