Abstract:The role of the prefrontal cortex (PFC) in consciousness is hotly debated. Frontal theories argue that the PFC is necessary for consciousness, while sensory theories propose that consciousness arises from recurrent activity in the posterior cortex alone, with activity in the PFC resulting from the mere act of reporting. To resolve this dispute, we re-analysed an EEG dataset of 30 participants from a no-report inattentional blindness paradigm where faces are (un)consciously perceived. Dynamic causal modelling was used to estimate the effective connectivity between the key contended brain regions, the prefrontal and the posterior cortices. Then, a second-level parametric empirical Bayesian model was conducted to determine how connectivity was modulated by awareness and task-relevance. While an initial data-driven search could not corroborate neither sensory nor frontal theories of consciousness, a more directed hypothesis-driven analysis revealed strong evidence that both theories could explain the data, with a very slight preference for frontal theories. Specifically, a model with backward connections switched off within the posterior cortex explained awareness better (53%) than a model without backward connections from the PFC to sensory regions. Our findings provide some support for a subtle, yet crucial, contribution of the frontal cortex in consciousness, and highlight the need to revise current theories of consciousness.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is about the role of the prefrontal cortex (PFC) in the generation of consciousness. Specifically, the research aims to evaluate two major consciousness theories through computational modeling: the sensory theory and the frontal lobe theory. 1. The **sensory theory** holds that consciousness is entirely generated by recurrent activities in the posterior cortex, and the activities of the prefrontal cortex only report the results of consciousness. 2. The **frontal lobe theory** believes that the prefrontal cortex is necessary for the generation of consciousness. To verify these two theories, the author re - analyzed an EEG data set, which was from an experiment of no - report inattentional blindness paradigm, in which 30 participants had conscious or unconscious perception of human faces. Through Dynamic Causal Modelling (DCM), the author estimated the effective connections between key brain regions (prefrontal cortex and posterior cortex) and used the Parametric Empirical Bayes (PEB) to determine how these connections were affected by consciousness and task - relevance. ### Summary of main research questions: - **Does the prefrontal cortex play an important role in the generation of consciousness?** - **Which of the sensory theory and the frontal lobe theory is more in line with the experimental data?** Through the above methods, the research found that although the initial data - driven search failed to clearly support either theory, the directed hypothesis - driven analysis showed that both theories could explain the data, and slightly favored the frontal lobe theory. In particular, the model with the backward connection within the posterior cortex turned off performed better (53%) in explaining consciousness than the model without the backward connection from the prefrontal cortex to the sensory area. Therefore, the research results provide partial support for the subtle but crucial role of the prefrontal cortex in the generation of consciousness and emphasize the need to revise the existing consciousness theories. ### Formula representation: The models and statistical methods involved in the paper mainly include DCM and PEB. The specific formulas are as follows: - **Connection strength in the DCM model**: \[ y(t)=A x(t)+B u(t) \] where \(y(t)\) is the observed neural activity, \(A\) is the intrinsic connection matrix, \(B\) is the input matrix, and \(u(t)\) is the external input (such as visual stimulation). - **Parameter estimation in the PEB model**: \[ p(\theta|y)\propto p(y|\theta)p(\theta) \] where \(p(\theta|y)\) is the posterior probability of the parameter \(\theta\), \(p(y|\theta)\) is the likelihood function, and \(p(\theta)\) is the prior probability. These formulas help researchers quantify the connection strength between different brain regions and their changes under different conditions.

Computational modelling shows evidence in support of both sensory and frontal theories of consciousness