Evidence for Globalized Consciousness

[zoobyshoe]

You know from the literature that people have to learn to see them, to attend to them, to be able to predict their presence and so find them.

I don't know what you mean by this.

For example, a couple years ago Evo posted reporting the startling experience of a visual migraine aura. She had never experienced this before, or heard of it, and had no idea what was happening.

When and where did she "learn" to predict the presence of it and look for it?

 

[zoobyshoe]

But your "visual neuron" will be getting more top-down inputs from higher levels than it gets coming up as its "data stream".

Under normal circumstances. In the case of seizure and migraine or direct electrical or magnetic stimulation, I have to suppose all normal inputs from above or below are rendered moot, the neuron isn't firing in response to them. According to what I've read, during seizures the neurons are setting each other off by induction, not through synaptic connections.

 

[apeiron]

I don't know what you mean by this.

For example, a couple years ago Evo posted reporting the startling experience of a visual migraine aura. She had never experienced this before, or heard of it, and had no idea what was happening.

When and where did she "learn" to predict the presence of it and look for it?

And I just happened to have my first ever visual migraine just a few weeks back. It took me a little while to figure out it was there. I was thinking one eye seemed a little tired. The TV picture was not quite right. Then I eventually realized I was "seeing" a wriggling curve of lights - which for about half an hour I had just been trying to look around and ignore.

As soon as I thought, "that could be a visual migraine", that is when I could really focus on it and experience it as an object in itself. I could relate it to pictures that people had tried to draw of them.

So even with such a startling kind of "bottom-up, data driven" phenomenon, I found that top-down attention and expectation had to come into play for the sensation to be consciously reportable.

 

[apeiron]

Under normal circumstances. In the case of seizure and migraine or direct electrical or magnetic stimulation, I have to suppose all normal inputs from above or below are rendered moot, the neuron isn't firing in response to them. According to what I've read, during seizures the neurons are setting each other off by induction, not through synaptic connections.

If a seizure does indeed render the normal organised connectedness of the brain moot, then the result is...unconsciousness.

 

[atyy]

fMRI studies of these tasks show that perceptual switches are instigated by activity in the prefrontal cortex, indicating that a top down influence does cause the reordering of activity in the visual cortex as expected from predictive coding models. It has been mostly disproven that perceptual bistability arises from bottom-up rivalry in the early stages of visual processing alone.

The fact that there is a percept which varies independently of the stimulus to me implies that there must be an internal model. The only question is whether is constructed in sequential processing stages from the sensory input or whether some form of predictive coding is involved.

In the paper of the OP, isn't there a unique percept to each stimulus, although maybe not a unique stimulus to each percept? So it seems to be simply a many-to-one mapping. Why is an internal model required?

Even in the case where there are "top-down" influences, couldn't that just be apparent stochasticity due to long-transients from previous "bottom up" stimuli? Again, no "internal model" seems to be needed.

Example of long transients
http://arxiv.org/abs/cond-mat/0603154
http://arxiv.org/abs/0705.3214

 

[zoobyshoe]

If a seizure does indeed render the normal organised connectedness of the brain moot, then the result is...unconsciousness.

Obviously not. In a simple partial seizure we know the neurons are not communicating in the normal way, and yet consciousness is preserved.

Somatosensory-primary sensory cortex seizures usually elicit positive or negative sensations contralateral to the discharge.[3] Symptoms associated with seizures from the postcentral gyrus include the following:

Tingling
Numbness
Pain
Heat
Cold
Agnosia
Phantom sensations
Sensations of movement
Abdominal pain usually originates from the temporal lobe, and genital pain from the mesial parietal sensory cortex. The posterior parietal cortex is the likely origin of limb agnosia.

Supplemental sensory-secondary sensory cortex seizures may have ipsilateral or bilateral positive or negative sensations or vague axial or diffuse sensations.

Visual-calcarine cortex discharges produce elemental hallucinations including scintillations, scotomata, colored lights, visual field deficits, or field inversion. The visual association cortex is the probable location of origin of complex visual hallucinations and photopsias.

Auditory SPS from the auditory cortex typically are perceived as simple sounds, rather than words or music. Olfactory-uncinate seizures originate from the orbitofrontal cortex and the mesial temporal area. Perceived odors are usually unpleasant, often with a burning quality.

Gustatory seizures usually are associated with temporal lobe origin, although the insula and parietal operculum also have been implicated. Perceived tastes are typically unpleasant, often with a metallic component.

Vestibular seizures originate from various areas, including frontal and temporal-parietal-occipital junction. Symptoms include vertigo, a tilting sensation, and vague dizziness.

