Attention and awareness: A two-stroke model of consciousness

Posted on 28 October 2023 by Cube Flipper

The overarching point of this video, which is that in the model that I am currently working on, attention and awareness work through this more general phenomenon that I call oscillatory complementarity – so attention and awareness are the oscillatory complements of each other.

Andrés Gómez Emilsson published a video last year on attention and awareness – which I have to say I found quite confusing. Attention and awareness are oscillatory complements of one another? What could that mean?

This came not long after I’d taken Michael Ashcroft’s Alexander Technique course, which had handed me a fairly nuts and bolts model of attention and awareness that I got a lot of utility out of – so these were terms I’d been thinking about. Here’s how I usually explain it to people: You have awareness, which corresponds to everything currently in your sensorium. Then you have attention, which is a subset of that – like the beam of a spotlight – and most importantly, you have agency over it, you can choose where to point it and how wide or narrow you would like it to be.

This had prompted me to do a little reading into attention and awareness, and somebody recommended a Victor Lamme paper which lead me to associate awareness with fast but unconscious feed-forward processing, and attention with slow but conscious recurrent processing of stimuli. An example might be – I’m driving down the street and catch an object out of the corner of my eye. I can’t recognise it because it’s outside my attentional spotlight, but it triggers enough of a prediction error to drag my attention over – oh, it’s just a seagull.

I’d also been experimenting with DMT, and what I understood of DMT was that it reifies the object of attention – easily observable by an increase in geometric symmetry. I had to infer that DMT was simply showing me an extreme version of a process which was always playing out when sober, and that attention corresponded to some process of reification – and perhaps this could also be thought of as like allocation of computational resources.

Anyway, these are the circles which I’d been trying to square for some time. Back to Andrés:

In meditation… there’s these very exotic things – you can actually turn awareness into attention and attention into awareness! So let’s say you’re focusing on a point very heavily, and that’s the oscillatory complement to the rest of your experience. And you start to expand and expand that, and the region that is associated with the oscillatory complement of the experience becomes larger and larger, until you get to this critical threshold where half of your experiential field is the oscillatory complement of the other half. And then you can start to narrow down the size of the oscillatory complement on the other side until you get a fixed point there.

So, isn’t that bizarre? That actually means that the awareness became the attention and the attention became the awareness. And you can do that while meditating and it’s really trippy.

And I think that non-dual states like you experience on LSD or higher doses of concentration meditation are actually kind of when you find this perfect balance where neither of the two sides is actually something you could say is more narrow than the other. So all of a sudden that breaks the paradigm of the flashlight of attention.

Well, there’s no substitute for first hand experience, and I thought that if I was ever going to get some perspective on this, I ought to develop my meditation skills. Well, as I found out – similar insights might be accessible while on 5-MeO-DMT.

I recently had the opportunity to experiment with five, mostly taking doses of about 2.5 mg. On the come up I would typically see some kind of turbulent patterns rendered in black and white – and these would then gradually merge over time.

On one occasion, these patterns stabilised such that I saw a region at the center of my visual field that was strobing black and white, surrounded by another region which was strobing one half-cycle out of phase. Suddenly the size and shape of my attentional aperture was made clear to me in the most straightforward possible terms:

The attentional spotlight, replicated at 5 Hz.

I’m not sure how I knew for sure that these corresponded to my attention and awareness – I kinda just did, and I seemed to have similar volitional control over them.

What happened next was fun. The attentional aperture spontaneously slid over and grew until it occupied one half of my visual field, and all I could see were these two strobing regions splitting my visual field horizontally:

Attention and awareness become interchangeable with one another.

It wasn’t until some time later that I realised this could be what Andrés had described in his video! Indeed the two regions had lost any phenomenal character which might distinguish them from one other.

On a separate occasion, attention and awareness merged entirely, and all I saw was a single pure strobe:

Attention and awareness become as one.

I have some discomfort with the term non-dual – but I guess I feel comfortable labelling this experience as such. The two-stroke cycle becomes a one-stroke cycle.

I should note that there are a few, possibly important details which I am having a hard time recalling:

  • I can’t remember how fast it was strobing – but my sense of time is distorted on five, anyhow.
  • I can’t remember if there was a sense of the waves arising and passing in time.
  • I can’t remember if there was a sense of the waves moving in space.
  • I could tell that the boundary overlapped different regions from cycle to cycle, but I’m not quite sure what it looked like. I used Perlin noise in my replication, but I think it might have been more jagged than that.

