A study published in Journal of Experimental Psychology: General found that when people are unaware of a visual cue, their brain processes it separately from their expectations, but when they become aware of it, these two processes start to interact.
Understanding how conscious awareness influences perception is a central question in neuroscience and psychology. The global neuronal workspace theory suggests that conscious perception emerges when top-down (TD) decision-driven processing and bottom-up (BU) sensory-driven processing integrate within overlapping neural networks.
Prior research has established that both objective sensory information and subjective interpretation influence behavior, but how these streams interact depending on conscious awareness is unclear.
Research using priming paradigms has demonstrated that subliminal primes (stimuli not consciously perceived) can still influence responses to subsequent targets. However, whether the neural mechanisms of TD and BU processing remain separate or integrate during different levels of awareness has not been thoroughly tested.
In a new study, Ze-Fan Zheng and colleagues investigated this question by designing two experiments to determine whether TD and BU processing remain independent in unconscious conditions or merge when stimuli are consciously perceived.
The researchers utilized two experiments using variations of the Stroop priming paradigm. The first experiment involved 31 participants. This study used a continuous flash suppression (CFS) technique to render the prime word—either the Chinese character for “yellow” or “blue”—invisible by presenting it to one eye while rapidly changing visual noise (Mondrian patterns) was shown to the other.
This technique ensured that the prime remained undetected in certain conditions while its visibility varied across different trials. Following prime exposure, participants viewed a color target (either yellow or blue) and indicated its color. They then attempted to identify the prime word in a two-alternative forced-choice (2AFC) discrimination task. To manipulate awareness systematically, the researchers adjusted the prime’s contrast and categorized trials based on how well participants could consciously perceive it.
The second experiment reanalyzed data from a previous study (Sand & Nilsson, 2017) that used backward masking to manipulate prime visibility. Here, 67 participants were exposed to an English color word (either “RED” or “BLUE”) for six milliseconds, followed by a visual mask that prevented conscious perception in some trials. The target stimulus (a colored frame in either red or blue) appeared immediately afterward, requiring participants to identify its color.
Unlike the first experiment where prime contrast varied, in this experiment, the physical characteristics of the prime remained unchanged across trials. This allowed researchers to assess subjective visibility independently of stimulus strength. Participants reported their perceived visibility of the prime using a three-level scale (“no percept,” “unclear percept,” or “clear percept”) and attempted to identify the prime word in a 2AFC task. Researchers categorized trials as either “invisible” or “visible” based on subjective ratings.
The findings were compelling. In the first experiment, when participants were unable to consciously perceive the prime (as determined by low prime discrimination accuracy), TD and BU congruency effects influenced reaction times independently, meaning the two types of processing operated in separate neural systems. However, as prime visibility increased, an interaction between TD and BU effects emerged, indicating the two processes started to merge when the prime was consciously perceived.
Specifically, reaction times in conditions where both TD and BU representations were incongruent showed a reduced interference effect. This aligns with the hypothesis that conscious awareness enables overlapping neural networks to integrate these representations. The transition from independent to interactive effects across different visibility levels provides behavioral evidence supporting the global neuronal workspace theory, which predicts that conscious perception results from integrating sensory and decision-driven processing in a shared neural workspace.
The second experiment replicated these findings using a different paradigm, confirming that TD and BU processes interact only when the prime is consciously perceived. When participants reported no percept of the prime, TD and BU effects remained separate, with no significant interaction. However, when participants subjectively reported seeing the prime—even with unchanged physical characteristics—TD and BU effects combined non-independently, mirroring the pattern in the first experiment.
Results also indicated that BU-driven processing did not significantly influence reaction times when the prime was entirely suppressed. This suggests unconscious perception may not always generate measurable behavioral effects in certain masking conditions.
By demonstrating that TD and BU interactions depend on conscious awareness across two different experimental approaches, this research provides evidence that merging these processes is a hallmark of conscious perception rather than a byproduct of stimulus strength or task design.
Of note is that while the behavioral approach provides an accessible alternative to neuroimaging, it does not directly measure neural mechanisms.
The research, “Interaction Between Top-Down Decision-Driven Congruency Effect and Bottom-Up Input-Driven Congruency Effect Is Correlated With Conscious Awareness,” was authored by Ze-Fan Zheng, Shu-Yue Huang, Shena Lu, and Yong-Chun Cai.
