Multisensory perceptual awareness: Categorical or graded?

Area of Science:

• Cognitive Neuroscience
• Computational Neuroscience
• Psychophysics

Background:

• Neural mechanisms for conscious access are often described as all-or-none.
• The nature of phenomenal consciousness (subjective experience) as categorical or graded remains debated.
• Existing models often link consciousness to global neural broadcasting after a threshold is met.

Purpose of the Study:

• To investigate whether phenomenal consciousness is categorical or graded.
• To model consciousness using artificial neural networks with feedback connectivity.
• To test model predictions with human psychophysical experiments.

Main Methods:

• Constructed artificial neural networks with varying feedback connections between multisensory and unisensory cortices.
• Simulated network responses to unisensory and multisensory stimuli.
• Conducted a multisensory cue-congruency task with 29 human subjects over 10 hours.

Main Results:

• Artificial neural networks showed intermediate reaction times for multisensory stimuli associated with unisensory feedback.
• Human psychophysical testing revealed intermediate reaction times for multisensory cues reported as unisensory.
• Results align with models predicting graded phenomenal consciousness.

Conclusions:

• Phenomenal consciousness can exhibit graded properties, not strictly all-or-none.
• Simple neural networks, even those based on all-or-none principles, can instantiate graded consciousness.
• Findings challenge the dichotomy of conscious vs. unconscious states, supporting intermediate subjective experiences.