I thought I would try a little cannabinoid-assisted meditation while I am writing this. I took a big rip of ∆8-THC, put an eye mask on, put attention on my breath until my system settled down – and then concentrated my attention really hard on a single point in my fovea. I can just perceive individual cycles by studying the background noise in my visual field, and I can adjust my attentional aperture, but I don’t get any real sense of a hard boundary between attention and awareness, nor a sense of any antiphase strobing.

I guess if I wish to seek further insight into this phenomenon, I’ll just have to do more five – and do my meditation homework.

Open questions

I’ve been curious lately about whether or not consciousness might have a refresh rate, and what the timeline of a refresh cycle might look like. I felt that observing oscillatory complementarity provided me with my first clue as to the spatiotemporal shape of these cycles.

I will note that I have no reason to believe that an attention-awareness cycle is isomorphic with or even fits within the context of a consciousness refresh cycle, so far – but it seems like a reasonable starting assumption.

I felt that by writing a list of open questions I’d be able to provide myself with direction for future research – and perhaps some of the wider community could help illuminate the finer details of this system’s architecture. As always, I mostly just want to build a precise model of the system in question.

What might the exact refresh rate be?

I recently ran an informal survey on Twitter, asking the crowd what they thought the refresh rate of consciousness might be. The question was open to interpretation and I received a wide variety of answers, from 1 Hz to the Planck interval – with a couple of clusters at around 10 Hz and 40 Hz.

We have some amount of informal consensus within our circles that five generates vibrations at around 40 Hz. We also know that five generates smooth tracers, compared to other psychedelics which generate various strobing effects with different frequencies. Going by the control interrupt model, this might imply that five is interrupting consciousness at its actual refresh rate – or at an exact integer fraction thereof.

Does the refresh rate drift over time?

A handful of commenters including Nick Cammarata suggested that the refresh rate might be adaptive, changing its rate over time in response to various moods or stimuli. Brad Caldwell also suggested that NMDA antagonism slows down the refresh rate – and I guess this aligns with what Andrés has said elsewhere about ketamine slowing down the wave propagation rate.

Additionally, from 5-MeO-DMT vs. N,N-DMT: The 9 Lenses:

It is worth mentioning that 5-MeO-DMT is probably not hitting the right frequencies merely by chance. It’s probably more that it is activating a system whose attractor is self-correcting and results in the kind of symmetrical crystallization that gives rise to deep feelings of bliss.

It could be interesting to see if people provide different estimates of the frequency of five in different situations. I feel this would provide credence to the idea that it is in fact phase-locking, rather than operating at its own frequency.

The obvious route is mixing five with ketamine, but my instincts say this is a horrible idea. This said, I’m just as interested to see whether other psychedelics phase-lock in this manner as well, and many of these do mix nicely with ketamine. 2C-B generates nice slow strobing tracers, and so I think it could be educational to record some 2C-B tracers both with and without adding ketamine. Perhaps there’s a threshold dose, beyond which phase-locking breaks and the characteristics of the tracers change.

If we do find consistent phase-locking, perhaps we could then start looking at the binding affinities of different psychedelics, and speculating about what the purpose of different serotonin receptor networks might be insofar as they correspond with different phase-locked control interrupts.

Do different sensory modalities have different refresh rates?

Other commenters also suggested that the somatic and visual fields might have different refresh rates. Brad Caldwell said he thinks the frames and paint – which he says correspond to amodal and modal perception – refresh at 10-20+ Hz and 40-100+ Hz, respectively.

I think the control interrupt is likely felt in the somatic fields as body load. I find my body load frequency changes while on different psychedelics in a manner correlated with the tracer tool results – psilocybin is quite high frequency, and ayahuasca is higher still. I have not, however, tested whether or not these frequencies are the same as those reported by the tracer tool – perhaps if they were different, we could conclude that temporal aliasing effects are involved.

It could be interesting if we found that different psychedelics phase-lock with different sensory refresh rates. Perhaps different serotonin receptor networks are responsible for annealing different sensory modalities, or other systems which have less of an influence upon our phenomenal fields – or, perhaps there is a global refresh rate, and its frequency is dependent upon which sensory modality attention is focused on at the time.

I’ve only spoken about somatic and visual qualia. I’m not sure what’s going on with auditory qualia. Those are weird.

How does this work with multiple attentional streams?

It makes intuitive sense that it’s possible to juggle multiple attentional streams, but it would be nice to have some idea how these work at a more granular level. There’s a more recent video Andrés published where he discusses the insights he had on a jhāna retreat:

On the previous retreat it became clear to me that we actually have something like seven attention heads, and the way in which you access the first jhāna is by synchronising the sensory doors enough that they acquire the proper vibe – this would be the pīti – such that wherever the attention heads settle, they will be vibrating exactly the same way. And when they’re vibrating the same way, they actually collapse into just one, and as a consequence that gives rise to ekaggatā – which is single-pointed attention.

