PerspectiveNeuroscience

The controversial correlates of consciousness

See allHide authors and affiliations

Science  04 May 2018:
Vol. 360, Issue 6388, pp. 493-494
DOI: 10.1126/science.aat5616

The mechanism of consciousness is one of the most fundamental, exciting, and challenging pursuits in 21st-century science. Although the field of consciousness studies attracts a diverse array of thinkers who posit myriad physical or metaphysical substrates for experience, consciousness must have a neural basis. But where in the brain is conscious experience generated? It would seem that, given this remarkable era of technical and experimental prowess in the neurosciences, we would be homing in on the specific circuits or precise neuronal subpopulations that generate experience. To the contrary, there is still active debate as to whether the neural correlates of consciousness are, in coarse terms, located in the back or the front of the brain (1, 2). On page 537 of this issue, van Vugt et al. (3) provide evidence that the prefrontal cortex is one of the brain regions that mediates visual consciousness. Additionally, Joglekar et al. (4) provide evidence that the prefrontal cortex is important for igniting neural networks that contribute to visual signal processing. Both studies support a model for consciousness that involves distributed and reciprocal interactions across the cortex.

The investigation of nonhuman primates by van Vugt et al. was motivated by a joint consideration of signal detection theory and global neuronal workspace theory. Signal detection theory attempts to explain the processing of stimuli that are around the threshold of perception and the reasons why sometimes we perceive such stimuli and other times we do not. Global workspace was originally a psychological framework for consciousness that has, in the past decade, become more neurobiologically informed (5). Global neuronal workspace theory posits that a subset of excitatory neurons and long-range tracts in the cortex, including prominent involvement of the prefrontal cortex, amplify, sustain, and broadcast specific representations for widespread cognitive processing.

Van Vugt et al. hypothesized that the threshold required for signal detection and the neural activity required for the broadcasting of information through the neuronal workspace would be the same. Neuronal activity in the visual cortex (areas V1 and V4, in the back of the brain; see the figure) and dorsolateral prefrontal cortex (in the front of the brain) of awake monkeys was recorded to establish neural correlates of visual stimuli that were perceived and reported with a specific eye movement. Reported stimuli were associated with strong and sustained prefrontal cortex activity, whereas nonreported stimuli were associated with weak and transient prefrontal activity. The investigators also assessed where information was lost for nonreportable stimuli and found that propagation failures could occur at various stages in the feedforward pathways en route to the front of the brain. It was concluded that stimuli cross the threshold for reportable signal detection when a critical trigger (or “ignition”) for broadcasting occurs in the prefrontal cortex. This empirical conclusion was supported by a model in which reciprocal activity between frontal and parietal cortices enables the signal to become a self-sustained representation.

But how do the long-range excitatory neurons of the global neuronal workspace amplify sensory signals without corrupting them or leading the brain into unbridled activation? Joglekar et al. examined a model of feedforward (back-to-front) and feedback (front-to-back) processing in nonhuman primate brain networks. They demonstrate that excitatory feedback connections can amplify signals but are balanced by local inhibitory processes. This “global balanced amplification” fits current empirical data and is novel because it integrates local, feedforward, and feedback processing as well as the inhibitory neuronal brakes that appropriately constrain the network. Furthermore, Joglekar et al. modeled sensory input to area V1 in the visual cortex and found that weak signals activate local cortex whereas stronger signals lead to activation of prefrontal cortex and a reverberation of cortical networks, all of which are consistent with global neuronal workspace theory. These modeling data support the hypothesis of van Vugt et al. that the threshold of signal detection is the same threshold that activates prefrontal cortex and ignites a reciprocally connected cortical network.

Processing consciousness

Visual processing (left) becomes consciously accessible (right) after “ignition” in the frontal cortex leads to reciprocal interactions that allow the representation of a stimulus to become self-sustaining and widely broadcast.

GRAPHIC: A. KITTERMAN/SCIENCE

How do these data inform the current controversy regarding the location of the neural correlates of consciousness? To answer this question, it is necessary to consider the distinction between phenomenal consciousness and access consciousness (6). Phenomenal consciousness is the purely qualitative aspect of experience, whereas access consciousness shares that experience with other cognitive systems for further action (for example, memory, motor activity, or report). The empirical data of van Vugt et al. were focused explicitly on reportable conscious events and provide further support to the hypothesis that the prefrontal cortex is important for access consciousness. The study does not speak directly to the question of phenomenal consciousness, but the authors do allude to evidence that prefrontal cortical neurons can represent consciously perceived visual stimuli even in the absence of report (7). Furthermore, a recent study on dreaming found that, although the neural correlates of phenomenal consciousness appeared to be primarily in the posterior cortex, dreaming during rapid eye movement sleep was also associated with high-frequency activity in the frontal and prefrontal cortices (8). In other words, anterior cortex might also contribute to the pure experience that constitutes phenomenal consciousness.

The studies of van Vugt et al. and Joglekar et al. also help to explain a related and long-standing scientific issue, namely, the mechanisms of general anesthetics. Models in which a reciprocally interacting frontal and parietal cortex is deemed necessary for access consciousness align with the finding that diverse general anesthetics depress metabolism and/or disrupt connectivity in frontal-parietal networks (9). Indeed, there is evidence that diverse general anesthetics preferentially suppress feedback connectivity from the frontal cortex in humans (10), potentially representing a failure of ignition. General anesthetics also show dose-dependent effects on local and long-range processing, two critical elements of global balanced amplification. At lower doses, local network connectivity is enhanced while the long-range connections of the global neuronal workspace are disrupted; higher anesthetic doses suppress both local and long-range connectivity (11). Thus, general anesthetics could prevent ignition in the prefrontal cortex while mitigating signal strength, depending on the dose. This independent line of investigation on the neural correlates of unconsciousness provides further evidence supporting the conclusions of the two studies.

Many questions remain. How do the studies of van Vugt et al. and Joglekar et al. apply to sensory processing outside of the visual system or to endogenous experiences (such as dreams) that do not require external sensory input at all? How does this work in nonhuman primates translate to humans or to animals that have a less developed prefrontal cortex? Further, how do these presumed correlates of consciousness inform the actual causes of consciousness? Solving—or dissolving—controversies related to the neural correlates will require transparent definitions of consciousness (access, phenomenal) and carefully designed paradigms (report, no report) as well as techniques that causally manipulate neural circuits with the dependent variable being a principled, possibly behavior-independent measure of consciousness. Only rigorous neuroscientific investigation will be able to reveal if the brain can ultimately explain both itself and its most precious function.

References

View Abstract

Stay Connected to Science

Navigate This Article