In 2024, a team at Tohoku University in Japan arranged nine wooden blocks on soil in the shape of either a cross or a circle, placed a fungus called Phanerochaete velutina in the center, and watched what happened. The fungus grew outward from the center, consumed the resources in the central blocks, and then—here’s the part that got the headlines—selectively extended toward the outer blocks, distinguishing between inward and outward directions. The researchers described this as a form of spatial pattern recognition. Microbial ecologist Yu Fukasawa said, with the kind of understatement that makes a mycologist sound like a neuroscientist at a cocktail party: “You’d be surprised at just how much fungi are capable of. They have memories, they learn, and they can make decisions.”
A 2025 study by Yin et al. expanded this further, demonstrating context-dependent food preferences in the slime mold Physarella oblonga—an organism that can evaluate multiple food sources, adjust its preferences based on context, and violate basic principles of rational choice theory in the same ways humans do. A July 2025 paper on Physarum polycephalum showed that slime mold memory isn’t just a reflex—it’s overwritable in light of new information, which ticks the box for a widely accepted criterion of navigational memory.
Meanwhile, the most popular story about fungal intelligence—the “wood wide web,” the idea that forests are connected by a cooperative underground network of mycorrhizal fungi through which trees share nutrients, send warnings, and nurture their offspring—has been substantially dismantled by the very scientists who helped build it. The state of fungal cognition research in 2026 is that the organisms nobody thought were interesting are turning out to be fascinating, and the narrative everybody loved is turning out to be mostly wrong.
What the wood wide web actually claimed
The term “wood wide web” was coined in a 1997 Nature article. The concept, popularized extensively by University of British Columbia forest ecologist Suzanne Simard and German forester Peter Wohlleben’s bestselling The Hidden Life of Trees, goes like this: mycorrhizal fungi form a symbiosis with tree roots, trading soil nutrients for photosynthetic sugars. As fungal filaments spread through forest soil, they physically connect the roots of neighboring trees into “common mycorrhizal networks,” or CMNs. Through these networks, proponents argued, trees transfer carbon and nutrients to each other—sometimes across species—with mature “mother trees” preferentially sending resources and defense signals to their offspring. The forest, in this telling, is less a Darwinian battleground of competing organisms and more a cooperative commune in which trees communicate, share, and care for each other through an underground fungal internet.
The story was irresistible. It combined genuine science with a narrative that resonated with environmental values, appeared in Avatar and The Last of Us, sold millions of books, and fundamentally changed how a generation of nature enthusiasts understood forests. Simard’s 2021 memoir Finding the Mother Tree became a New York Times bestseller. The wood wide web became one of the most successful science communication stories of the 21st century.
What the evidence actually shows
In February 2023, a paper in Nature Ecology & Evolution by Justine Karst, Melanie Jones, and Jason Hoeksema—three mycorrhizal ecologists with decades of combined field experience—examined the evidence behind the wood wide web’s central claims and found them “largely disconnected from evidence.”
The paper evaluated three claims. First, that common mycorrhizal networks are widespread in forests. The researchers concluded that with current technology, it’s difficult to confirm that continuous, non-transient fungal connections between trees actually persist in the field. Mycorrhizal fungi are extraordinarily delicate—dig up a root to study the connection and you’ve destroyed it. DNA sequencing of fungal networks had been achieved in only five field studies, on a limited range of fungi and tree species. The networks may exist, but their prevalence and permanence have not been established.
Second, that resources are transferred through these networks in ways that boost seedling growth. The researchers found that in the best-controlled experiments, fewer than 20 percent showed that fungus-connected seedlings performed better than disconnected ones. In the remaining 80 percent, connected seedlings performed the same or worse. Alternative explanations—nutrients moving through soil pores, direct root-to-root transfer—could account for the observed results without invoking the network at all. A critical nuance: even when tagged carbon from one tree showed up in a neighbor, much of it stayed in the mycorrhizal roots themselves rather than being transferred to the recipient tree. The fungi were receiving the carbon. Whether they were passing it along, and whether the amounts mattered ecologically, remained undemonstrated.
Third, that mature trees preferentially send resources and defense signals to their offspring through CMNs. The researchers stated flatly: “The claim that mature trees preferentially send resources and defence signals to offspring through CMNs has no peer-reviewed, published evidence.” Zero field studies support it.
