Reading Rewires Your Brain. Scrolling Unravels It.

There's a study from Carnegie Mellon where scientists put kids through 100 hours of reading instruction, then scanned their brains. New wiring had physically grown inside their language regions. The kids who didn't read? Zero change. Nothing.

That was 2009. It gets wilder.

In 2013, Emory researchers scanned students every morning for 19 straight days while they read one novel chapter each night. The morning after reading, areas responsible for understanding other people's emotions had formed new connections. So had the region that processes physical sensation. Their brains were simulating what the characters felt, as if it were happening to them. Those changes stuck around for five days after they finished the book.

Now flip to scrolling. A meta-analysis published in Psychological Bulletin pulled together 71 studies covering 98,299 people. Heavy short-form video use (TikTok, Reels, Shorts) showed a consistent pattern: worse attention, weaker self-control, more anxiety. Across teenagers and adults. Across every platform tested. Oxford didn't name "brain rot" its 2024 Word of the Year for nothing.

None of this is metaphor. It's measurable, scannable, structural change inside the human skull. And the research spans two decades, from DTI white-matter imaging of children to EEG recordings of TikTok users. The medium you feed your brain determines the brain you build.

Here's what the studies actually found.

100 hours of reading grew new brain wiring in children

The Carnegie Mellon study was led by Timothy A. Keller and Marcel Adam Just, published in Neuron (Vol. 64, No. 5, pp. 624–631). They enrolled 72 children aged 8 to 10 and split them into three groups: 35 poor readers who got 100 hours of intensive reading instruction over six months, 12 poor readers who got nothing, and 25 good readers as a second control.

Using Diffusion Tensor Imaging on a 3T scanner, they measured fractional anisotropy (FA), basically a measure of how well-organized and structurally intact white matter is, before and after the intervention.

The results weren't subtle. Before remediation, poor readers had significantly lower FA in the left anterior centrum semiovale, a white matter region carrying fibers between the frontal, parietal, and temporal cortex. After 100 hours of instruction, the treated children's FA had increased to normal levels. The structural deficit was gone. Both control groups showed zero change over the same period.

What happened at a cellular level: FA increased because radial diffusivity decreased (water movement perpendicular to nerve fibers slowed), with no change in axial diffusivity. That pattern is the signature of increased myelination: oligodendrocyte cells building thicker myelin sheaths around axons. Myelinated fibers transmit signals roughly ten times faster than unmyelinated ones. Marcel Just described it as increasing "communication bandwidth" between connected brain regions by an order of magnitude.

The amount of structural rewiring correlated directly with how much reading ability improved. More white matter change, more reading improvement. And the absence of change in both control groups ruled out natural maturation. This was the reading instruction doing it.

I keep coming back to how clean this result is. Two separate control groups, zero change in both, dose-dependent response in the treated group. Thomas R. Insel, then-director of NIMH, called it "an exciting new approach" showing behavioral treatment could alter brain connectivity.

Reading a novel rewired brains for days after the book was closed

Gregory Berns and colleagues at Emory University designed something clever. They took 19 undergraduates (ages 19–27, from an initial 21), gave them daily resting-state fMRI scans every morning for 19 consecutive days, and structured the experiment in three phases. Five days of baseline scans with no reading. Nine days where participants read roughly a ninth of Robert Harris's Pompeii each evening (verified by quizzes). Then five "washout" days with no reading.

The critical bit: participants were not reading during the scans. They were just lying at rest. Yet two categories of change showed up.

First, short-term connectivity increases appeared in the left angular and supramarginal gyri and the right posterior temporal gyri, regions tied to perspective-taking and Theory of Mind (understanding other people's beliefs and emotions). These tracked with reading nights and faded after participants finished the novel.

Second, and this is the part that stuck with me: long-term connectivity changes emerged in the bilateral somatosensory cortex, the region that processes physical sensation and body movement. These changes persisted for the entire five-day washout period. Berns interpreted this as evidence for "embodied semantics," the idea that reading about physical experiences activates the same circuits you'd use to actually have those experiences. He called the lingering traces "shadow activity," describing them as almost like muscle memory.

So reading a thriller about a volcanic eruption left structural traces in the brain region that processes physical sensation. For days. After the book was finished.

Kidd and Castano's 2013 Science paper added another angle: reading literary fiction temporarily improved Theory of Mind performance across five experiments. (Fair warning on this one, replication has been mixed. Samur et al. 2018 and Panero et al. 2016 didn't replicate it, though Kidd and Castano's own 2019 preregistered replications supported the correlational finding.)

98,299 people and the cognitive cost of short-form video

The biggest assessment of what short-form video does to cognition came from Nguyen, Walters, Paul, and colleagues at Griffith University in Australia. Their systematic review and meta-analysis in Psychological Bulletin (Vol. 151, No. 9, pp. 1125–1146) synthesized 71 studies covering 98,299 participants across five continents. Platforms examined included TikTok/Douyin, Instagram Reels, YouTube Shorts, and Bilibili.

