Tuesday, July 08, 2014

Information integration without awareness.

I want to point to an excellent review by Christof Koch and colleagues. It contains some useful summary graphics.
•Empirical data suggest that consciousness is not necessary for integration.
•Unconscious integration is nevertheless limited.
•Consciousness enables integrations over extended spatiotemporal windows.
•Consciousness may be needed for novel and high-level semantic integrations.
Information integration and consciousness are closely related, if not interdependent. But, what exactly is the nature of their relation? Which forms of integration require consciousness? Here, we examine the recent experimental literature with respect to perceptual and cognitive integration of spatiotemporal, multisensory, semantic, and novel information. We suggest that, whereas some integrative processes can occur without awareness, their scope is limited to smaller integration windows, to simpler associations, or to ones that were previously acquired consciously. This challenges previous claims that consciousness of some content is necessary for its integration; yet it also suggests that consciousness holds an enabling role in establishing integrative mechanisms that can later operate unconsciously, and in allowing wider-range integration, over bigger semantic, spatiotemporal, and sensory integration windows.


  1. Anonymous9:07 AM

    One thing has been puzzling me ever since I read about relation between consciousness and information integration: how does information and its integration, and hence consciousness, come about from the molecular-level interactions of neurons.

    Attributing consciousness, information, integration of information and such to emergence is very common, but actually it's pretty much like saying "God did it". If any emergent phenomenon does indeed have any features that cannot be reduced to its parts and their interactions, it has an unsettling ex nihilo tone to it. And, so far, I don't know of any part or interactions that has already "consciousness" or "information" or "integration" in it.

    Let me elaborate. Say, we have balls whose conductivity depends on pressure. Put the balls in a bowl and fill the bowl with liquid similar to cerebrospinal fluid. Connect two wires in two balls on the opposite sides of the bowl, and turn electricity on. Shake the container, and let the balls squeeze against each other. Could it be somehow possible that consciousness or information arise out of their interactions? Could the system integrate any information, or could it become minimally conscious even for a fraction of a second just because of the physical interactions, nothing else?

    I don't think so. It would sound quite absurd to claim so.

    What is the fundamental difference between the aforementioned bowl and the brain -- apart from us knowing that brain and consciousness are related due to first-person subjective experience -- that can change the game and allow mere emergence to explain appearance of consciousness without it already being a property of the parts or interactions?

  2. I suggest you check out Metzinger's (The Ego Tunnel) book and other writings on the issues you raise. If you enter "metzinger" in the MindBlog search box, you will find a series of 2009 postings that abstract some of the book chapters.

  3. Anonymous10:16 AM

    Thanks for the tip. That book has been on my to-read list for a while now, I guess I'll prioritize it now.