A recent paper examined a patient with deaf-hearing, analogous to blind-sight, where there can be detection of a signal without conscious awareness of it. (citation below) For example, a person with blind-sight may avoid an obstacle without awareness of it; and, a deaf-hearing person may be startled and orient towards a noise without consciously hearing it.
Deaf-hearing seems to be the more rare condition and so this stroke victim was examined extensively. The path from within the ear through the brain stem to the thalamus was normal. But bilateral damage to the auditory areas of the cortex seemed to completely disrupt the processing of the signal. There appeared to be a problem with the communication between the thalamus and the cortex as well as damage to the cortex.
However, despite this break in the usual path, a signal designated P3 did occur within the cortex. The discussion has this passage:
Bernat et al. (2001) offer evidence that subliminal stimuli can evoke consistent P3 waves. They speculated that P3 could represent a link between unconscious and conscious awareness in the context updating processes. In our patient the generation of P3-like potentials implied that deviant stimuli were selectively processed bypassing networks involved in conscious perception. Schonwiesner et al. (2007) and Pandya (1995) hypothesized that association areas in and adjacent to the auditory parabelt might form an independent circuit from thalamo-cortical projections in the auditory system. These alternate pathways could be preserved in our patient and responsible for the generation of P3-like potentials.
What does a P3 wave indicate?
P3 offers a covert and indirect measure of attentional resource allocation that represents an index of change detection. P3 is related to the activity of associative cortical areas and is sensitive to complex processes around recalled information, stimulus significance, recognized auditory information and memory context updating. The sources of P3 are believed to be located in heteromodal areas of the fronto-parietal cortex and their activation might reflect an attention switch to an environmental change …. some authors have demonstrated an asymmetrical cortical activation of P3 by using unilateral auditory stimulation. Among others, Gilmore and colleagues (2009) argued that in normal condition the right hemisphere is more prominently engaged during working memory and updating processes underlying P3 …the patient showed robust P3-like components over the left posterior areas and a significantly lower distribution of the potentials over the right fronto-temporal and central areas in response to right ear stimulation. The left ear stimulation could not evoke any detectable responses.
In other words, P3 is about an event rather then a sound, switching attention to unusual events. In this patient there was a left hemisphere P3 but not the right hemisphere one and it appears to be the right hemisphere P3 that engages consciousness.
The authors use these results to give input to the choice between two hypotheses. A simple hierarchical model where each level of processing is necessary for the next is not consistent with the findings. But the reverse hierarchy theory, which asserts that neural circuits mediating a certain percept can be modified starting at the highest representational level and progressing to lower-levels in search of more refined high resolution information to optimize perception, would not be ruled out by the paradoxical findings in this patient.
Personally, although not explicit in the paper, I find the results are more evidence for: one, the necessity of thalamus-cortex communication to consciousness; two, attention being a much more complex entity when just the focus of consciousness; three, that there is a difference in how the left and right hemispheres handle sound; and four, the importance of top-down inputs (expectations) in forming perceptions.
Cavinato, M., Rigon, J., Volpato, C., Semenza, C., & Piccione, F. (2012). Preservation of Auditory P300-Like Potentials in Cortical Deafness PLoS ONE, 7 (1) DOI: 10.1371/journal.pone.0029909