The paper also documented a structural problem in the scientific literature itself. The researchers reviewed 1,676 citations of original CMN field studies and found that among papers published in 2022, fewer than half the statements made about the original studies were accurate. A 2009 study that mapped fungal distribution was routinely cited as evidence of nutrient transfer—even though it never investigated nutrient transfer. Alternative hypotheses provided by original authors were consistently omitted in subsequent citations. The wood wide web had become, in the researchers’ words, a scientific game of telephone.
Karst described the reception of their paper as “a bit of a relief” within the mycology community. Jones, one of the three authors, noted: “I’d like to see more old-growth forests protected but this ‘wood wide web’ distorts the evidence.”
What fungi actually do (which is strange enough)
The irony of the wood wide web correction is that the real science of fungal behavior is arguably more interesting than the debunked narrative—it just doesn’t fit as neatly into a story about cooperative forests.
Mycorrhizal symbiosis is real and ecologically essential. Fungi grow inside and on tree roots, forming relationships that are fundamental to tree nutrition. The fungi access soil nutrients—particularly phosphorus and nitrogen—that roots can’t reach on their own, and in return receive photosynthetic sugars from the tree. This mutualism has existed for over 400 million years and underlies the normal growth of virtually all land plants. It is not in dispute.
What’s in dispute is whether the relationship is primarily cooperative or primarily transactional—and whether fungi have their own agenda. The framing of the wood wide web cast fungi as benevolent infrastructure, passively shuttling resources between trees for the forest’s collective benefit. The alternative, which the evidence increasingly supports, is that fungi are active agents pursuing their own nutritional interests. When carbon moves from tree to fungus, the fungus may keep most of it. When a fungal network connects two trees, it may be exploiting both of them rather than facilitating communication between them. Some mycorrhizal relationships are parasitic—certain orchids and understory herbs use CMNs not to cooperate but to steal sugars from connected trees. The network isn’t necessarily a commune. It might be a marketplace, or a protection racket, or something with no human analogy at all.
The cognition research is where things get genuinely weird. Physarum polycephalum—a slime mold that is technically not a fungus but occupies a similar ecological and conceptual niche—can solve mazes, construct transport networks with efficiency comparable to human-engineered systems (it famously replicated the Tokyo subway), make cost-benefit trade-offs, habituate to harmless stimuli, form spatial memories stored in its extracellular slime trails, and override those memories when new information makes them obsolete. It does all of this without a single neuron, using rhythmic contractions of its protoplasm to propagate chemical signals across a network of tubules. A 2024 study showed that actual fungi—not just slime molds—can recognize spatial patterns in their resource environment and adjust their growth strategy accordingly.
Whether any of this constitutes “thinking” depends entirely on your definition. If thinking requires neurons, fungi don’t think. If thinking means adaptive information processing that integrates sensory input, memory, and decision-making to produce flexible behavior—then fungi do something that, functionally, is difficult to distinguish from thinking, and they do it using mechanisms that predate the evolution of nervous systems by hundreds of millions of years.
The Underground Atlas
In 2025, the Society for the Protection of Underground Networks—SPUN—released the first high-resolution predictive biodiversity map of Earth’s mycorrhizal fungal communities, using over 2.8 billion fungal DNA sequences sampled from 130 countries. The map, called the Underground Atlas, represents the most comprehensive picture of below-ground fungal diversity ever assembled. Among its findings: 83 percent of Earth’s climate-critical fungi remain unknown to science, identified only by DNA sequences with no corresponding described species. The underground world, it turns out, is vastly more complex and vastly less understood than even the most enthusiastic mycologist suspected—and the parts we don’t know about may be more important for carbon cycling and ecosystem function than the parts we do.
The real story of mycelial networks in 2026 isn’t cooperative trees whispering through the soil. It’s a kingdom of organisms that process information without brains, make decisions without neurons, form networks whose structure and function we’re only beginning to map, and play roles in global carbon and nutrient cycling that we can’t yet quantify because we haven’t identified most of the species involved. The wood wide web was a beautiful story. The truth is stranger and, honestly, better.
We cover fungal cognition, mycelial network research, and the neuroscience of organisms without nervous systems across our Neurozoology course—including why the most interesting question in cognitive science might not be “how does the brain think?” but “what was thinking before brains existed?”