Some time ago there was an NS article about vision, concluding with the idea that everything we see is an internal construct that we then externalise and map onto the real world. Maybe consciousness is similar - a construct - and so Tallis is right that you will never see it in the neurons because the activity is all about maintaining the construct rather then consciousness directly manifesting in neural activity.
I don't like using techno analogies, a big flaw in that way of thinking being we map our current beliefs onto reality, hence Newtonian clockwork universes and computer technologies 'Reality is a Simulation', and so on. However, I think a reasonable comparison might be that consciousness is a virtual instance running on hardware, and so therefore it would be interesting to compare the activity of a brain when there is no consciousness to when there is, and see what the differences are. No doubt this has already been done many times!
This also made me wonder, bearing in mind the contingent nature of memory, if memory is more to do with remembering the pattern of a construct (i.e. the memory), rather than explicitly remembering the memory itself. Would it be easier to remember the instructions to order the words in a book than the book itself? Plans and blueprints take up less room and resource than the objects they describe.