Psychic SPS arise predominantly from the temporal and limbic region, including the amygdala, hippocampus, and parahippocampal gyrus. Perceptual hallucinations or illusions are usually complex, visual or auditory, and are rarely bimodal.

Déjà vu and jamais vu phenomena may occur. Fear is usual, but SPS can elicit happiness, sexual arousal, anger, and similar responses. Cognitive responses include feelings of depersonalization, unreality, forced thinking, or feelings that may defy description.

None of these sensations is produced in the ordinary way. They result from the populations of neurons in the given particular area firing wildly in the absence of a legitimate stimulus. The sensations produced are illusions, hallucinations as it were.

http://emedicine.medscape.com/article/1184384-clinical

 

[apeiron]

Obviously not. In a simple partial seizure we know the neurons are not communicating in the normal way, and yet consciousness is preserved.

In what way are the neurons communicating if not the usual way? You mentioned something about "EM" as if you think they somehow bypass action potentials and synapses completely.

See - http://www.wisegeek.com/what-is-the-pathophysiology-of-seizures.htm

There is certainly a failure of normal activity. But it merely confirms the point that neurons exist within a network of excitatory and inhibitory influences.

Disrupt that feedback-based balancing act and a part of the brain can generate vivid imagery - suppressed states can become expressed states. And the rest of the brain can try to integrate this chaotic activity into the general web of its activity.

Consciousness is of course not preserve if the disruption becomes more generalised or hits areas key to organising the top-down orchestration of what should be coming bottom up.

Remember that pixels are different because they will glow the same regardless of what happens around them, but will glow only if they have a driving input.

Neurons do it the other way round. They are always spiking regardless of whether they have input (top-down or bottom-up). But they also always reflect what is happening around them in terms of firing rates, synchrony and as Pythagorean mentions, a whole host of sub-level processes as well.

 

[zoobyshoe]

You mentioned something about "EM" as if you think they somehow bypass action potentials and synapses completely.

It seems they do in some cases:

Increased synchrony between neurons due to ephaptic interactions
Electrical fields created by synchronous activation of pyramidal neurons in laminar structures, such as the hippocampus, may increase further the excitability of neighboring neurons by nonsynaptic (ie, ephaptic) interactions. Other possible nonsynaptic interactions include electrotonic interactions due to gap junctions or changes in extracellular ionic concentrations of potassium and calcium. Increased coupling of neurons due to permanent increases in the functional availability of gap junctions might be a mechanism that predisposes to seizures or status epilepticus.

What I meant by the usual connections being moot, as well, is that a seizing neuron obviously isn't going to be at liberty to respond to the usual sort of input. It's tied up seizing.

The thing to attack about my notion is not the claim the neuron isn't responding to input, but the unspoken assumption it's not generating output. It is, of course. The thalamus will get entrained into any cortical seizure. The thalamus could then communicate the experience to the frontal lobes where it becomes a conscious experience, if you're into the frontal lobes as the location of consciousness.

 

[Evo]

And I just happened to have my first ever visual migraine just a few weeks back. It took me a little while to figure out it was there. I was thinking one eye seemed a little tired. The TV picture was not quite right. Then I eventually realized I was "seeing" a wriggling curve of lights - which for about half an hour I had just been trying to look around and ignore.

As soon as I thought, "that could be a visual migraine", that is when I could really focus on it and experience it as an object in itself. I could relate it to pictures that people had tried to draw of them.

So even with such a startling kind of "bottom-up, data driven" phenomenon, I found that top-down attention and expectation had to come into play for the sensation to be consciously reportable.

Classic ocular migraines are unmistakeable and shock the heck out of you when you have one. It isn't something that is vague or hard to notice. They continue to rotate and grow until they disappear, usually in 15-20 minutes.

Picture below is most like the comon crescent with geometric patterns inside, it is very bright and pulsing and the designs inside undulate. Not something you can ignore or is hard to see.

Sounds like you might have had a minor visual disturbance.

 

[apeiron]

The thing to attack about my notion is not the claim the neuron isn't responding to input, but the unspoken assumption it's not generating output.

Who was saying that?

What was being stressed was the need to move away from a computational view of neuronal function - one where input produces output in some simplistic mechanical fashion.

Instead, as Madness says, you have a neuronal network that is predicting its inputs and adapting its state in the light of prediction errors, pattern matches, or other sources of surprisal. So any individual neuron is already in a state of meaningful output - it is already contributing fully to the global state of the brain - before it responds to any input.

