Ask people what they saw and you get two kinds of answer. One kind is about shapes: lattices, tunnels, spirals, webs, patterns that breathe and repeat. The other kind is about things: figures, landscapes, faces, whole scenes with a story in them.
The science has a good account of the first kind. It has almost nothing to say about the second. That gap is the honest shape of this article.
Four families of shapes, described a century ago
In the 1920s the psychologist Heinrich Klüver studied the visual effects of mescaline, first in a journal paper in 1926 and then in a short book in 1928. He noticed that the simple imagery people reported kept falling into a small number of recurring geometric types, and he called them form constants.
Later authors summarise his classification as four groups. In the wording of a 2002 paper by Bressloff and colleagues, these are tunnels and funnels, spirals, lattices (they add honeycombs and triangles as examples), and cobwebs. Two details are worth keeping in mind. First, Metanoia has verified Klüver's publications bibliographically and has not yet read his original text, so the four-group list here comes from the later mathematical literature rather than from Klüver's own page. Second, two papers by the same later group list the lattice examples slightly differently, which is a small reminder that even careful summaries drift from the source.
Others had described these visions before him. Physicians and essayists in the 1890s had already written about mescaline imagery. What Klüver added was a classification, and that is a different kind of contribution. A description tells you what one person saw. A classification tells you what keeps coming back.
Why the shapes are probably the brain's own
Here is the finding that reorganises the whole question. The same four families of shapes turn up in conditions that have nothing to do with psychedelics.
A 2012 review in Psychological Bulletin by Billock and Tsou gathered the evidence: migraine aura, some forms of epilepsy, flickering light, magnetic stimulation and pressure on the eye all produce a shared set of elementary hallucinations, and most of them fall into Klüver's groups. Bressloff and colleagues list further triggers in their 2002 paper: bright flickering lights, the moments of falling asleep and waking up, and near-death experiences.
They also record a simple observation with a large consequence. The images are seen in both eyes and move with the eyes. From this they infer that the images are generated in the brain, not in the eye and not in the world. The observation is a finding. The inference from it is an interpretation, and a very widely shared one.
If the same geometry appears whether the trigger is a drug, a migraine or a strobe light, then the geometry is not a property of the drug. It is a property of the visual cortex, and the drug is one of several ways of revealing it.
A model that turns cortex into pattern
Why those shapes and not others? Since 1979 there has been a mathematical answer, and it should be introduced for what it is: a model, not a measurement.
Ermentrout and Cowan proposed that when the primary visual cortex becomes unstable, activity across it can organise spontaneously into stripes and hexagons, the way ripples organise on a vibrating plate. The visual field is mapped onto the cortex in a particular geometric way, so a stripe of activity on the cortex does not look like a stripe to the person. Run the mapping backwards and a set of parallel stripes in the cortex becomes a set of concentric rings or a spiral in what is seen.
Bressloff, Cowan and colleagues extended this in 2001 and 2002, using the known wiring of the cortex and its symmetries. Their result is that the patterns which emerge when the resting state of the visual cortex becomes unstable correspond, once translated into the visual field, to all of Klüver's form constants. They also suggest that the machinery producing these hallucinations is closely related to the machinery the cortex uses every day to process edges, contours and textures.
That is elegant and it is well published. It is still a hypothesis. We have not identified a study that watched a spiral form on a human cortex and matched it to the spiral the person reported. What the model explains is the geometry. What it does not explain is content. There is no spiral in the world that a spiral in the cortex is a picture of, and there is no obvious route from cortical stripes to a face, a jaguar or a city.
The same shapes, with no drug at all
One of the most useful facts in this area is also one of the least known. In a 2021 study of 24 volunteers, stroboscopic light flickering against closed eyelids produced vivid visual hallucinations of the simple type, strongest at one of the two frequencies tested, and participants rated their intensity as similar in strength to effects induced by psychedelic substances. The study used standard validated questionnaires, and a personality trait called absorption correlated strongly with how much people experienced.
A follow-up imaging study in 2023 by the same group found that flicker increases connectivity between the thalamus and visual cortex in a way that tracks the intensity of the visual experience. That is a single study, and we have not identified a replication.
For Metanoia this matters beyond curiosity. It means there is a legal, drug-free way to study the elementary visual phenomenon in healthy adults with validated tools, and it is one of the research directions we have on our list.
