the mind's inner eye's inner mind

wired UK #2.04, 04.96

It's a particularly bizarre feature of hallucinations that whilst on the one hand they are supposed to give access to your innermost thoughts and desires, on the other they always seem to be fairly generic; hippy culture depends for its very existence upon people being able to bond over the "shared" experience of getting high. Ever since Freud, the interpretationists have dominated the field, at least as far as popular culture is concerned, with their various recyclings of the basic idea that a hallucination has meaning by dint of the fact that it gives expression either to your unconscious or to some profound philosophical truth (such as the idea that we're all really descended from Venusian Mushroom Men.) The fact is however that most psychoactive experience does not produce surreal Daliesque dreamscapes but various combinations of geometric figures. Noting this fact, Dr. Mario Markus of the Max Planck Institute in Dortmund recently conducted a n investigation the findings of which suggest that the reason that certain types of hallucinatory experience are very common is not due to any shared cultural or personal experience but rather to the nature of the architecture of our brains. His claim is that because our brains are made up of densely interconnected networks - related to, if not identical with,"neural net" computer architecture - certain conditions can cause the neurons to fire "in sync." When this happens in the visual cortex - the area of the brain thought to be largely responsible for the processing of sight - we "see" particular patterns mentioned, notably spirals, tunnels, and concentric circles.

Dr. Markus argues that the perennial popularity of these figures in the art and imagination of almost every epoch and culture is down to the fact that they are caused by the simplest possible examples of ordered firing in the visual cortex. They occur when neurons arrange themselves to fire in stripes.

So how come mere stripes can produce all these different patterns? It has been well-established for some time that the cells in the retina project themselves onto the brain in such a way that the retina is reproduced spatially in the visual cortex (polar coordinates (radius and angle) in the cortex correspond to cartesian coordinates (x and y) in the cortex). If you stimulate the appropriate area of the visual cortex of a conscious person with an electrical probe, s/he will "see" spots of light. Dr. Markus has taken this fact and linked it with two other ideas. The first of these is the theory the brain operates in a largely similar manner to the parallel computers commonly known as "neural nets." The second is a theory originally worked out by code breaker, computer genius and all-round good bloke Alan Turing on the back of an envelope in the 1950's - that there is a general mechanism for the way spatially-ordered patterns form in systems as diverse as those responsible for pigmentation (a zebra's stripes for example, or the markings of a clown fish), plant structure (e.g. the way leaves order themselves along a stem), or the regular segmentations in mammaillian embryos.

Turing's mechanism hinged on the behaviour of a phenomenon known as an auto-catalytic reaction. This is any chemical reaction which produces more of its own "fuel" as a by-product, thus continually spurring itself onwards. In other words, it is a chemical reaction with a built in positive feedback loop (a microphone picking up output from its speaker forms a positive feedback loop, resulting in the howling noise so beloved of Jimi Hendrix and The Who.) Turing hypothesised that under certain conditions, a plethora of these reactions acting in parallel (as occurs in many biological systems) would spontaneously generate a coherent and stable spatial order.

A necessary condition for this is that the auto-catalytic substance (or "fuel") produces its own inhibitor (or "poison") and that this inhibitor diffuses faster than the autocatalytic reactand. This results in so-called "lateral inhibition," meaning that no autocatalysis can occur around an auto-catalytic reaction, thus clearly defining the pattern in space. For example, we have spaces between the fingers on our hands because at some stage of development the formation of fingers in those spaces was inhibited.

Markus reasoned that if these pattern forming mechanisms existed in chemical systems, then they should also exist in neural networks. Neural nets work by distributing inhibiting and activiting potentials across their synapses, and since the patterns produced by Turing's auto-catalytic reactions are effectively patterns of inhibition and excitation it makes sense that they should be able to exist on networks. And if they can exist in neural nets - and if we subscribe to the opinion that a neural net is a fairly good, if simplified model of a biological brain - then they should, by extension, be able to exist in the brain itself.

But surely it's impossible to test whether or not Turing patterns are occuring in the brain? Unless of course that pattern were to exist in the visual cortex, in which case - thanks to the convenient mapping of the retina onto the brain - we would be able to see it. Because of this mapping, a Turing pattern in your visual cortex corresponds to a significantly deformed pattern in the retina. It is this deformed pattern that the person interprets as a "vision." In the case of a simple, striped Turing pattern, the final image depends on the orientation of the stripes. Horizontal cortical stripes would produce a visual pattern of concentric circles. Vertical stripes in the cortex would appear as lines converging to the centre. Diagonal stripes, on the other hand, would give rise to images of spirals. A more complex but still common type of Turing pattern, known as "labyrinthine" because of its maze-like appearance, would look very much like a tunnel receding into the distance.

To test his hypothesis, Markus investigated sketches made by artist Gerald Oster - sketches he made of the hallucinations he experienced under the influence of LSD. Markus then digitised the images, fed them into his computer, and applied his transformation algorithms to them in order to work out how these visions looked when mapped out according to the topography of the visual cortex. Pleasingly, the spirals and circles were found to correspond to exactly the simple striped Turing patterns that Markus had predicted.

Electricity can produce visual hallucination as readily as drugs or sensory deprivation. At the turn of the century, when electricity was still regarded as something of a novelty, a popular form of party game involved a group of people joining hands in a circle and receiving a shock from a high-voltage generator. Participants would see what were known as "high voltage circles" dancing before their eyes. And Max Knoll, co-builder of the first electron microscope, used to enjoy applying low-voltage square-wave pulses to his temples, wrapping the electrodes in felt and soaking them in a salt solution to increase their conductivity. He reported seeing patterns whose shape depended upon the frequency of the pulses, a finding which supports Markus's suggestion that they are caused by regularities in the firing patterns of the neurons.

Markus goes as far as suggesting that the tunnels to the underworld experienced by shamen during the performance of drumming rituals or the long tunnels with a light at the end seen by many people who have "near-death experiences" are both examples of hallucinatory form constants, the appropriate pattern being setup in the visual cortex by the resonances from the drums in the first instance and by a lack of oxygen to the brain in the second. "The dramatic visionary episodes involving spiralling tunnels leading to the 'beyond' can be reduced to a dry numerical calculation of a physico-chemical process," he maintains. "For myself this does not influence my views about the 'beyond,' but it seems that one should be sceptical about world-like descriptions of the non world; the transition to death is probably more indescribable than just a tunnel with a light at its end. Note in this connection, that in some religions (e.g. the Jewish or the Mohammedan) concrete pictorial imaging of the transcendental is forbidden."

(This is the original version of the article that appeared in the magazine).