So the spike train of a neuron is saying something before it changed its rate or timing, just as much as it is after the change. You could say that the neuron might go from saying something insignificant and ignored to something suddenly distinctive and attention-worthy.
But that is confirming the importance of those contextual judgements - it is the larger brain that is saying there is a difference between the two. The neuron itself is just switching rates. What does it know about anything?

 

[madness]

In the paper of the OP, isn't there a unique percept to each stimulus, although maybe not a unique stimulus to each percept? So it seems to be simply a many-to-one mapping. Why is an internal model required?

Even in the case where there are "top-down" influences, couldn't that just be apparent stochasticity due to long-transients from previous "bottom up" stimuli? Again, no "internal model" seems to be needed.

Example of long transients
http://arxiv.org/abs/cond-mat/0603154
http://arxiv.org/abs/0705.3214

I don't have access to the article from home, but as far as I understand from the abstract it is an example of binocular rivalry. There is only one binocular stimulus, but it is rivalrous between the two eyes and the brain suppresses the input from one eye.

To understand the sense in which the brain generates an internal model, read these: http://philosophyandpsychology.com/?p=1013
http://en.wikipedia.org/wiki/Bayesian_brain
http://mrl.isr.uc.pt/pub/bscw.cgi/d27540/ReviewKnillPouget2.pdf

The final link is a peer reviewed article by one of the top guys in this subfield.

 

[Evo]

Who was saying that?

What was being stressed was the need to move away from a computational view of neuronal function - one where input produces output in some simplistic mechanical fashion.

Instead, as Madness says, you have a neuronal network that is predicting its inputs and adapting its state in the light of prediction errors, pattern matches, or other sources of surprisal. So any individual neuron is already in a state of meaningful output - it is already contributing fully to the global state of the brain - before it responds to any input.

So the spike train of a neuron is saying something before it changed its rate or timing, just as much as it is after the change. You could say that the neuron might go from saying something insignificant and ignored to something suddenly distinctive and attention-worthy.
But that is confirming the importance of those contextual judgements - it is the larger brain that is saying there is a difference between the two. The neuron itself is just switching rates. What does it know about anything?

Hey Apeiron, please link to the peer reviewed studies for this. Thanks.

 

[apeiron]

Hey Apeiron, please link to the peer reviewed studies for this. Thanks.

https://www.physicsforums.com/showpost.php?p=3964058&postcount=19

(or actual paper - http://www.nature.com/nature/journal/v398/n6725/abs/398334a0.html)

https://www.physicsforums.com/showpost.php?p=3965700&postcount=41

 

[atyy]

I don't have access to the article from home, but as far as I understand from the abstract it is an example of binocular rivalry. There is only one binocular stimulus, but it is rivalrous between the two eyes and the brain suppresses the input from one eye.

To understand the sense in which the brain generates an internal model, read these: http://philosophyandpsychology.com/?p=1013
http://en.wikipedia.org/wiki/Bayesian_brain
http://mrl.isr.uc.pt/pub/bscw.cgi/d27540/ReviewKnillPouget2.pdf

The final link is a peer reviewed article by one of the top guys in this subfield.

I see, so by internal model you meant something different from a model of the self. It just means that there is "spontaneous" activity in the brain.

 

[madness]

I see, so by internal model you meant something different from a model of the self. It just means that there is "spontaneous" activity in the brain.

In the context of perception it generally means a model of the external world. For self-awarenses and introspection there should also be a model of the self. I don't know why you think the activity would be spontaneous. Under bayesian theories of perception the activity encodes the probabilities of different stimulus configurations based on the incoming data. This activity is not spontaneous but is very orderly and stimulus-dependent. This is one reason why multistable perception is so useful experimentally - by maximising the ambiguity of the stimulus we can investigate perceptual inference and the neural mechanisms which cause us to settle on a unique interpretation.

 

[zoobyshoe]

The neuron itself is just switching rates. What does it know about anything?

The obvious counter to that is, "What does the "larger brain" know about anything?" As madness admitted, he has no idea how consciousness might arise from the Bayesian brain. How does the result of a calculation become aware of itself? The predictive modeling described is deemed to be a good description of what kind of calculation the brain is performing and nothing more. Consciousness isn't explained by it.

Disrupt that feedback-based balancing act and a part of the brain can generate vivid imagery - suppressed states can become expressed states. And the rest of the brain can try to integrate this chaotic activity into the general web of its activity.

