Subliminal Perception: Just How Fast Is The Brain?

By Neuroskeptic | January 8, 2015 3:43 pm

Subliminal perception has long been a hot topic. The idea that something (generally an image) could appear and disappear before us so quickly that it escapes conscious perception, and yet affect us subconsciously, is a fascinating (and scary) one.

Psychologists and neuroscientists are fairly skeptical of any grand or sinister claims for the power of subliminal advertising or propaganda, but on the other hand, many of them use the technique as a research tool.

So what’s the absolute speed limit of the brain? What’s the minimum time that a stimulus needs to appear in order to trigger a measurable brain response?

In a new study, Swiss researchers Holger Sperdin and colleagues say that they’ve detected neural activity in response to images presented for just 250 microseconds – that’s 1/4 of a millisecond, or 1/4000-th of a second. Sperdin et al. say that these ultra-brief stimuli are undetectable on a conscious level, yet still evoke a brain response – albeit a small one.

Here’s the paper: Submillisecond Unmasked Subliminal Visual Stimuli Evoke Electrical Brain Responses.

The authors recorded brain activity using EEG and presented the brief stimuli using a device they invented called the LCD Tachistoscope.


It’s built around two LCD monitors and a mirror. Sperdin et al. published the design last year. A normal screen just isn’t fast enough to present stimuli at microsecond durations. (Other tachistoscopes exist.)

While the apparatus was complex, the stimulus Sperdin et al. used was a simple checkerboard pattern of black and white squares. The researchers presented the stimuli for 250, 500 and 1000 microseconds. These are, respectively: consciously undetectable, only sometimes detectable, and easy to spot.

Here’s what happened. The black lines are the visual evoked EEG signal in response to the checkerboard; the red lines were a control stimulus:


Both the 500 and 1000 microsecond checkerboards clearly evoked a neural response which began roughly 80 milliseconds after the checkerboard appeared. The 250 microsecond stimulus did so too, albeit marginally. (Note that the graph’s scale varies.)

The bottom half of the diagram shows the statistical significance of the difference between the brain responses to the checkerboard and the control stimuli. For the 250 microsecond stimuli, only a few short periods were significant.

Still, even if the effect of a 250 microsecond stimulus on the brain isn’t huge, I find it pretty impressive that the 500 microsecond stimuli can evoke such marked neural responses. The brain is able to detect a flash of black and white that occupies just one two thousandth part of one second.

This doesn’t mean, of course, that we could process a message that only appeared for such a brief time, but still, it’s rather cool.

ResearchBlogging.orgSperdin HF, Spierer L, Becker R, Michel CM, & Landis T (2014). Submillisecond unmasked subliminal visual stimuli evoke electrical brain responses. Human Brain Mapping PMID: 25487054

CATEGORIZED UNDER: EEG, papers, select, Top Posts
  • Jan Moren

    I rather suspect the minimum flash time is pretty much a function of stimulus brightness (relative contrast, really). The actual time doesn’t matter as long as it pumps out enough energy to reliably affect the light-sensitive compounds in the retinal cells.

    As a thought experiment, take a very high-intensity laser and shine a pulse into the eye for just a few nanoseconds. If the intensity is just right to affect the retinal pigments in the same way as a moderate light over a period of milliseconds, you’d see the same kind of response. And you could sort-of argue that we can detect events in the nanosecond range.

    This is why you use predefined gratings, not single edges, for determining spatial resolution.

    • Neuroskeptic

      That’s a very interesting idea!

    • temblor4

      Have you tested something like this in the lab?

      • Jan Moren

        I’m not an experimentalist, so no. It shouldn’t be too difficult to test, I think, though you would need some specialized and perhaps purpose-built equipment in order to generate the stimuli in a well-controlled (and safe) manner.

        • David Palmer

          This would be easy to do with an Arduino or other micro controller.

          You don’t need a laser, if you are looking right at an LED in a dark room it should be visible in a flash that’s much shorter than 250 microseconds.

          My guess is that the eye-brain system would be completely incapable of differentiating between a 250 microsecond pulse of light, a 500 microsecond pulse at half the brightness, or a couple of 125 microsecond pulses a few milliseconds apart (assuming that these all put the same number of photons into the eye at a few milliseconds timescale). It would not be hard to test.

    • Michael Kovari

      I agree.

  • Charles Barnard

    The brain may be a limiting factor, but the actual question is in regards to the ability of the mind to assimilate data…which depends greatly upon things like the familiarity of the data presented. With optical data, the amount and type of data preceding and following the flash will affect the time that the mind actually spends on the data.

    Training in observational techniques can affect this sort of processing, as will many conditions of the body itself. Retention is also dependent upon frequency of presentation–not merely time per se.

  • Pingback: Friday 1/6/2015- Neuro OnAir (draft) | Neuroscience Hub()

  • Pingback: How Fast Is The Brain? | Extrano's Alley, a gun blog()

  • Felonious Grammar

    Assuming that we didn’t blink in time to miss it. It’s both creepy and cool that our brain does so much work subliminally.

  • David Daniels

    all well and good, but did anyone consider that it’s not the reaction that takes the most time? That it is in fact the things the brain does directly after the initialization of a response that would accurately and purely tell us how long the brain takes to, for e.g. determine danger, initiate movements that are part of a defensive response and also how long it takes other parts of the brain to correlate an appropriated response, and how accurate on average the response is. These would be better gauging values than simply: the brain responds to this kind of stimuli this quickly, all in all, much more data needs pulled from more instances and more direct testing of specific capabilities before we just come out and say: this is this.

    • S.Garland

      This is an excellent point. Some reactions are purely chemical responses and do not involve any complex cognition. Studies have been done on neural ganglia that are stimulated in various ways and react. A lot of our autonomic responses work this way.

  • Pingback: Morsels For The Mind – 16/01/2015 › Six Incredible Things Before Breakfast()

  • Pingback: Neuroscience | Pearltrees()

  • Pingback: subliminal image recognition – Magf()

  • robert easterbrook

    What the research is talking about is synaptic firing – it all depends whether the stimulus was registered by a neuron. EEG tech is still not sensitive enough to detect a single synaptic firing, but only large numbers firing in a particular region. True? What if the person imagined they saw the light and a synapse fired? Was it the external light they responded to or was it the imagined light?



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar