Differentiating levels of awareness: an EEG and time-frequency analysis of subliminal, preconscious and conscious processing

Despite decades of intense investigation across philosophy, psychology and neuroscience, the precise neural signatures and dynamics underlying the conscious awareness remain a significant challenge. Not to mention the attempt to delineate the different levels or states of visual awareness, or the transition between unconscious states and full conscious perception. The present study aims to investigate the neural correlates of consciousness in the time-frequency domain, as well as measuring them in the two stages of unconsciousness to characterise exclusively subliminal, preconscious and conscious states. In accordance with the most widely accepted theories supporting visual awareness, the hypotheses are as follows: firstly, the total absence or strongly reduced presence of certain patterns in the subliminal context and secondly, the presence in the preconscious state of significant variations almost superimposed on the conscious state, following the presentation of a visual stimulus. The experiment involved 66 participants who were subjected to the presentation of masked visual stimuli in order to simulate the two stages of visual unawareness, using two paradigms already documented in the literature. The subjects' brain responses to the stimuli were recorded by an EEG headset (Emotiv EPOC Flex) worn by the participants throughout the test. Depending on the expected contribution of each task, participants were asked either to interact with the keyboard or respond to a brief questionnaire. All signals were filtered, processed and divided into epochs using the Matlab's EEGLab extension, which was also used to extract the wavelet transform coefficients for power calculation. In the time-frequency domain, the five oscillations (delta, theta, alpha, beta and gamma) were analysed in terms of amplitude, complexity and phase coherence inter-trial and between different regions. The data extracted from the questionnaire were utilised in the statistical analysis alongside the time-frequency characteristic variables. Non-parametric tests were adopted for the statistical analysis to detect potential differences between the variables associated with the two stages of unawareness. The analysis revealed distinct patterns of brain activity across the investigated states of awareness. Specifically, average power demonstrated significant differences in various frequency bands, showing unique profiles in distinguishing between subliminal and preconscious stages. Furthermore, differential power responses were observed for certain oscillations across various preconscious and subliminal phases, with some patterns aligning with existing literature. Beyond power, preliminary explorations into inter-trial phase coherence and connectivity between different brain regions suggested additional neurophysiological distinctions across these states of awareness. Finally, spectral entropy provided further evidence for significant statistical differences between the subliminal and preconscious stages, supporting expected dissociations.