What has been measured in the brain
Now to the drugs themselves. Three findings carry most of the weight.
Blocking the receptor blocks the visuals. In a double-blind study published in 2013, Kometer and colleagues gave healthy volunteers psilocybin and recorded the brain's electrical activity. Psilocybin strongly reduced a resting rhythm over the back of the head called alpha, and strongly reduced an early visual response called N170, and the reduction of N170 went together with the appearance of visual alterations including hallucinations. In the 15 volunteers analysed, each person's N170 reduction correlated with their rating of complex imagery (scenes and pictures) and of sound-driven visual changes. For elementary imagery (patterns, colours, light flashes) the correlation was in the same direction but did not reach statistical significance. When the volunteers were pretreated with ketanserin, a drug that blocks the serotonin 2A receptor, all of these effects were blocked. This is the closest thing the field has to a causal chain: receptor, brain signal, experience, and the whole chain removed by blocking the first link. It establishes the receptor as necessary. It does not establish it as sufficient, and ketanserin is not perfectly selective, which the authors and later reviewers acknowledge.
The visual changes are specific to the visuals. In 2016, Carhart-Harris and colleagues imaged the brain under LSD with three methods at once. Increased blood flow in the visual cortex, decreased alpha power over the visual cortex, and a greatly expanded pattern of connections from the primary visual cortex all correlated strongly with how intense the visual hallucinations were. They did not correlate with the drug's other effects on consciousness, which tracked a different circuit. So it is not simply that the whole brain changes and everything is reported more intensely. The visual changes belong to the visuals.
With the eyes shut, the visual cortex behaves as if it were seeing. In 2012, de Araujo and colleagues asked experienced ayahuasca drinkers to imagine images with their eyes closed while in a scanner. Under ayahuasca, activation in the primary visual area reached a level comparable to looking at a real picture with the eyes open, and the effect correlated with the perceptual changes participants reported. Areas linked to memory and to internally generated imagination were also more active. This is a small, single study, and we have not identified a replication. The authors close their abstract with a sentence about why shamans might have selected the brew for visual revelation. That sentence is cultural interpretation, and we quote it here only to say that it is not a finding.
Together these say something specific. The visuals are produced by real activity in the parts of the brain that normally do seeing, they depend on one receptor, and their intensity tracks measurable changes in that visual machinery. None of this decides what the pictures mean or where their content comes from.
About the brain network everyone mentions
Popular accounts say psychedelics switch off, dissolve or quiet the brain's default mode network, and that this is the mechanism of the experience. The evidence does not support that as the central story.
In April 2026, a mega-analysis in Nature Medicine pooled 11 independent resting-state imaging datasets across five psychedelics from research groups on three continents, and processed them with one uniform pipeline. The core shared signature was an increase in connectivity between the brain's higher-order networks and its sensory networks, including the visual network. Reductions in connectivity within networks were weak to moderate and selective, and varied a great deal between drugs and sites. The accurate short version is that psychedelics increase integration between brain networks and reduce the separation between them. The dramatic version is not what the pooled data show.
Are the visuals different for different substances?
People are certain they are. The evidence is thinner than the certainty.
The direct, controlled comparison we have identified is a 2023 study by Ley, Holze, Liechti and colleagues in 32 healthy participants, in two sequential groups, who received mescaline, LSD and psilocybin on separate occasions in a double-blind crossover design. LSD and psilocybin were given to all 32 participants, while each of the two mescaline levels was given to 16, so the mescaline arm of the comparison rests on half the sample. We have not identified another. The subjective effects were comparable at the higher of the two mescaline levels tested, across a range of validated scales, and the authors found no evidence of qualitative differences in the altered states produced at that level. The one clear and consistent difference was duration: mescaline lasted longest, psilocybin shortest, LSD in between.
Two cautions. That study measured the overall altered state with general scales. It was not a study of visual experience specifically, but its published figure does break the scales down, including the visual ones, and the caption reports comparable effects on all subscales at the higher mescaline level. So the similarity extends to the visual subscales as those scales are built. The authors also note that their instruments may not have been sensitive enough to capture the complex phenomenology, and that subtle qualitative differences cannot be excluded. What the study does not tell us is whether the content of the visuals differed. It does raise a fair suspicion that much of the "personality" attributed to each substance is duration and dose wearing a disguise.
The second caution is about where the imaging comes from. Almost all of the brain imaging in this field is of psilocybin and LSD. We have not identified modern imaging of the visual effects of mescaline, of 5-MeO-DMT, of iboga or of Amanita muscaria, which acts on a completely different receptor system. Where the evidence is missing we say so rather than filling the gap from the psilocybin literature. A small number of laboratories in Zurich, Basel and London produce a large share of everything cited above, and we have not identified a quantification of what that concentration does to the picture.
We measure how strong, not what it depicted
Every quantitative study above measures the visuals with a rating scale. The oldest and most used instrument in the field was re-analysed in 2010 across 43 experimental studies, and its original scales turned out to be multidimensional, so the authors rebuilt it as eleven lower-order scales that behave consistently across substances and settings. That is good psychometrics.
It also exposes a structural gap. The rebuilt scales do separate elementary imagery, meaning patterns, colours and light flashes, from complex imagery, meaning scenes and pictures, and the eleven lower-order scales also include audio-visual synaesthesiae and changed meaning of percepts. So the person who saw lattices and the person who saw a landscape do score differently. What the scales rate is how intense each class of imagery was. We have not identified a validated instrument that characterises what the complex imagery depicted, as distinct from how strong it was, which is one reason the complex, narrative vision that people most want to talk about is the part the science has barely modelled. That gap can be worked on without any controlled substance, and it sits close to the centre of what Metanoia is building in its measurement work.
What the traditions say they see
Everything above is the view from the laboratory. The traditions that have used these plants and fungi for centuries have their own account, and it belongs in a different register: what a tradition says, recorded as what it says, without our ruling on whether it is true.
One item stands out because it comes from inside the tradition and not from an outside interpreter. In 1959 the mycologist Gastón Guzmán, showing freshly collected mushrooms to Nahuatl speakers in the Sierra Norte de Puebla in Mexico, recorded a living name for them. Analysing it in a 2012 paper, he concluded that the correct form is teotlacuilnanácatl, which he translates as "the sacred or divine mushroom that paints or describes through color figures". The root is the verb for painting, the same root as the word for the painters of the sacred codices. The name was given in response to three different Psilocybe species, not one.
Read that carefully for what it is. A tradition, in its own language, names the mushroom by what it does to sight: it paints, in figures and colour. That is a phenomenological claim from within the culture, and it is recorded here as such. Guzmán then connects the name to the neurotropic visual effects of the mushrooms. That connection is a mycologist's interpretation in 2012, and it is not something the tradition said. Keeping those two sentences apart is the whole discipline of this section.
The colonial record adds a second glimpse, from a hostile witness. The Franciscan Bernardino de Sahagún, compiling his account of Aztec life in the sixteenth century in order to help the Church identify and eradicate such practices, described merchants eating mushrooms at a feast and, the next morning, telling one another the visions they had seen. Everything in that record passed through two filters, what informants were willing to tell a hostile questioner and what the compiler chose to write down. It survives as evidence that the visions were shared, discussed and valued, not as a description of their content.
Several living traditions have visual art that is often said to depict or to derive from visions: textile patterns in the Peruvian Amazon, painted visions from the same region, yarn paintings in western Mexico, and the ritual space of the Bwiti in Gabon. Metanoia has not yet verified those links against primary sources, so we do not describe them here. We also flag a tempting shortcut, because it appears in a great many books: indigenous art contains geometric motifs that resemble Klüver's form constants, therefore the form constants are universal and the art records them. That is a claim that needs demonstrating, image by image and source by source. It is not a background fact.
Seen, or real?
Every reader arrives at this question and it deserves a straight answer about what the evidence can and cannot do.
The closed-eyes finding under ayahuasca shows that the visual cortex can be driven, without external input, to a level comparable to real seeing. That is why the experience feels like seeing rather than imagining. It says nothing about whether what is seen has an external source, in either direction. A vivid internal image and a perception of something outside would look the same in that measurement. Proposals at the edge of the literature, such as a 2017 suggestion that the phosphenes arise from amplified biophotons inside the eye and brain, exist and are cited here so that they are on the record. They have no direct experimental support.