I believe it's much more literally an experience of, consciousness of, the cortex as such:

Though migraine causes great suffering for millions of people, there has been much success, in the last decade or two, in understanding what goes on during attacks, and how to prevent or minimize them. But we still have only a very primitive understanding of what, to my mind, are among the most intriguing phenomena of migraine — the geometric hallucinations it so often evokes. What we can say, in general terms, is that these hallucinations reflect the minute anatomical organization, the cytoarchitecture, of the primary visual cortex, including its columnar structure — and the ways in which the activity of millions of nerve cells organizes itself to produce complex and ever-changing patterns. We can actually see, through such hallucinations, something of the dynamics of a large population of living nerve cells and, in particular, the role of what mathematicians term deterministic chaos in allowing complex patterns of activity to emerge throughout the visual cortex. This activity operates at a basic cellular level, far beneath the level of personal experience.

http://migraine.blogs.nytimes.com/2008/02/13/patterns/

The same would be true of any dedicated area of the cortex that is seizing or experiencing migraine hyperexitability. The sounds, smells, tactile sensations, tastes that happen during these episodes manifest something elemental about that area of cortex. The cortex isn't modeling the outside world during these episodes, obviously, but it also isn't just producing incoherent noise: it's demonstrating something about itself. The firing of visual neurons in the visual cortex, firing produced by any means whatever, produces a visual experience. Visual neurons fire-->visual experience. Olfactory neurons fire-->olefactory experience. Etc. During these paroxysmal episodes these neurons aren't listening to any top-down or bottom up input. The thalamus get entrained into the paroxysm and starts helping to sustain it.

Whatever thing you experienced that you are citing as a migraine aura doesn't concur with what I've read and heard. I don't know what you experienced but the following is typical of the description I usually encounter:

I have had migraines for most of my life; the first attack I remember occurred when I was 3 or 4 years old. I was playing in the garden when a brilliant, shimmering light appeared to my left — dazzlingly bright, almost as bright as the sun. It expanded, becoming an enormous shimmering semicircle stretching from the ground to the sky, with sharp zigzagging borders and brilliant blue and orange colors. Then, behind the brightness, came a blindness, an emptiness in my field of vision, and soon I could see almost nothing on my left side. I was terrified — what was happening? My sight returned to normal in a few minutes, but these were the longest minutes I had ever experienced.

http://migraine.blogs.nytimes.com/2008/02/13/patterns/

Evo's was the same. This is not an experience that has to be learned to be attended to to become conscious of it. Attention to it is automatic and involuntary: it's an extreme experience.

It seems to me the Bayesian calculation is thrown into disarray during a paroxysm: both the reality and the model to compare it to become unavailable for the duration. Yet consciousness is preserved.

 

[atyy]

In the context of perception it generally means a model of the external world. For self-awarenses and introspection there should also be a model of the self. I don't know why you think the activity would be spontaneous. Under bayesian theories of perception the activity encodes the probabilities of different stimulus configurations based on the incoming data. This activity is not spontaneous but is very orderly and stimulus-dependent. This is one reason why multistable perception is so useful experimentally - by maximising the ambiguity of the stimulus we can investigate perceptual inference and the neural mechanisms which cause us to settle on a unique interpretation.

I didn't mean that spontaneous activity is unpatterned. Fundamentally, if we assume classical physics as a sufficient basis for neural function, there is of course no such thing as spontaneous activity - it depends on a choice of coarse grained variables, such that experimental preparations that are identical at a coarse scale are different on a fine scale. Anyway, the coarse scale is convenient, and described by a probability model. In Knill and Pouget's proposal, the spontaneous activity I am thinking about is similar in spirit to the Poisson noise they mention - but their proposal is more specific.

 

[zoobyshoe]

I had an English class and this blonde girl with large jowls sat across from me. All I could do was stare.

And not paying attention, I was asked what is the cadence (or something) Shakespeare's sonnets are written in.

I gave a good look into the top of my head and answered iambic pentameter. We still haven't f'd. But I've alway wondered where the answer came from.

There are only 26 letters in the alphabet. nouns and verbs, nouns and verbs. Richard Dietrich.

Ultimately the world hits a rhythm. Dip-switch logic falls into play and people follow the politics.

Besides, animals can communicate with linguistic competency. They tend to follow the common language, as it is as free as you might possesses it.

You read a lot of James Ellroy?

 

[madness]

I didn't mean that spontaneous activity is unpatterned. Fundamentally, if we assume classical physics as a sufficient basis for neural function, there is of course no such thing as spontaneous activity - it depends on a choice of coarse grained variables, such that experimental preparations that are identical at a coarse scale are different on a fine scale. Anyway, the coarse scale is convenient, and described by a probability model. In Knill and Pouget's proposal, the spontaneous activity I am thinking about is similar in spirit to the Poisson noise they mention - but their proposal is more specific.

