Hacker News new | comments | show | ask | jobs | submit login
The spotlight of attention is more like a strobe, say researchers (princeton.edu)
131 points by daddy_drank 13 days ago | hide | past | web | favorite | 53 comments

Long time mediators are familiar with this. When you drill down attention to smaller and smaller time resolution, eventually you notice that attention flickers.

Buddhist Abhidhamma sutra describes this phenomenon. Smallest flicker of consciousness is called ceta. I think 250 ms is closer to vithi (kind of small molecule of mind- moments), than ceta.


> We have seen that the Abhidharma’s analysis of sentient experience reveals that what we perceive as a temporally extended, uninterrupted flow of phenomena is, in fact, a rapidly occurring sequence of causally connected consciousness moments or cittas


> The Sarvāstivādins use the term “moment” (kṣaṇa) in a highly technical sense as the smallest, definite unit of time that cannot be subdivided, the length of which came to be equated with the duration of mental events as the briefest conceivable entities. There is no Sarvāstivādin consensus on the length of a moment, but the texts indicate figures between 0.13 and 13 milliseconds in modern terms

(My personal subjective estimate is something like to 40 ms)

Oh wow! I just recently stumbled on this in my personal meditation practice. I noticed that it's almost like there's a game loop that periodically hits the CartesianSelfAssertion function. Or something. Everyone I've tried to talk to so far has just given blank stares, so it's kind to nice to run across someone with similar experience. Thanks for sharing.

> I noticed that it's almost like there's a game loop that periodically hits the CartesianSelfAssertion function. Or something.

Could you go into a bit more detail? What does CartesianSelfAssertion feel like? How often is it called---on the order of 40ms per OP, or less frequently? Do you notice this only during/after meditation?

> Could you go into a bit more detail?

I'm not super lucid on it yet. Just recently encountered the experience and still getting a feel for what it is, but I'll try to share what I can.

> What does CartesianSelfAssertion feel like?

A bit disorienting, actually. Though that could just be a function of it's current novelty for me. It has aspects of feeling like waking up, or remembering something crucial, or turning a corner and finally finding your keys.

To philosiphize a bit, I'm tempted to say it's like recurively having a realization, "Oh! This is the real deal... Oh wait, that was just a dream, THIS is the real deal... Wait, no, THIS is it..." which seems quite utterly silly.

And completely independent of all that, I also feel like I'm spinning clockwise, or at least at the center of some spinning thought-sensation storm. And the angle of 30 degrees is in the mix as well for some reason. I have no idea what to make of either of these.

Then there's also a part of me that wonders if all the above is just nonsense, because it just sounds so tripped out. lol

> How often is it called---on the order of 40ms per OP, or less frequently?

Honestly, timing the interval never really occurred to me until reading the OP's comment. In my limited experience, there's definitely some bit of flexibility. In extreme cases I've even been able to halt the process altogether for tens of seconds or so, but modulo doing weird stuff, my gut feeling is to say somewhere on the order of 100ms---40ms definitely doesn't clash with my experience.

> Do you notice this only during/after meditation?

It's been most blatant in some of my more recent meditation sessions. Durning the normal course of the day it's mostly way in the background, though I can also sort of "consciously access" that awareness space to some level. Though the mathematician in me feels like "conscious access of sub-awareness processes" should be a contradiction.

The spinning at an angle is very likely you reaching self awareness with enough fidelity to become aware of your own inner ear noise floor.

Like being in a silent room eventually relaxes your ear muscles and lack of other stimulation will bring out whatever birdies, tones, noise in your hearing, no movement and concentration can bring out same in your inner ear.

As we age, these artifacts tend to accumulate.

Happens in eyes too, but manifests a bit differently. What happens there is we need more photons to register the same sensory brightness levels. When we are in the dark for a while, the retinal noise floor can be more easily seen, like when star gazing.

Secondly, blood flow, other things can be seen in that noise floor. That happens better in consistent, lower light, but not dark environments, in my experience.

All of these are normally filtered in much the same way we filter our nose out of our visual field, unless we concentrate awareness there.

How do I know?

Well, we know nothing. But, my experiences come from three things:

I have a very good sensory recall. I know what it used to sound, hear, feel like with high fidelity.

My own meditative activities.

Some built in degree of self awareness. When I relax a bit, those filters come down fairly easily for me. Have always been that way and was confused by it more than once growing up.