The effect sizes were consistent.

Inhibitory control (self-control): r = -0.41
Attention: r = -0.38
Overall cognition: r = -0.34
Stress: r = -0.34
Anxiety: r = -0.33
Overall mental health: r = -0.21

These patterns held across teenagers and adults, across every platform. One interesting moderating variable: studies using addiction scales (measuring compulsive use patterns) found stronger negative associations than studies measuring simple screen time, which suggests how people engage matters more than raw duration.

One surprise: body image and self-esteem showed no significant association, which goes against what broader social media research typically finds. The authors attributed this to the diverse content and creators on short-form platforms compared to, say, Instagram's curated-life aesthetic.

The theoretical framing here matters. The researchers used dual-process theory: repeated exposure to fast, highly stimulating content habituates users to slower, effortful tasks (reading, problem-solving), while algorithm-driven variable rewards push the brain toward impulsive, instant-gratification-seeking behavior.

Weakened brain waves in heavy scrollers' prefrontal cortex

If the meta-analysis shows the pattern, a 2024 EEG study from Zhejiang University shows the mechanism.

Tingting Yan and colleagues fitted 48 participants (ages 18–33; 45 retained for analysis) with 64-channel EEG caps and measured theta-band brain activity (4-8 Hz), an established marker for cognitive control, while participants performed an Attention Network Test.

Participants with higher short-form video addiction scores showed significantly weaker theta power in the frontal region during executive control tasks (r = −0.395, p = 0.007). This survived controls for gender, age, anxiety, and depression. Higher addiction also correlated with lower self-control scores (r = −0.320, p = 0.026).

The spatial specificity is worth noting. The negative correlation was strongest in frontal electrodes and weakened progressively moving backward, becoming nonsignificant at parietal sites. And resting-state theta power showed no correlation with addiction. The deficit only showed up when cognitive control was actively demanded. The brain looked normal at rest. It only faltered under load.

This parallels what addiction researchers find with substance use, where impaired prefrontal functioning relates to impulsivity and poor response inhibition.

Additional structural evidence from Gao et al. (2025, NeuroImage) used 3T MRI to show short-video addiction correlated with increased gray matter volume in the orbitofrontal cortex and heightened spontaneous activity in the dorsolateral prefrontal cortex. The reward system appears to be getting physically restructured by habitual scrolling. A functional near-infrared spectroscopy study (2025) further found that people addicted to short-form video showed higher risk-taking behavior and increased right orbitofrontal activation during risky decisions. The prefrontal system's ability to regulate behavior looks like it's being overridden.

So deep reading strengthens the prefrontal cortex. Heavy scrolling weakens it. The same brain regions, opposite directions.

Humans were never born to read

This is the part that reframes everything else, and it comes from Maryanne Wolf, Professor-in-Residence at UCLA's School of Education and Information Studies.

Her core insight, laid out in Proust and the Squid (2007): there is no reading gene. Written language is roughly 6,000 years old, far too recent for evolution to have built dedicated reading hardware. Instead, reading hijacked the brain's existing plasticity, repurposing neurons that were originally wired for other tasks.

This connects to Stanislas Dehaene's "neuronal recycling" hypothesis. Dehaene, at the Collège de France, showed that the Visual Word Form Area (VWFA) in the left fusiform gyrus, the region that specializes for recognizing written words in literate people, develops by repurposing neurons originally dedicated to face and object recognition. His 2010 Science paper confirmed this directly: as the left fusiform activates for reading, face processing shifts rightward toward the other hemisphere. Reading literally colonizes existing brain architecture. There's a neural trade-off.

What this repurposing builds, according to Wolf, is a reading circuit that ties together visual regions, language areas (phonology, semantics, syntax), cognitive systems (executive function, working memory, attention), and affective centers (empathy, perspective-taking, emotional processing).

But because this circuit isn't innate, it only survives through use. Stop reading, and it fades.

Wolf discovered this herself. In Reader, Come Home (2018), she describes trying to reread Hermann Hesse's The Glass Bead Game after years of heavy screen reading. She couldn't do it. It took her two full weeks of daily 20-minute reading sessions to recover her deep reading capacity. A neuroscientist who spent her career studying reading had to retrain herself to do it.

That scares me more than any of the other studies.

Wolf's concept of "cognitive patience," the sustained attentional capacity that deep reading builds, is what she thinks is really at stake. Her worry is specific: when the reading brain skims, it reduces time for deep processing. You lose time for grasping complexity, understanding other people's feelings, perceiving nuance, and forming your own original thoughts.

Her proposed solution isn't to burn all screens. It's what she calls a "biliterate reading brain", analogous to bilingualism. A brain that can deploy deep reading processes in print while also possessing digital fluency. She advocates teaching children deep reading first through physical books, then progressively introducing digital tools.