There is certainly a lot of noise in the nervous system - that is, variability in response to a stimulus etc., although there is debate over whether it really is noise. In general, however, spontaneous activity refers to activity which is not related to the onset of a stimulus or some other cognitive task (such as attention, remembering). See this article for an interesting example of spontaneous activity http://www.unicog.org/publications/PNAS-2008-Hesselmann.pdf.

The noise in response to a stimulus is generally considered to be unpatterned, although it can covary with other variables. For population coding models such as Pouget's, the neurons respond with some stimulus-specific firing rate (the tuning curve of the neuron) plus some noise. For a large population the noise can actually increase the information content of the network in some cases. However, it wouldn't exactly be to correct that the stimulus variables are represented by spontaneous activity alone.

 

[atyy]

There is certainly a lot of noise in the nervous system - that is, variability in response to a stimulus etc., although there is debate over whether it really is noise. In general, however, spontaneous activity refers to activity which is not related to the onset of a stimulus or some other cognitive task (such as attention, remembering). See this article for an interesting example of spontaneous activity http://www.unicog.org/publications/PNAS-2008-Hesselmann.pdf.

The noise in response to a stimulus is generally considered to be unpatterned, although it can covary with other variables. For population coding models such as Pouget's, the neurons respond with some stimulus-specific firing rate (the tuning curve of the neuron) plus some noise. For a large population the noise can actually increase the information content of the network in some cases. However, it wouldn't exactly be to correct that the stimulus variables are represented by spontaneous activity alone.

Thanks for the Hesselmann reference. It's terrific!

Do you have any more recommendations for reading work of similar quality about spontaneous activity? Just glancing at Hesselmann's references, it looks like (4) and (5) are in the same spirit.

 

[madness]

Thanks for the Hesselmann reference. It's terrific!

Do you have any more recommendations for reading work of similar quality about spontaneous activity? Just glancing at Hesselmann's references, it looks like (4) and (5) are in the same spirit.

I would say (26)-(30) from that paper, if you are interested in what spontaneous activity actually is and what it relates do (which is mostly unknown).

There are some more by Hesselman (note that they claim their studies as evidence for predictive coding, which I mentioned earlier):

http://www.unicog.org/publications/sadaghiani-fnsy.pdf
http://www.jneurosci.org/content/28/53/14481.full
http://www.jneurosci.org/content/29/42/13410.full

The motion one is basically the same study as the Hesselman study you already read but based on motion coherence rather than the Rubin vase.

 

[atyy]

I would say (26)-(30) from that paper, if you are interested in what spontaneous activity actually is and what it relates do (which is mostly unknown).

There are some more by Hesselman (note that they claim their studies as evidence for predictive coding, which I mentioned earlier):

http://www.unicog.org/publications/sadaghiani-fnsy.pdf
http://www.jneurosci.org/content/28/53/14481.full
http://www.jneurosci.org/content/29/42/13410.full

The motion one is basically the same study as the Hesselman study you already read but based on motion coherence rather than the Rubin vase.

Thank you very much! I'm like it that they are also looking at auditory tasks! I am very interested in spontaneous activity and neural variability, but was completely unaware of the functional imaging work on it. (I did know the approaches associated with Knill and Pouget.) One of the things that spontaneous activity (or neural variability) is supposed to do is enable "trial and error" learning. It's a long line of thought from Sutton and Barto's reinforcement learning, and also the theme of recent work in singing birds, and motor learning. Is there any functional imaging work on this?

Hmm, glancing through the Sadaghiani review on which Friston is a co-author, I don't see any mention of it, although Friston has worked on cholinergic plasticity, which I associate with theories of the synaptic rules underlying reinforcement learning.

 

[madness]

Thank you very much! I'm like it that they are also looking at auditory tasks! I am very interested in spontaneous activity and neural variability, but was completely unaware of the functional imaging work on it. (I did know the approaches associated with Knill and Pouget.) One of the things that spontaneous activity (or neural variability) is supposed to do is enable "trial and error" learning. It's a long line of thought from Sutton and Barto's reinforcement learning, and also the theme of recent work in singing birds, and motor learning. Is there any functional imaging work on this?

Hmm, glancing through the Sadaghiani review on which Friston is a co-author, I don't see any mention of it, although Friston has worked on cholinergic plasticity, which I associate with theories of the synaptic rules underlying reinforcement learning.

I'm not sure about the relation to reinforcement learning, it's not something I've ever read into very deeply.