Thanks for sharing, I can definitely relate. Good to know others have the same experience.

I've also been trying to think how this can be used to design AGI, the goal being a machine that is able to recursively reason about it reasoning about itself etc... it would be logical if this was in discrete blocks that contained some representation of the previous state so as to avoid just going in circles. Maybe this is part of the reason why we so easily get distracted when observing ourselves observing ourselves - without some noise the mind would just endlessly go meta without any insight.

Thank you for sharing.

> A bit disorienting, actually. Though that could just be a function of it's current novelty for me. It has aspects of feeling like waking up, or remembering something crucial, or turning a corner and finally finding your keys.

Ok, that's definitely what I experience too. I have also described it as "disorienting" and "like waking up". I like your name, CartesianSelfAssertion. I might steal that.

The weird bit is that I experience this much less frequently. Like on the order of minutes or hours instead of ms. And typically after but not during meditation. During meditation I'm busy catching my mind wandering off.

> And completely independent of all that, I also feel like I'm spinning clockwise

Ok that bit's a little weird :-).

> but modulo doing weird stuff, my gut feeling is to say somewhere on the order of 100ms---40ms definitely doesn't clash with my experience.

One idea I had for measuring this (because I can't read about something without jumping to how to measure it) is to try to compare the spacing of the moments against the peaks of an auditory sine (or similar) wave:


Here's an excerpt from my meditation journal, about what it feels like to be in this state of mind (the period of time just after CartesianSelfAssertion):

A change in perspective. Before:

       past    present    future

    # - concrete (senses)
    @ - abstract (memories; hypotheticals)

    past present future
      @  ←  #  →  @      past
      @  ←  #  →  @      present
      @  ←  #  →  @      future

    It's a chain of links, but you're only ever one link at a time.
    There's a lack of sense of self, probably due to the isolation of the links.
(This has now happened a _lot_ of times. Why does it not feel ordinary? Jesus fuck. Ok, I guess I'll just run with it.)

I've always felt it more akin to garbage collection triggered by hitting a resource limit.

(not being sarcastic)

Would you mind elaborating a bit? This is something I've recently been trying to get more intimately familiar with and new metaphors are a really helpful rubric.

With the caveat that I think stretching the metaphor too much further would exceed its utility, my point is only that awareness is not a smooth process, but occurs in fits and starts, slightly delayed from actual reality- it takes time for a brain to assemble meaning from its sensorium.

Different parts of the brain can work semi-independently, while other parts are strictly single-use. And 'resetting' one's mental machinery after this kind of interference takes a measurable amount of time, much like the triggering of garbage collection.

The Stroop Effect is a good demonstration of this kind of cognitive collision.



That’s a nice way to put it. Descartes wrote “I exist, that’s certain—but how often?”

Does anyone else notice the length of a moment changing? I especially notice a difference if something causes me to trigger an adrenaline rush; I noticed it a lot when playing DDR and similar, since the physical motion would trigger adrenaline, and then the tempo of whatever I was playing would appear to slow.

It was like ... to use a game developer's terms, like my perception was suddenly running at a higher framerate. Of course as a musician this meant the music felt "wrong" somehow after that. "Why is it so slow? Was this song always this easy?" and relatedly, "Why is it harder to move my legs quickly?" etc. This persisted for maybe 20 minutes and then returned to normal gradually. Once I pieced together the criteria, I started to notice that sort of tempo... kind of a "beat" in everything that I do. I feel it especially when I'm visually analyzing a scene, as it's the speed at which I can control my eyes darting about a room. If I'm tired that slows down considerably.

The brain is just... so fascinating.

> Does anyone else notice the length of a moment changing?

Sure- a couple beers makes a couple hours feel like a couple minutes.

Similar to the time distance of an Elvis style slapback vocal echo. Less than ~40ms and it’s not perceived as a distinct sound, more than ~120 ms and it starts to sound obvious like a reggae dub echo. Same sound perception time constants show up all over audio recording.

I did not know exactly what Elvis-style slapback echo is, so here's the fruit of my moments of poking around: https://www.youtube.com/watch?v=FuStmPbG528 About 1:30 long.

It's generically echo or delay.

Normally, it's called slapback when it's used for "rockabilly" (yes, basically sounding like Elvis--specifically his guitarist Scotty Moore).