I had a few finer details I wished to ask him about:

Is a single background field of awareness shared between several attention heads, or are there several pairs of attention heads and awareness fields?

There is an awareness side to each of the attention streams, so ±7 as well.

Are these attention heads normally phase-locked with one other, or do they run at independent frequencies?

Very good question! This is where the richness of this paradigm shines. Any number of attention heads can be synchronized with each other. Often they will do this via harmonic relations (so they don’t have to have the exact same frequency). One attention stream could be vibrating twice as fast as the other and still sync up properly with it. More so, this also adds a certain degree of non-transitivity. Attention streams A, B, and C would be flickering at, say 6, 12 and 15 Hz, respectively. Here A would be strongly connected to B and C, but B and C will be only slightly connected due to their less harmonic relationship.

This also seems to happen with the awareness components of the cycles/strokes. Meaning that two attention heads could be pointed at very different things and doing so at different frequencies, but still overlap on where that gets broadcasted. Awareness can pool the contents gathered by different attention heads.

I think this is what makes DMT so exotic: You get to experience the marvelous tapestry of all of these possible combinations. Where, say, three of your attention heads are interlocked, but their awareness side are different, but then one of those awareness sides could be interlocking with the awareness side of yet another attention stream. Or suddenly five of your attention heads fully sync up, and a sixth plays the role of an observer of this process and the seventh plays the role of observing this oberservation.

Each of the streams functions as a frequency and location specific container of qualia and information, and their synching up (harmonious or not, via awareness or not) is what determines how the contents in these containers are shuffled around and blended with one another.

It looks as if our model is getting quite complicated – attention and awareness can implement just about any network topology with weighted coupling! Do we now need to be talking about a fourteen-stroke model of consciousness?

It’s fun to consider: What might these attentional streams look like for different neurotypes, or even different animals – such as a chameleon, with its independently mobile eyes?

What happens within a cycle?

I have no first hand perspective to offer, here. However, Rival Voices recently linked a post by Kenneth Shinozuka which I had not seen before, Shinzen Young’s 10-Step Model for Experiencing the Eternal Now – which contains plenty of great descriptions and diagrams.

Now, we start to get wavy. I recently experienced these ripples and vibrations when I was practicing kundalini yoga. After the session ended, I noticed that my entire visual field was flickering at a high frequency while my eyes were open. It seemed like my visual field had arranged itself into a series of uniformly distributed Gabor patches that were fading in and out very quickly. I have yet to perceive the wavy essence of mental imagery or mental talk, though.

This is some extremely dank phenomenology. Much to ponder, here.

What neural correlates might these cycles have?

Brad Caldwell has been thinking about this sort of thing for a long time. He takes a very QRI-esque phenomenology first approach to his research which I find myself quite sympathetic to, and he seems fairly comfortable spitballing what the neural correlates of what he observes might be. I’ve been telling him I’ll read his book for some time, after which I’d like to have a call with him; but I’ve sadly lacked the time to do so while I’ve been engaged with other work.

I find myself wondering, is the boundary between attention and awareness one and the same as the rings of fire that he talks about? His phenomenological model doesn’t quite align with ours, but I think this is what tends to happen when phenomenologists work in isolation. I suggest that some cross-pollination would be fruitful.

What electromagnetic correlates might these cycles have?

By this, I mean: What might be happening in the topological pocket during one of these cycles?

I’m just going to leave this here:

I still only have a fairly naïve understanding of electrodynamics, but I’m not going to let that stop me rotating some shapes in my head. Let’s say the control signal dictates the shape of the pocket; do attention and awareness – whatever they may be – oscillate back and forth within that? Maybe the boundary between attention and awareness corresponds to the nodes between standing waves – where their amplitude is zero? Could it be as simple as that?

Mapping back from physical space to qualia space – why would one appear black to me while on five, while the other appears white?

What are attention and awareness?

Electromagnetic radiation? Electric field lines, aware of themselves? Electrons trapped in the topological pocket? Or something more quantum, like wavefunction collapse?

I have no idea, and plan to cultivate agnosticism towards settling on a particular physical correlate for now. I’m starting to work with a model of qualia as probability distributions, and I guess I’ll see where that leads.

What makes attention and awareness different?

So we now understand that attention and awareness are two sides of the same coin; so what makes it reasonable to discuss them in separate terms?

I discussed this with Andrés the other day, and he suggested the following definitions:

  • Attention is where the waves converge.
  • Awareness is where the waves radiate out and bounce off the context.

We have not yet formalised this further, however. These definitions could describe one of two things, which might ultimately turn out to be the same thing:

In the context of this post, I’m working with the former definition.

