For decades, cognitive psychology and neuroscience have studied perceptual consciousness, which refers to the subjective experience each person has of information arriving through their senses. Two of the most interesting questions in this field are the brain regions involved in this process and what distinguishes information we perceive consciously from that which never reaches our awareness.
A recent study by researchers from the Information and Communications Technologies Research Center (CITIC) and the Mind, Brain and Behavior Research Center (CIMCYC) at the University of Granada used a combination of electroencephalography (EEG) and machine learning (ML) techniques to investigate how our brain represents information we perceive both consciously and unconsciously.
To achieve this, a group of volunteer participants was asked to identify visual stimuli that appeared and disappeared quickly, indicating whether they had seen them. While they performed the task, their brain activity was recorded using EEG. The main results show that it is possible to reliably decode both the presence of the stimulus (whether it appeared or not) and the subjective perception of the individuals (whether they saw it or not) from the EEG signals. In addition, the results demonstrated that decoding improves significantly when using time-frequency representations of the EEG signals, which underscores the importance of brain oscillations, especially in the theta and alpha bands.
Regarding unconscious information, the study found evidence that our brain processes stimuli even when we are not aware of them. Although these unconscious representations were less stable and more fleeting than conscious ones, they were observed in the early stages of perception (~100 ms) and during response preparation. Interestingly, the presence of unconscious information could speed up people's responses, but without improving the accuracy of their decisions (i.e., they responded faster but with low precision).
A Range of Potential Applications
The ability to reliably decode the presence of stimuli and subjective perception, even distinguishing between conscious and unconscious processing, provides a powerful tool for future research on the minimum neural mechanisms required for conscious experience. Furthermore, the methodological advance of using time-frequency representations is crucial, as it offers a richer and more sensitive approach for analyzing EEG data in general, applicable to various areas of cognitive neuroscience.
In summary, this research highlights the great potential of ML algorithms with EEG to understand how conscious and non-conscious information are represented in the brain, showing distinct neural dynamics for each.
Reference
Rodríguez-San Esteban, P., Gonzalez-Lopez, J.A. & Chica, A.B. (2025) Neural representation of consciously seen and unseen information. Scientific Reports, 15, 7888. https://doi.org/10.1038/s41598-025-92490-y
Contact
Pablo Rodríguez-San Esteban (@email)