Metanoia's position is to map the claims, state what would count as evidence for each, and not to rule. The people who hold that the visions disclose a hidden reality and the people who hold that they are the visual cortex talking to itself are describing the same brain measurements and drawing incompatible conclusions from them. The measurements do not adjudicate.
A word on visuals that do not stop
For a small number of people, visual disturbances persist after the substance has left the body, sometimes for a long time. This is a recognised clinical entity, hallucinogen persisting perception disorder (HPPD), which has its own diagnostic code in the ICD-10 coding used in health records. One EEG study compared 38 people with persisting visual disturbances after LSD use with 33 controls and found altered patterns of electrical coherence over the visual cortex with the eyes closed, which the authors themselves describe as suggestive rather than explanatory. A 2026 retrospective cohort study identified 25,778 people diagnosed with HPPD through electronic health records and characterised the health conditions that often accompany it. That study is the largest to date on HPPD, but as a cohort built from people already diagnosed, it does not by itself establish how common HPPD is in the wider population either. How common HPPD is, and who is at risk, is not well established, and we have not identified a study that settles either question. We include it because a page about what people see would be incomplete without saying that for some people the seeing does not end on schedule. Anyone concerned about persisting visual symptoms should talk to a doctor, our safe use page has more on getting help.
What we do not know
We have not read Klüver's own text, and the four-group classification attributed to him rests on later summaries until we do.
We have not identified a study that compared what the visual experiences of different psychedelics depicted, as distinct from how intense each class of imagery was.
We have not identified brain imaging of the visual effects of mescaline, 5-MeO-DMT, iboga or Amanita muscaria.
We have not identified a mechanistic model of complex, narrative visions. The mathematics covers the lattices. It does not cover the jaguar.
We have not identified a study measuring what healthy adults see as distinct from what clinical populations see, though most of the imaging work cited here was in fact done in healthy volunteers.
We have not identified any study that tested the chemistry, the geometry and the network dynamics together in the same participants, so the tidy three-layer story above remains a reasonable synthesis and not a demonstrated one.
Sources
-
1. Klüver H. Mescal visions and eidetic vision. Am J Psychol 1926;37(4):502–515.
JSTOR stable/1414910 · Level 1. Metadata verified against three independent records. Text not read by Metanoia, stated in the article.
-
2. Klüver H. Mescal: The 'Divine' Plant and its Psychological Effects. London: Kegan Paul, Trench, Trubner; 1928.
Level 1. Metadata verified. Text not read, stated in the article.
-
3. Bressloff PC, Cowan JD, Golubitsky M, Thomas PJ, Wiener MC. What geometric visual hallucinations tell us about the visual cortex. Neural Comput 2002;14(3):473–491.
PMID 11860679 · DOI · Level 1. Source of the four-group wording, the trigger list, and the "seen in both eyes and move with them" observation. Abstract read.
-
4. Bressloff PC, Cowan JD, Golubitsky M, Thomas PJ, Wiener MC. Geometric visual hallucinations, Euclidean symmetry and the functional architecture of striate cortex. Philos Trans R Soc Lond B 2001;356(1407):299–330.
PMID 11316482 · DOI · Level 1. The full mathematical model, presented as a model, not a measurement.
-
5. Ermentrout GB, Cowan JD. A mathematical theory of visual hallucination patterns. Biol Cybern 1979;34(3):137–150.
PMID 486593 · DOI · Level 1. Origin of the cortical pattern-formation account.
-
6. Billock VA, Tsou BH. Elementary visual hallucinations and their relationships to neural pattern-forming mechanisms. Psychol Bull 2012;138(4):744–774.
PMID 22448914 · DOI · Level 1. Cross-condition review: migraine, epilepsy, flicker, magnetic and mechanical triggers share Klüver's forms.
-
7. Bartossek MT, Kemmerer J, Schmidt TT. Altered states phenomena induced by visual flicker light stimulation. PLoS One 2021;16(7):e0253779.
PMID 34197510 · DOI · Level 1. n=24, within-subjects, validated instruments. Small sample.
-
8. Amaya IA, Schmidt ME, Bartossek MT, et al. Flicker light stimulation induces thalamocortical hyperconnectivity with LGN and higher-order thalamic nuclei. Imaging Neurosci 2023;1.
PMID 40799697 · DOI · Level 2. Single study, not yet replicated, stated in the article.