People emulating Brian Setzer is the more modern thing: https://www.youtube.com/watch?v=-PY4CSScCKY https://www.youtube.com/watch?v=EW54TX4Pwl8

When playing software instruments with latency, most people say the threshold for percieving the latency is between 10-30ms

> When playing software instruments with latency, most people say the threshold for percieving the latency is between 10-30ms

I really challenge the lower end of this. Sound travels about a foot per millisecond. Bands regularly play at 10ft distances.

The problem is that 10ms latency on a digital audio workstation generally isn't if you start measuring with an oscilloscope.

There’s a huge difference between the feedback loop of playing an instrument and the one-way experience of listening to a band, though.

Bands typically are using in-ear or floor monitors. Main PA latency won't affect their playing because it's not what they're paying attention to.

It's always mind-blowing to hear the precision and scientific rigor of the ancient Buddhist monks from two millennia ago and how they managed to create what is ultimately technology to observe the phenomena that comprise consciousness to the smallest perceivable level.

How do you know what 40 ms feels like as opposed to say 60?

If you're used to noticing the dots-and-dashes from flicking your eyes past a 60hz (non-incandescent, say LED) light (each about 17ns apart), - something that happens to most of us many times per day - 40 vs 60 is a pretty obvious difference, and 17ns an obvious measure for comparison.

> My personal subjective estimate is something like to 40 ms

This is fascinating. I'm waiting for the time in my meditation practice when I start to notice it.

Could you (or anyone else who experiences individual moments) go into a bit more detail about what you experience? Do the moments tend to all be from the same sensory modality (e.g. vision), or are they mixed? If you can tell, does any moment combine more than one sensory modality? How often are there moments of thinking (e.g. narration or a visually imagined scenario), or does thinking tend to end the experience? How isolated do the moments feel: are they like isolated snapshots, or is it like a movie running slow where you can tell that there are frames but they partially blend together?

Cannot strongly recommend enough "The Mind Illuminated" by John Yates.


It's been discussed on HN before (it's how I found it a few years ago) and breaks down meditation in a systematic way while relating the phenomena described in Buddhist texts to current psychological principles. This "moments of consciousness model" of the mind is discussed at length and a short answer to you question is yes, different sensory moments are integrated in "binding moments".

Yeah, I own it. "Binding moments" is why I asked that question :-).

I'm partly wondering if peoples' experience matches the book's description, and partly looking for more detail about what the actual experience is like than the book provides.

Would second the recommendation though: it's a great book.

To see the "spotlight of attention" masterfully exploited, I recommend Apollo Robbins[1]. Yes, his slight of hand is very impressive, but his act is fundamentally about how he fills up your attention input buffer, notices where you are allocating your "attention spotlight", and works with a blatant flourish just outside of that spotlight.

The overlap with "walk in with confidence holding a clipboard" approah to security is obvious; Apollo Robbins has even given a talk at DEFCON[2] about the implications of this approach to attention & security.

I find his demonstrations a very humbling lesson in just how weak our actual perception of reality really is.

[1] https://video.newyorker.com/watch/apollo-robbins-tricks-of-t...

[2] https://www.youtube.com/watch?v=1kkOKvPrdZ4

Ooh that's nice, thank you.

I recently figured out that keeping two separate items in one hand, say my bicycle keys and my house keys, makes it easier for me to lose one of them without realizing it since I'm still feeling the other. So there's no sense of loss.

I am curious if this has any relation to saccades. It feels to me like it would be easier to retain attention on a moving object than a stationary one.

Not substantial but found this in a quick search of any support for this conjecture: https://www.tandfonline.com/doi/abs/10.1080/0004953760825526...

Our perception of reality is just a bunch of hacks, God is a programmer with a deadline confirmed.

In all seriousness, it's interesting to think about the design tradeoffs evolution came up with for humans, and how those tradeoffs may no longer be the right tradeoffs given our radically changed lifestyles. If only we could flick focus mode like a switch when sitting at a computer, that would be nifty.

> If only we could flick focus mode like a switch when sitting at a computer, that would be nifty.

But you do. When you engage in a task, the dopamine and noradrenaline systems in your brain modulate which parts are active, directing your attention. If these systems don't work well enough, then you get ADHD.

If what you mean is to be able to disable involuntary attention control, then that's not good either. You'd be at great risk of under-responding to emerging situations.

Plot twist: the human brain is actually running JavaScript and attention is executed on each tick of the event loop.

The article is paywalled, but it looks like the concept of flow was not considered in the human study conducted by the authors, judging only by the short summary of the paper.

I suspect this changes under conditions of flow, where a human can spend hours engrossed in one task and totally lose track of his surroundings.

Flow as described by Mihaly Csikszentmihalyi.¹

It would be interesting to see if experimental results differ between when a person is in flow-state and when they are not.

1. https://en.wikipedia.org/wiki/Mihaly_Csikszentmihalyi#Flow

Makes me wonder if there might be any advantages to modulating Softmax attention in deep artificial neural nets, with some kind of mechanism for recomputing attention based on each previous attention.

I came here hoping to read this. I'm not clever enough to implement it, but my first thought was that someone would see a connection here.

They say this flicker likely provides an evolutionary advantage-a chance to switch attention. I wonder if it’s optional. Wonder also about cases like attention disorders and attention of autistics.

Attention deficit disorders are known to be caused primarily by problems with neurotransmitter systems which modulate attention (e.g. dopamine and noradrenaline). This seems to be more like a clock signal for the brain.

I imagine problems with this system would manifest more as Sensory Processing Disorder, which has differences noticeable in an EEG.

Could this be in any way related to the strobing effects (and afterimages / discrete tracers) of LSD and other psychedelics?

(see also http://psychedelic-information-theory.com/The-Control-Interr... - not sure how much of it is just pseudoscience nonsense though)

“Every 250 milliseconds, you have an opportunity to switch attention,” said Ian Fiebelkorn,

Really makes you rethink the speed of website load time. I know there is a ton of research around site abandonment after x seconds (ms in some cases) but I’ve tended to think drop rates were because the time spent was the upper bound of what that user would spend of “continuous” attention. Reading this article it makes me think that if the site takes 4 seconds to load, the user had about 16 opportunities to go do something else. While the abandonment outcomes are what they are, the cause seems to be very different than what I had always thought.

I think it's also important to note that when looking at a blank screen there is nothing stimulating, and stimulation (external or internal) drives our attention control. Putting a progress indicator on something gives us something to pay attention to, and so we're more likely to respond with frustration than inattention.

Sounds like a timer interrupt invoking a scheduler. Interesting!

Given my (anecdotal) experience with audio books, I'd have to agree. I tried listen to mostly non-fiction in the car while on the highway (read: not local traffic) and despite the fact it was only me and the road I found myself fading in and out so often that I stop listening and went back to reading.

Since, I've heard / read that setting play back to a faster speed helps - because of the novelty? - but I haven't tried it myself.

The first time I heard the spotlight analogy from a neuroscientist, I asked for what backs this up and the response I got was, "it's been shown" with a tone that made it clear no argument or discussion was allowed. Later I found out it comes from some behavioral studies that are pretty contrived.

I think I experience the opposite, focusing on nothing in particular and only occasional brief flashes of focus, when I need to focus on something. Are people really more or less unaware of what is going on around them all the time?

I don't really understand any practical consequences of this.

Apparently there aren't many: no one (except perhaps a few buddhists who spent an unreasonable amount of time meditating) noticed until now. The interesting bit is that we don't notice.

The finding highlights and perhaps provides somewhat of a meaningful answer to philosophical arguments around what we perceive to be reality around us in fact being very much an illusion of continuity. The fact that we so readily trick ourselves into believing, and only through intense mindfulness/introspection are we able to contradict that sense of internal consistency, is quite profound in my opinion.

For folks who are trying to be more focused, it seems like this emphasises the importance of training yourself to re-focus your attention

Is the attention spotlight synchronised with the brain's theta waves?

I saw a researcher present a model of theta as a siren light (vs strobe), that moves from the back of the brain to the front, in an oscillatory loop. Where theta is active, the whole of attention is integrating with the local input of the part. As a general principle, we look for "phase amplitude coupling" to describe the integration of fast (gamma) local oscillations with slower (theta) global oscillations.

Another well established oscillatory model describes binding in visual perception -- which is based on the synchronization of 40hz inhibitory interneurons in different parts of the visual field. Different parts of the same object have a synchronized inhibitory input. Theory is called "binding by inhibition gating" or " communication through coherence"

“Theta phase acts as a clocking mechanism, organizing alternating attentional states”

From the research paper linked to by article.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact