EM Information Theory of Consciousness | Hacker News Reader

EM Information Theory of Consciousness

academic.oup.com

30 points8bitsrule13 comments

I find this deeply confusing.

- The distinction between integrating information in time vs in space is not clear to me. The author talks about how EM fields "instantly computes" solutions to physical equilibrium problems (how will iron filings be arranged, etc), but also how "dynamic EM field information" can do computations like AND. Doesn't that mean we're talking about computation/information processing that unfolds in both time and space (like everything else?)

- But more fundamentally, I don't see why consciousness is better explained by neurons influencing each other through their EM fields rather than just the normal, slower, ion gate / axon / synapse / neurotransmitter / etc way we already know they interact?

- And I don't get the local vs global distinction the author tries to draw:

> I estimated that the EM field electrical perturbation from the firing of a single neuron extends into a volume of only about 80 µm encompassing a maximum of about 200 neurons

> within the brain, the light cones of all the trillions of emitted photon will almost entirely overlap. Any charged particle in the brain (or outside, but with rapidly diminishing probability) can potentially be the receiver of any of the trillions of emitted 10 bit packages of information available in the entire field

... so it sounds like the author thinks both locality and global reach are important?

But I think mostly, I don't get why the distance-weighted-sum way that EM field information gets integrated together would give you anything better than a noisy jumble. The complexity of our behavior and experience seems much more likely to be explained in the connection and interactions among neurons, where the 'summation' is both complex and informed by the complicated way neurons have wired together. But I don't see how any of this begins to approach the subjective side of things, even though he name-drops Nagel.

I read about half the paper but couldn't find an answer to my first objection (perhaps I missed it), but if this was the case, wouldn't we observe sudden changes in cognition in those who were subject to the strong magnetic fields of an MRI?

You might find this an interesting read: https://en.wikipedia.org/wiki/God_helmet

> wouldn't we observe sudden changes in cognition in those who were subject to the strong magnetic fields of an MRI?

It was mentioned, the theory is only EM waves of the right shape will couple with the brains neural networks.

Other studies have proven it is possible to manipulate mind state from external electromagnetic fields

Maxwell's equations are linear, so probably not complicated enough to support the idea that they can generate human-like actions before they interact with matter.

They're linear coupled PDE s are they not? which allow for a significantly wider range of behaviours that linear ODEs?

As far as I understand, there is no 'closed form solution' for the initial value problem of Maxwell's equations, despite them being linear?

So I'm not sure it's correct to say that "linearity ~ simplicity" in PDE theory. My understanding is the correct analog of "linear" in PDEs is "elliptic" [which is more akin to 'operator with full spectrum'], which Maxwell's equations are not?

With infinite possibilities -attempts possible, with enough time, isn't it probable to happen? Why couldn't intelligence exist energetically before it manifests via a physical path?

That is only true in vacuum you get non-linear effects from (among other things) the medium and cell-membranes.

Isn't the Schrodinger equation also linear in a certain sense? In that, the derivative of the wave function with respect to time is the result of applying a linear operator to the wave function.

my point being that, if the Schrodinger equation is a correct description of physics, and what is needed is just the right hilbert space and the right hamiltonian, then all of physics would be linear in a certain sense, and so the fact that a particular thing is linear, wouldn't prevent that thing from doing what humans do (unless humans are an exception to ordinary physics, which, maybe, but if so I don't expect that we'll see proof of it any time soon).

But how one defines consciousness (without relying on synonyms - "sentience", "awareness", etc.).

Read several sentences. Seems like dualistic pseudoscience to me. Maybe I missed something.

I don't believe IIT is the right approach to consciousness, I rather prefer the RL approach. IIT feels like trying to fit the problem into a ready made concept, and doesn't start from where it should - why should there be consciousness in the first place? for survival (my answer). RL with its agent centric setting is much more suited to answer such questions. Agents with goals, embedded into the environment are a better starting point than integrated information that exists who knows why.

Reply on news.ycombinator.com