A recent paper by C. Sergent and others has been commented on by R. Kentridge (citations below). They showed that attention to the visual space where a stimulus was, but is now gone, can bring that stimulus into consciousness. This retroperception effect can occur as late as 400 ms after stimulus presentation ends.
Here is the abstract of the Sergent paper:
Is our perceptual experience of a stimulus entirely determined during the early buildup of the sensory representation, within 100 to 150 ms following stimulation? Or can later influences, such as sensory reactivation, still determine whether we become conscious of a stimulus? Late visual reactivation can be experimentally induced by postcueing attention after visual stimulus offset. In a contrary approach from previous work on postcued attention and visual short-term memory, which used multiple item displays, we tested the influence of postcued attention on perception, using a single visual stimulus (Gabor patch) at threshold contrast. We showed that attracting attention to the stimulus location 100 to 400 ms after presentation still drastically improved the viewers objective capacity to detect its presence and to discriminate its orientation, along with drastic increase in subjective visibility. This retroperception effect demonstrates that postcued attention can retrospectively trigger the conscious perception of a stimulus that would otherwise have escaped consciousness. It was known that poststimulus events could either suppress consciousness, as in masking, or alter conscious content, as in the flash-lag illusion. Our results show that conscious perception can also be triggered by an external event several hundred ms after stimulus offset, underlining unsuspected temporal flexibility in conscious perception.
The Kentridge commentary concludes:
Although the new study does not directly address underlying mechanisms, the effect must depend on attention acting on some neural trace that persists after the offset of the target. We know that attention modifies the neural response elicited by targets so it is, perhaps, unsurprising that attention can affect neural responses that continue after target offset. Neural activity elicited by transient visual stimuli persists for long periods. What is surprising is that retro-active attention brings otherwise unseen stimulus into consciousness.
Attention plays a role in many theories of consciousness. Both Lamme and Dehaene et al. propose that attention can amplify the neural trace of a stimulus so that it has long-lasting effects spreading from sensory areas of cortex to frontal regions. They accommodate findings that attention can act on stimuli that do not elicit consciousness by suggesting that attention only promotes stimuli to conscious report whose sensory neuronal representation persists through feedback of signals between areas. When attention produces a behavioural effect in the absence of consciousness the strength of neural response is enhanced but no recurrent feedback takes place. … The neural traces that attention acts on in Sergent et al.s experiments persist for so long that they are likely to depend on feedback of neural signals, so it appears that without attention these recurrent signals do not elicit consciousness, as Dehaene et al. suggest.
The philosopher Ned Block, however, distinguishes between two forms of consciousness: phenomenal consciousness, which corresponds to the experience elicited by a stimulus, and access consciousness, in which the properties of the stimulus become available to cognitive processes. He explains that the strong but still losing coalitions in the back of the head are the neural basis of phenomenal states (so long as they involve recurrent activity). The contrary position, for example, is that experience of the unreported items is incomplete and so there is no dissociation between experience and cognitive access. For this to occur we need to have, in Blocks own words, unconscious representations that are specific enough to do the task with the observed accuracy. the cue is supposed to promote attentional amplification of the cued unconscious specific representation, which, when combined with the conscious generic representation, results in a conscious specific representation of the cued item. That is, of course, exactly what Sergent et al. have found (except that their subjects do not even appear to report a generic representation of the unseen stimulus).
Sergent et al.s result does not necessarily invalidate the distinction between access and phenomenal consciousness, but it does lend weight to the alternative, and perhaps simpler, position that consciousness is just consciousness.
I suspect that the thalamocortical traffic has a lot to do with sustaining perception of unattended stimuli for some time and the directing of attention to them after the original stimuli have disappeared. There is more going on than feedforward and feedback confined to the neocortex.
Sergent, C., Wyart, V., Babo-Rebelo, M., Cohen, L., Naccache, L., & Tallon-Baudry, C. (2013). Cueing Attention after the Stimulus Is Gone Can Retrospectively Trigger Conscious Perception Current Biology, 23 (2), 150-155 DOI: 10.1016/j.cub.2012.11.047
Kentridge, R. (2013). Visual Attention: Bringing the Unseen Past into View Current Biology, 23 (2) DOI: 10.1016/j.cub.2012.11.056