19 April 2002Vibrating the Brain
by Art WinfreeThis column started in November with an unexpected observation about how we see directions and lines. That was partly for tactical convenience: the observation had provided my daily "GamesWorth" exercise on that day. And partly it made strategic sense: I have no professional background or experience in "psycho-physics," and the corresponding naiveté is exactly what is needed for this literary venture, in which reinventing the wheel is perfectly acceptable (just so it is done in an unfamiliar way.) The unexpected observation involved the Moon at a particular phase that recurs twice monthly. Since readers need to see the same thing and make what they will of it before any second report from me, I meanwhile filled in with mysteries unexpectedly arising in the course of thinking about stellar magnitudes and about moonlight, cued by Alan MacRobert’s primer on "The Stellar Magnitude System" in the same issue of E-Bulletin.
And so this column came to seem mis-titled: it seemed really about Very Amateur Astronomy. Not so. I want to get back on track now, providing you the advertised variety of project starters for Adventures in Discovery, on promising terrain reconnoitered by daily GamesWorths in 2001. Each might again sprout digressions, but at least they won't all be about the Moon and stars! Meteorites provided a gentle transition bringing us back to Earth. Now let's get back inside our heads to consider more psychophysics.
Evenings were brightening in January 2001, so I arrived home from work before my outdoor lights turn themselves on. They are screw-in fluorescents inside a 1-foot globe of white plastic, activated when a photoresistor inside the globe says it is getting dark out there. If I am a little late for dinner, I see them in the first 10 minutes or so of operation, when their own light still tends to turn them back off. The appearance of the globe is really strange. It seems covered with vaguely patterned purple blotches that shift and flicker, separated by yellow worms like electrical discharges that squirm around. They seem finer-grained nearer the center of my visual field. All are hard to describe, like convection patterns on the surface of the Sun or maybe more like fibrillation on heart muscle in that they change faster than you can identify them. How fast is that? I suppose no faster than the time scale of cerebral processing for awareness, something like 50-100 msec. And how big are these purple blotches? No identifiable size: I think they get finer more centrally. They look big from a distance, but small close up. Very roughly, they subtend about the same range of angular diameters from any distance. I'd say the foveal ones are in the range of 1/4 degree across (1/4 little-finger width at arm's length), which means about 1/8 mm on my fovea, 25 mm from the lens, and half that distance on my primary visual cortex, occipital area V1 at the back of the head. This 1/16 mm is getting close to the dimensions of cells (or of receptive fields?). You see what I am getting at: maybe these purple blotches are not on the lamp, but only in my brain, and maybe they reflect the spatial scale of propagation across functional units in the brain when tickled at high frequency.This presents a conundrum. How do you investigate, how do you communicate to others, and how do you describe without pictures, something that has no "objective" existence? Well, what does "objective" mean? I think it means that if others do the same thing they will get the same sensations. But I wish it could mean more than that. I would like it to mean that I can photograph the phenomenon and compare it quantitatively to similar instances. I can't. Can one do science without quantification? Well, let's try. This evidently starts with bringing others to observe the same, so this topic seem a natural for this public forum.
If anyone can be troubled to screw a fluorescent bulb into a light-controlled socket and put it behind a diffuser screen, I would be happy to know what you see when the ambient light is just low enough to barely turn on the fluorescent. You can fake it with a flashlight: in a dark room: shine the flashlight on the photodetector until the fluorescent turns off, then back off a little. It begins to turn back on, but immediately has trouble as its own light hits the photodetector and tends to turn it back off. You can magnify this feedback by placing a white reflector or even a mirror to bounce more fluorescent light into its controlling eye. This does not work with all incandescent lamps, I presume because of their thermal inertia: some bulbs just steadily fadeout as the flashlight brightens the eye, then switch off. But it does work with some. "Lifting the hood" to observe the bulb directly while flickering in response to feedback from a nearby mirror, I have the impression that the physical light source has uniform intensity: the mottled texture is invented by eye and brain.
