What do people see when they’re tripping?: 5 Questions for neuroscientist Sean Noah

Noah discusses what analyzing Erowid's trip reports can tell us about psychedelics and human vision.

The existence of synesthesia blew Sean Noah’s mind the first time he learned about it in high school biology class. He found it fascinating that, for synesthetes, sounds could generate images, and that some drugs, including psychedelics, could elicit that effect in people who don’t usually experience synesthesia. What did that mean about how our brains process the senses? This question led Noah to study psychology and neuroscience as an undergrad at UC Berkeley. After receiving a doctorate in psychology from University of California at Davis, where he studied the relationship between attention and visual perception.

Noah recently returned to UC Berkeley for a postdoctoral fellowship, where he’s studying psychedelics to better understand how the drugs affect visual perception. He began a research project with the website Erowid, a repository of some 60,000 pages of information about psychoactive plants and chemicals. Together with the website’s founders, Fire and Earth Erowid, Noah is analyzing trip reports posted to Erowid’s site to evaluate the range of visual effects psychedelics can produce. The Microdose spoke with Noah about what analyzing these reports can tell us about psychedelics and about human vision.

How did this collaboration with Fire and Earth Erowid come about, and what’s it like to work with them?

Erowid has a collection of publicly available experience reports — they’ve been collecting them since the 1990s, and their site allows anyone in the world to submit a self-report of their experience with a psychoactive substance. The result is this amazing database with reports about tons of different substances, including obscure ones that haven’t been rigorously studied by scientists, and for which there aren’t good sources of information elsewhere.

The Erowids state on their website that researchers cannot “mine” data from their site but that they’re open to discussing projects with researchers, provided they’re properly credited and cited. So we reached out to them to ask if we could work with their dataset, and explained our project: we planned to use a large language model to essentially read through the trip reports and describe and analyze the visual effects described in them, which is a novel approach researchers haven’t tried before.

They were willing to work with us, so we were able to download all the reports — about 40,000 of them — and we collaborated with them in writing the research paper manuscript as well. And they’ve been great collaborators; I think they once described themselves as “squares among the weirdos, and weirdos among the squares.” They’re eccentric but professional, diligent, and dedicated to this project; they have a lot of great ideas, and we’re fortunate that they’re sharing their expertise and their database with us. I grew up reading Erowid online in the early days of the internet, when it was one of the most interesting websites to visit, so it’s been really fun to work with them.

What’s the ultimate goal of this model?

We just published a first paper where we looked at the distribution of different types of common visual effects that we defined in advance based on our reading of the literature — for instance, seeing colors changing, surfaces like walls “breathing,” visual trails, or geometric patterns. But what motivated this study is that we got the sense that standard inventories, like the Altered States of Consciousness survey that’s commonly used by researchers, are really only scratching the surface of the realm of effects psychedelics can cause - and ultimately, defining the categories in advance, of course, is a source of bias. Rather than letting the data speak for itself, we’re defining what to look at rather than letting the rich dataset speak for itself.

So ultimately, what we want to do is to identify visual effects in a bottom-up way rather than top-down way. We’re developing that method now, and what’s cool about it is that the model will be adaptable to other domains, too — so we could modify it to identify the emotions people describe in trip reports, or cognitive effects.

I think of this study as a form of psychedelic natural history, where we have to adopt the attitude of a naturalist — we’re doing careful observation and study of organisms and their processes in their natural environment, their behaviors, and their interactions with one another. The experiences our minds can produce are diverse, beautiful, and ephemeral.

In that first study, did you observe any differences between visual effects generated by different drugs?

We determined the percentage of sentences in each drug report that could be classified as a description of visual effects. Less than 5% of trip report sentences are about visual effects, so part of the challenge is filtering them out from the rest of the report so we can study them.

Unsurprisingly, psychedelics, by far, had the highest percentage of sentences describing visual effects; opioids had the lowest. We also looked at different types of psychedelics and didn’t see any systematic differences there.

If only around 5% of reports are visual description, what’s in the other 95%? And is there anything about those reports that surprised you?

Erowid’s submission guidelines are open-ended, so you get a huge diversity in how people tell their stories. Erowid encourages people to include information about their motivation, the context in which they’re taking the drug, and as much information about dosage, timing, and onset of different effects, but aside from that, people write whatever they want. There’s often a lot of narrative information; some people submit a very literary description of their experience.

I was surprised by how some of the visual effects I would expect to see more weren’t common. Like, you often hear stories about encounters with entities, especially when taking DMT — there’s a common trop about encountering elves, but I didn’t see many reports that described that. And then there were other effects I saw more often than I expected — people seeing symbols or letters, and machinery. I hadn’t heard of those motifs before but they showed up again and again in the data.

What could this work tell us about vision and cognition?

If we had a more useful tool for identifying the elemental or specific visual effects in individuals, we could use that tool to study how the brain produces visual perception in the first place. We know psychedelics are powerful tools for altering consciousness and perception but don't yet have a comprehensive understanding of all their effects, either — like how changes in brain activity are associated with their effects, or even the full realm of possible visual effects that occur. Once we have more of a handle on the taxonomy of effects — which is what this study is trying to do — we can trace them back to individual changes in brain activity. And that’s actually another study we’re starting right now; we’ll be administering psilocybin to volunteer participants, putting them in an fMRI scanner, and giving them a visual illusion to see how psychedelics affect their brain’s processing of the visual stimulus. The stimulus is basically an optical illusion, something you can perceive in two different ways — like if you remember the viral phenomenon of The Dress, where some people see it as blue and black while others see it as white and gold. Our task isn’t The Dress, but it’s a similar concept, where people can perceive it in different ways, and the scanner will allow us to understand how psilocybin might change that perception.’

This interview has been edited and condensed for clarity and length.