The philosopher John Searle once remarked: “Studying the brain without studying consciousness would be like studying the stomach without studying digestion.” Ten years ago, few researchers would have taken him seriously, but times are changing. Although ‘Consciousness Studies’ remains a somewhat desultory field—as anyone knows who has attended the biennial Tucson conferences on the subject—a consensus is now starting to emerge, at least among neuroscientists, that the problem can be approached experimentally. This was apparent from a recent meeting at Cold Spring Harbor Laboratory, where some thirty representatives of what might now be considered the scientific ‘mainstream’ gathered to discuss recent progress in the field.

Most of these researchers have chosen to sidestep the ‘hard problem’ of explaining the subjective aspect of consciousness, and to focus instead on identifying neural processes that correlate with third-person reports of conscious experience, particularly sensory awareness. They hope that by combining psychophysics, neuroimaging and electrophysiology, it will eventually be possible to understand the computations that occur between sensory input and motor output, and to pinpoint the differences between cases where a stimulus is consciously perceived and those where it is not.

Much of this work is best done in humans, because it is difficult to train animals to report their experiences. It is not sufficient that a stimulus should influence an animal's behavior: there are many examples in the neurological literature of dissociation between awareness of a stimulus and the ability to localize it or make appropriate motor responses. The distinctive feature of consciousness is that it provides a flexible representation of the world that can be used to control diverse behaviors, including of course answering questions as to whether a stimulus was perceived.

It thus seems likely, as Francis Crick and Christof Koch have argued, that the neural correlates of sensory awareness are to be found in areas with connections to the prefrontal cortex (PFC), the region responsible for controlling complex behaviors. Lesions to the PFC, however, do not abolish consciousness, and indeed it is implausible that any single area is the locus of all conscious experience. Instead, most researchers believe that different conscious states depend on activity in different brain regions, perhaps including subcortical as well as cortical structures.

What types of activity? Two schools of thought were apparent at the Cold Spring Harbor meeting. One view, advocated by Koch, is that neurons exist whose specific function is to support conscious representations of the world. Although many other neurons are necessary to support these representations, their activity corresponds to ‘behind-the-scenes’ computations, of which we are not directly aware. If ‘consciousness neurons’ can be identified, it might be possible to manipulate them selectively in order to test causal theories of consciousness. Not everyone finds this hypothesis compelling, however, and an alternative view, summarized at the meeting by Giulio Tononi, is that conscious states arise from a ‘dynamic core’ consisting of correlated activities of large numbers of neurons distributed across many brain regions. Such neurons would be defined not by any fixed attribute, but instead by their transient participation in a complex pattern of interaction with other neurons.

Not enough data exist to resolve this debate, but both views find some support from experiments with binocular rivalry, a popular testbed for ideas about consciousness. When different stimuli are presented to the two eyes, rather than experiencing a mixed stimulus, subjects perceive the two stimuli in alternation, allowing different patterns of neural activity to be correlated with different conscious states, independent of any change in the stimulus. Experiments in both humans and monkeys indicate that correlates of awareness under such conditions are not confined to one area but are widely distributed, consistent with the dynamic core hypothesis, and that in some areas (for instance, IT cortex), the majority of neurons track the perceptual changes. It is also true, however, that some of the modulated neurons are driven by the nondominant stimulus, and so are presumably not contributing directly to awareness.

Binocular rivalry is of course an artificial situation in which conscious states change spontaneously. In most real-world situations, however, awareness is at least partly subject to voluntary control, through attentional shifts. The power of attention to determine what enters consciousness is often under-appreciated: a particularly dramatic recent demonstration is that when observers view a complex scene, they often fail to detect even very large changes in the unattended components of the image, if the changes are timed to coincide with a transient discontinuity such as a saccade or eye blink. Clearly, attention is not sufficient to produce awareness—visual attention, for example, leads to enhanced activity in visual cortex even when no stimulus is present—but it may be necessary. In any case, the two phenomena are clearly intertwined, and it seems likely that understanding the neural mechanism of attention, along with related phenomena such as masking and neglect, will yield important insights into the basis of consciousness.

There are of course limitations to the current approach. Human consciousness encompasses not only sensory awareness but also imagery, emotions, silent speech and much besides. Moreover, experiments on sensory awareness are concerned only with the content of consciousness, and not with what distinguishes conscious from unconscious states in general. It is too early to say whether studying perception will eventually lead to a unifying theory of consciousness. But considering that the topic was hardly on the scientific agenda a decade ago, it is impressive that so much agreement has been reached on how to proceed. As one participant quipped, the most remarkable thing about the Cold Spring Harbor meeting was that most of the speakers were not yet tenured. It is no longer considered professional suicide for young experimentalists to study consciousness, and for a field that is itself so young, that represents progress.