I tried another thing. I brought home an oscilloscope, in the form of Velleman's PCO64i circuit board and WinDSO software that makes my Win98 laptop into a digital recording oscilloscope. To the circuit board's serial input I attach a 1.5 volt AA cell in series with a 2200 ohm resistor and a 12kohm CdS photoresistor, and send the voltage across either resistor to the oscilloscope. With the oscilloscope computing Fast Fourier Transforms to present a spectrum, here is what I see in sufficient darkness (left panel; if I remember right, the vertical scale is linear, not logarithmic, in signal power):
Make your screen display window wide enough to accomodate the two panels left and rightxxxxxxxxxxxxxxxxx
I expected fluorescent light pulses 120 times per second, plus harmonics of its non-sinusoidal waveform, further distorted through the CdS photoresistor. But we have instead multiples of 60 Hz, I don't know why. Anyway it looks to the eye like ordinary fluorescent light, only vaguely flickering in peripheral vision. Also notice on the left the unexpected 30 Hz peak lurking inconspicuously just above noise level...
Then with flashlight almost turning off the fluorescent, purple blotches begin flickering all over the white globe. The oscilloscope's FFT shows (right panel) what has changed about the light, integrated over the globe: Its frequency has gone through a classical period-doubling bifurcation to 30 Hz, that was maybe incipient in the left panel.
Below you see on the right the corresponding voltage(time) trace during half a second of "purple blotches light" with a peak every 17 msec. This represents the 60 Hz AC output from the lamp. But these flashes are alternately big and little, introducing a 30 Hz component. In contrast, on the left I back off with the flashlight again, so normal fluorescent glow resumes without (much) alternation
xxxxx
Make your screen display window wide enough to accomodate the two panels left and rightThe doubled-period flicker is far from sinusoidal. Being so angular, it has plenty of harmonics, the first adding to the 60 Hz peak above, then a new one at 90 Hz as you see in the FFT prior figure's right panel, then the old 120 Hz peak is augmented, then another new one at 150 Hz, and so on.
Conjecture: what brings out the purple blotches is not any patchiness of light on the globe, but simply its lower-frequency flicker at 30 Hz.
Test: get a strobe flashing xenon light inside the globe or on a white wall, and gradually dial down its frequency from 100 Hz or so.
Result: Nothing special happens until 30 Hz. Then in the range 30-18 Hz (33-55 msec cycles), purple blotches and yellow worms squirm all over the globe or wall! Their apparent size seems roughly proportional to my distance, as observed outdoors with the fluorescent-lamp globe.
Check: A nearby graduate student, unaware of what the professor is up to, has the same experience in the same range.If this represents some kind of interaction between adjacent regions of the brain when tickled at 30 Hz, and those units appear to be about 1/4 degree apart, so 1/16 mm on the cortex, then we are talking about some electrical disturbance propagating at 1/16 mm / 1/30 sec = 2 mm/sec, slower even than migraine scotomas, spreading depression, or Jacksonian epilepsy across the cortex. Maybe it happens elsewhere, e.g., in the retina or the lateral geniculate nucleus of the thalamus, where I don't know the distance scale.
Rummaging libraries, I did find one description of peculiar visual patterns induced by flicker in the range 4-25 Hz, by J.R. Smythies in 1959, reported in General Psychology 50, 305-324 under the title "The Stroboscopic Patterns".
Was there any decisive follow-up?
It may be noteworthy that this 30Hz is in the famous "gamma" band, broadly around 40Hz, thought to be essential for integration of conscious attention throughout the cortex. How might that idea be tested? I guess one would want to tickle the brain at 40 Hz and see something about attention and consciousness go wrong. Can 40 Hz even get into the CNS? One might expect not, if this band is reserved for internal synchronization to organize perceptions. There should be something like a "blood-brain barier" to exclude such interference from housekeeping routines. Are human senses in fact arranged, to keep it out interference? This might seem hardly necessary: natural selection of our genes would have made no such barriers unless high-frequency periodic stimuli were a common hazard during most of human evolution. In fact hearing does not seem protected: we sense tones up to at least 2000 Hz by sending 1:1 report of acoustic pulses into the brain. What about the sense of touch? Electromechanical massage vibrators, be they plugged into the AC line or operated from batteries, typically buzz at 60 Hz, or anyhow so I infer from watching as the strobe freezes their motion near 60 Hz and also at half that frequency (presumably illuminating every second motion). Is this an engineered optimum for something? Are such high frequencies actually getting into the CNS? If so, is this connected with the peculiar effect, comparable to epileptic seizure in some respects, for which vibrators are so cherished, viz. sexual orgasm? Or is the frequency unimportant and in any case not delivered into the CNS? Maybe the effect depends only on Pacinian corpuscles and Meissner bodies in the skin being most sensitive in that range (I think 50-400 Hz). I believe that touch receptors and hair cells of the inner ear do respond in a phase-locked way, individually perhaps not on every cycle, but collectively sending to the CNS a volley of action potentials at the intervals of the sensory vibration. Does it get into the CNS as such, or does something like frequency demultiplication strip it down to a mere place-coded indicator of tone or mechanical vibration? Another trail head for an Adventure in Discovery ...
Is our visual flicker fusion frequency below this gamma band? I confidently expected so, but anyway I took the trouble to measure it for myself. Surprise! Human flicker fusion cutoff is 20-60 Hz, the higher frequencies pertaining to brighter and more uniform light, especially in peripheral vision. Undergraduates in a lab exercise in my university typically report about 40 Hz cutoff for bright LEDs in central vision. I tried the following trick: I tinkered a feedback circuit using a BA728 dual operational amplifier chip to symmetrically drive two LEDs in alternation, both of them bright green:
You can build it using the techniques so clearly presented in Paul Dito's columns.
Spacing the bright green LEDs 3" apart on black paper, I drive my two eyes in alternation. Each eye separately (keeping the other eye shut or its LED covered) develops blobs, patterns, and waves in the 17-40 Hz range, then flicker fusion intervenes at about 41 Hz. Using both eyes and converging to click the two LEDs into one stereo image halfway between my eyes and the paper, I see exactly the same thing. So this fused LED is flickering as fast as 80 Hz, and presumably so is the lateral geniculate nucleus of my hypothalamus and so are my cortical vision areas by the time flicker fusion in each eye finally makes the green look steady. So it seems entirely possible to subject the brain to interference in the gamma band supposed crucial for conscious attention. And nothing peculiar happened (except for the 18-30 Hz wavy patterns.)
What does this mean? First of all, I suppose it means that the brain can keep up with faster flicker than the eye can, and that flicker fusion is a retinal, not cortical, phenomenon. I am assuming that the perception of flicker corresponds, not just to a report that "there is something flickering," but also to a signal at that frequency in the visual areas of the brain. If the injected stimulus ranges across the whole gamma band as I dial the LED frequency, yet it upsets no perception, then is it plausible that gamma band synchronization is crucial for perception? I thought not --- that was the point of trying this experiment --- but maybe a critic would argue that input along visual channels is strictly segregated, so this input interferes with nothing else. So what peculiarities do we observe in vision alone? Purple blobs and yellow patterns, but only well below gamma range, as noted above. Within the gamma range, nothing peculiar: maybe the gamma input merely accentuates attention to the diodes? I don't know what to make of it. Here is another trail head for an Adventure in Discovery.
But this is a trail you might not want to follow toward the low end of this interesting frequency band. It is peculiar that visual red/blue flicker at 12 Hz or so (24 frames per second) can induce epileptic seizures in about one person in 4000, e.g., in 600 Japanese children watching Pokemon on TV, 17 December 1997. There is even a report of one woman habitually exposing herself to such flicker (waving fingers of outstretched hand before the Sun) because it induces orgasm. Lesser effects in the 8-16 Hz band were explored in Tony Conrad's movie "Flicker" (1966).
You might be one of the susceptible individuals, so you would not want to play around with this at all, especially in a place where your difficulties would pass un-noticed.
But here is something you can do at no risk I know of. I have seen nothing about this in the published literature, but have observed it all my life, and I suppose that others might also:
When you first get up in the morning to stand before the brightness of daylight at the window, shift your gaze back to indoor darkness or a sheet of black velvety crepe. Do you observe a rapid flickering? Whatever could that be? Too fast to be the circular muscle of the iris overshooting around a new widened steady-state. Anyway, it is not uniform across the visual field, but looks richly structured, like complicated wave propagation. Does it happen independently in each eye (and connected visual cortex)? Mine does. What is the rate, roughly? You might be interested to compare your observations with mine in the next column.
Copyright 2002 by A. T. Winfree. All rights reserved. Used by permission.