Do attention and awareness follow a branched flow structure?

Does it make sense to think of attention and awareness as rays or waves? If so, then perhaps the shape of the boundary between attention and awareness could indicate whether or not it follows a branched flow process?

A laser propagating inside a soap membrane. Posted by Andrés in Qualia Computing Networking.

From Branched flow (Heller et. al., 2021):

Branched flow is the spatial pattern that forms when a wave or a bundle of rays is launched over a smooth, weakly refracting random potential. The pattern begins (but doesn’t end) with singularities called cusps. In optics, such singularities are called caustics and typically result from refraction or reflection by curved surfaces. Along caustics, the number of ray solutions passing through a point in coordinate space changes abruptly.

I guess we already take it that attention involves a large number of waves or rays focusing upon a single point – but perhaps if the followed a branched flow pattern, they could also focus upon every point of fixation?

The paper goes into detail on the prerequisites for branched flow:

To have branched flow in a system, its dimensions must be smaller than the mean free path, or, at least, the observation time must be shorter than the mean free time. Beyond the mean free path, rays will start to turn around, and their propagation will no longer be ballistic but diffusive. But in weakly refracting media, when ε/E ≪ 1, the mean free path l = 〈∣v∣〉τ ∝ (ε/E)−2d is much larger than the typical branching length scale, which scales like (ε/E)−2/3. Thus for a wide regime, wave propagation is dominated by branched flow.

As mentioned earlier, branched flow has two other prerequisites. First, the source needs to be restricted in phase space – for example, a point, parallel ray, or plane wave. Sources with fuzzy but still somewhat localized manifolds, however, also produce pronounced branched flows and lead to extreme events. The second prerequisite is that the flows’ wavelength needs to be smaller than the correlation length of the random medium.

If branched flow is indeed a requirement for correct functioning of attention and awareness, perhaps these constraints could help us narrow down the following topological pocket parameters:

  • The maximum size of the pocket
  • The wave propagation rate within the pocket
  • The size and depth of the topographic or topological defects which the waves may encounter

This could also explain how ketamine works. If ketamine slows down the wave propagation rate, branched flow would turn into diffusive flow more readily – impeding the process of reification, and turning things into stuff.

I really wish I’d paid more attention to that boundary, when I’d had the chance.

What are attention and awareness doing?

I have a strong impression that attention is tightly coupled with reification in some way. So I feel that we might also ask: What is reification?

I’m still not really sure, but I have a simple example of how attention modulates reification:

A Necker cube, the classic example of a bistable percept.

My experience of the Necker cube is that I can choose to reify it in one direction or the other by choosing to place my attention at different fixation points. It’s as if the depth map flows outwards from the center of my attention, individual faces of the cube stabilising independently until the entire gestalt is formed. I can ever so subtly percieve slight ringing artifacts around each face as it resolves. Brad Caldwell reports a similar experience.

It would be nice to have an experience of absence of reification to compare this to. I’ve already written about what ketamine does to my visual experience, but I also ran across this fantastic report from @itinerantfog on Twitter:

I’ve directly experienced the loss of ‘reification of depth’ and seen movement through space as just shifting colors in shikantaza. It is easily one of the wildest things I’ve ever seen. An infinite scintillation. I’ve never seen anything with such diversity and richness and life as the visual field without seeing depth.

Exactly what I was doing was sitting in a meditation posture with my eyes open looking at nothing in particular doing nothing in particular. This was about day three. I had done this for over two hours per sit, all but one day over four days. When I stood after I noticed space shifting and moving like one may have experienced on psychedelic drugs. The walls melting, the floor twisting. As I continued this in a standing posture, just observing it, for about fifteen more minutes, and then shifted my head back and forth, sense of 3D space was gone.

I recognized nothing. The sparkle, the way it all glittered. Scintillation is the perfect word. There was nothing still at all. Every ‘pixel’ of my vision was dancing.

This was somewhat concerning for a first time experience so I didn’t engage it all very much and haven’t since then. Many more details, but that’s the gist.

What do attention and awareness want to do?

As silly as it might sound, I sometimes like to ascribe an anthropomorphic notion of volition to physical systems, as it makes it easier for me to reason about them. Perhaps this is the one context in which I can get away with it – for after all, are we not one and the same with these particular physical systems?

Anyway – and never mind that free will might be fake news – might these waves of attention and awareness want something? By carrying out their desires – insofar as they might have any agency to do so – might they then perform valence gradient descent, and by extension, reification?

If they want something, what is it? Do they just want to avoid colliding with themselves? Do they like to have a little wiggle room, do they like to have space to breathe? Do they just want to avoid stepping on their own toes?