-
9. Kometer M, Schmidt A, Jäncke L, Vollenweider FX. Activation of serotonin 2A receptors underlies the psilocybin-induced effects on α oscillations, N170 visual-evoked potentials, and visual hallucinations. J Neurosci 2013;33(25):10544–10551.
PMID 23785166 · DOI · Level 1. Double-blind, within-subject, healthy volunteers (15 analysed), ketanserin pretreatment. Doses used in the study are deliberately not reported here.
-
10. Carhart-Harris RL, Muthukumaraswamy S, Roseman L, et al. Neural correlates of the LSD experience revealed by multimodal neuroimaging. Proc Natl Acad Sci USA 2016;113(17):4853–4858.
PMID 27071089 · DOI · Level 1. Visual cortex blood flow, alpha and V1 connectivity correlated with visual ratings and not with other effects.
-
11. de Araujo DB, Ribeiro S, Cecchi GA, et al. Seeing with the eyes shut: neural basis of enhanced imagery following ayahuasca ingestion. Hum Brain Mapp 2012;33(11):2550–2560 (online 2011).
PMID 21922603 · DOI · Level 2. Small single study, experienced users. The closing sentence of the abstract is cultural interpretation, labelled as such in the article.
-
12. Girn M, Doss MK, Roseman L, et al. An international mega-analysis of psychedelic drug effects on brain circuit function. Nat Med 2026;32(4):1543–1554.
PMID 41942645 · DOI · Level 1. Increased between-network connectivity is the core signature, within-network reductions weak to moderate and selective.
-
13. Ley L, Holze F, Arikci D, et al. Comparative acute effects of mescaline, lysergic acid diethylamide, and psilocybin in a randomized, double-blind, placebo-controlled cross-over study in healthy participants. Neuropsychopharmacology 2023;48(11):1659–1667.
PMID 37231080 · DOI · NCT04227756 · Level 1. n=32, in two sequential groups, 16 per mescaline level. Doses deliberately not reported here. The visual subscales are reported in Figure 2 of the full text.
-
14. Studerus E, Gamma A, Vollenweider FX. Psychometric evaluation of the altered states of consciousness rating scale (OAV). PLoS One 2010;5(8):e12412.
PMID 20824211 · DOI · Level 1. 43 studies pooled, eleven lower-order scales.
-
15. Guzmán G. New taxonomical and ethnomycological observations on Psilocybe s.s. from Mexico, Africa and Spain. Acta Bot Mex 2012;100:79–106.
SciELO · Level 1 for the paper, 4 for the tradition it records. Full text read. The translated name is quoted verbatim from the paper.
-
16. Sahagún B de. Historia general de las cosas de Nueva España (Florentine Codex), Book 9, sixteenth century.
Level 4. Colonial record by a hostile compiler, as stated in the article.
-
17. Kapócs G, Scholkmann F, Salari V, Császár N, Szőke H, Bókkon I. Possible role of biochemiluminescent photons for lysergic acid diethylamide (LSD)-induced phosphenes and visual hallucinations. Rev Neurosci 2017;28(1):77–86.
PMID 27732562 · DOI · Level 1 for the journal, the claim itself is a theoretical proposal without direct experimental evidence. Cited as an edge proposal only.
-
18. Abraham HD, Duffy FH. EEG coherence in post-LSD visual hallucinations. Psychiatry Res 2001;107(3):151–163.
PMID 11566431 · DOI · Level 2. 38 cases, 33 controls. Authors label their interpretation as speculative.
-
19. Butler et al. Characterising the clinical associations of hallucinogen persisting perception disorder: a retrospective cohort study. Transl Psychiatry 2026;16(1).
PMID 42031693 · DOI · Level 1. Largest cohort to date on HPPD, 25,778 people diagnosed via electronic health records, comorbidity analysis. A diagnosed cohort, not a population-prevalence study.
Bibliographic source for all identifiers above: PubMed, re-verified directly on 16 September 2026 (metadata for all, abstracts read for sources 3, 6, 7, 9, 10, 11, 12, 13, 17) and 23 September 2026 (source 19, and the SciELO address for source 15). The full texts of sources 9, 13 and 19 were consulted on 2 October 2026 for the statements on this page, and the scale list of source 14 was checked against its full text.