I am trying to understand what consciousness is, based on my basic knowledge of our senses, artificial intelligence (computer vision, specifically) and some philosophy. Here's my reasoning:

As far as I understand it from my Neural Network class, each eye captures the light and transmits it through electric signals to the brain. Each action potential can be seen as a binary 1, no activity as a 0.

This stimulates some of the neurons in the visual cortex, which allows our brain to interpret and understand what the eyes have captured.

In computer vision, this is where "sight" ends: the computer detects what's in the image, based on its knowledge base - its training set - and that's it.

So why does the human brain take the extra steps of combining both signals into an image - i.e translating the different waves of the electromagnetic spectrum into shapes, colors, etc -, turning it upright and displaying it to ourselves?

The same could be said, as far as I know it, about the other senses: hearing would be detecting vibrations in the air, but the brain also translate these into sounds.

Does this hint that Descartes' dualistic definition could be true, that our consciousness is separate from the brain? Or could consciousness be a sub product of the brain, a consolidation of what each one of its areas are processing, in order to better understand the world (as in Stanislas Dehaene's workspace neurons cognitive model)?

edit: I've been thinking a bit more about the subject. If there are workspace neurons responsible for consciousness, then there should be a region of the brain that is constantly active, at all times.

After a bit of - online - research, I've stumbled upon the Default mode network (DMN), which is anti-correlated to the Task-positive network (TPN). Interestingly, the DMN is active when the individual is not focused on an activity - at rest. Furthermore, there is a possible increase in connectivity between DMN's nodes during REM sleep - when we dream. On the other hand, TPM is most active during attention-demanding tasks. It has therefore been theorized that both are elements of a single default network with anti-correlated components (in an scientific paper behind a pay-wall, so I'm quoting Wikipedia).

Plus, both networks involve the fronto-parietal attention network, where it is believed the workspace neurons are located.

I couldn't help correlating the DMN and TPN to Carl Jung's theories of the Conscious and the Unconscious - which I am particularly fond about. If these wild conjectures I am forming are correct, then we could have the following (oversimplified) neural correlate of consciousness:


In this idea, exchanges between TPM and DMN occur through neurotransmitters, while exchanges between the conscious and unconscious happen using psychic energy (as in the energy from the psyche, not supernatural powers). Furthermore, the collective unconscious would then be inherited from human to human - embedded in our DNA. If true, then there would be discernible patterns in a newborn's brain...

So, going back to A.I, a conscious agent would not, theoretically, be a single agent, but instead a multi-agent collaborative system...

Now, I have very basic knowledge and lack scientific research on these topics, but I am interested in learning a LOT more. Is this theory absurd and blatantly wrong, or is it worth pursuing as a hypothesis?

  • 1
    $\begingroup$ Thanks for re-posting the question here, it makes more sense +1. I think you are considering the difference between objective analysis (peripheral sensory organs) versus perception and sensation (central mechanisms). Now it is getting more clear what you're after, I think this previous post contains the answer you are after - cogsci.stackexchange.com/questions/987/… $\endgroup$
    – AliceD
    May 14 '15 at 4:32
  • $\begingroup$ @AliceD yes, I think that's part of the answer. I'm wondering, specifically, how do each of these theories explain the brain creating an image for us to see. Partly to satisfy my curiosity, partly because I think this is crucial to strong AI. If consciousness is not a sub product of the brain, how do we simulate it on a computer? What is consciousness role in human thought, what's its relationship to the neural network? Could it be the Oracle black box Turing was referring to? $\endgroup$
    – Tarek
    May 14 '15 at 10:17
  • 1
    $\begingroup$ Your question is known as the hard problem of consciousness. The best answer right now is: we have no idea. $\endgroup$ May 14 '15 at 12:59
  • $\begingroup$ @Josh do we have an understanding of how these images are produced by the brain? How does it translate action potential into a visual image? Maybe we could start from that. $\endgroup$
    – Tarek
    May 14 '15 at 13:20
  • 1
    $\begingroup$ No, not really. There are some theories (which is what the question AliceD linked to is about), but there's no clear picture of why or how qualia exist. $\endgroup$ May 14 '15 at 13:23

Your question is about the hard problem of consciousness, which is basically the question of how qualia can be explained in a mechanistic way. As alluded to by the name of the problem, it's hard to give a satisfactory answer. The answer right now is: we don't know.

There are some theories about how qualia and consciousness could have a neural basis (see this question and this wikipedia article). But none of these theories are widely accepted by cognitive scientists and there is still considerable debate over the issue.

Your question also illustrates another philosophical issue in the philosophy of mind. You asked:

So why does the human brain take the extra steps of combining both signals into an image - i.e translating the different waves of the electromagnetic spectrum into shapes, colors, etc -, turning it upright and displaying it to ourselves

(emphasis mine)

The idea that conscioussness and qualia reconstruct the world and then we act and perceive on that world creates a problem: how do we act an perceive on that world? Do we reconstruct it? If so, then we are back to the original problem. This is called the homunculus argument, and it illustrates the fact that we haven't explained anything if we posit that the purpose of qualia is to reconstruct what's already out there.

  • 1
    $\begingroup$ cogsci.stackexchange.com/q/9839/2868 might also be interesting. $\endgroup$ May 14 '15 at 14:12
  • $\begingroup$ About that last part, i think that the qualia is not reconstructing reality (was that what you meant with the word reconstructing?), given that it's a mental image (there are no colours). Given that the brain does not need a visual representation to to act upon the world, I would theorize - without any scientific basis, a.k.a out of my ass - that qualia's purpose is to link the brain and consciousness. $\endgroup$
    – Tarek
    May 14 '15 at 14:19
  • $\begingroup$ @ChristianHummeluhr interesting indeed, thanks! Something to ponder: Intelligent Agents in AI use all these principles, yet are they conscious? $\endgroup$
    – Tarek
    May 14 '15 at 14:30

Not the answer you're looking for? Browse other questions tagged or ask your own question.