The principle she keeps returning to: the medium is the message to the cortex. If the dominant attributes of your media diet are speed, efficiency, and attention-switching, your reading brain will imperceptibly take on those characteristics and stop using the deeper ones.

The dopamine slot machine inside every feed

What connects all of this is the brain's reward system.

Short-form video platforms operate on variable ratio reinforcement, the same reward pattern that makes slot machines addictive. Each scroll delivers a small dopamine hit, but the unpredictability of encountering something highly engaging (a hilarious clip, a shocking reveal) creates intermittent reinforcement. This is the schedule most resistant to behavioral extinction. You keep pulling the lever because you never know when the next payout hits.

Sharpe and Spooner (2025) formalized this as "dopamine-scrolling," describing tolerance development where users need progressively more stimulation to get the same reward response. PET imaging studies have shown heavy social media users have lower dopamine synthesis capacity in the bilateral putamen, the same pattern seen in chronic stimulant drug users.

The cycle works like this: overstimulation leads to receptor downregulation, which creates a baseline state where normal activities (conversation, reading, walking) feel insufficiently rewarding. So you go back to the screen for relief.

Reading produces a different dopamine profile. Narrative anticipation, comprehension rewards, emotional identification with characters. It's a slow burn rather than a strobe light. The reward comes from sustained engagement, not rapid switching.

The structural data goes further. The CARDIA longitudinal study tracked 599 adults over 20 years and found that each additional hour of daily television was associated with roughly 0.5% reduction in total gray matter volume. The frontal cortex was particularly affected. Hutton et al. (2020, JAMA Pediatrics) reported that higher screen usage in preschoolers was associated with lower white-matter integrity and reduced early literacy skills.

On the other side, Li et al. (2024, Advanced Science) used the ABCD dataset of over 10,000 children with Mendelian randomization analysis to demonstrate that reading has a direct causal effect on brain volume in frontal and temporal regions. That's one of the strongest causal claims in this literature. And Huber et al. (2018, Nature Communications) showed an 8-week intensive reading intervention produced rapid, large-scale changes in core white matter pathways within weeks.

Oxford named it, and Thoreau saw it coming

On December 2, 2024, Oxford University Press announced "brain rot" as its Word of the Year, selected from over 37,000 public votes and expert analysis. They defined it as the supposed deterioration of mental or intellectual state from overconsumption of trivial content, particularly online.

Usage of the term had increased 230% between 2023 and 2024.

The term itself is 170 years old. Henry David Thoreau wrote in Walden (1854): "While England endeavours to cure the potato rot, will not any endeavour to cure the brain-rot?" But the modern version carries a specific digital weight. As Casper Grathwohl, President of Oxford Languages, noted, younger generations coined and spread the term through the very platforms said to cause the condition. That's both funny and slightly depressing.

The selection coincided with U.S. Surgeon General Vivek Murthy's call for warning labels on social media platforms in June 2024. The concern had moved out of neuroscience journals and into mainstream vocabulary.

What the science can't prove yet (and that matters)

I'd be doing this research a disservice if I didn't lay out the limitations honestly.

Most studies in the Nguyen et al. meta-analysis are cross-sectional. They capture one snapshot. They can't tell you whether short-form video causes cognitive deficits or whether people who already struggle with attention gravitate toward short videos as a coping mechanism. Both directions are plausible. The likely reality involves both feeding into each other.

Most studies rely on self-reported usage data. People are terrible at estimating how much time they spend on their phones.

The Yan et al. EEG study had only 48 participants (45 analyzed), predominantly female Chinese university students. That's a narrow sample. The Kidd and Castano fiction-and-empathy finding has had mixed replication, as I mentioned above. Oxford's definition of "brain rot" includes the word "supposed," which is a deliberate hedge.

That said, and I think this part is hard to shrug off, the Carnegie Mellon study had proper control groups with zero change. The CARDIA study tracked participants for 20 years. Li et al.'s Mendelian randomization analysis provides genuine causal evidence that reading affects brain volume. And the meta-analytic evidence spans 98,299 people across 71 studies on five continents, consistent across age groups and platforms.

When structural imaging, functional imaging, electrophysiology, behavioral measures, and longitudinal tracking all point in the same direction, the "it's just correlation" argument starts to feel like it's doing a lot of heavy lifting.

Two paths of neuroplasticity

Neuroplasticity is value-neutral. The brain will build whatever circuit you exercise most.

Feed it 100 hours of reading instruction, and oligodendrocytes wrap new myelin around language pathways, increasing signal speed tenfold. Feed it thousands of hours of 15-second clips, and the prefrontal cortex shows measurably weaker activation when cognitive control is demanded.

Wolf's framing is the one I find most useful: the medium is the message to the cortex. The reading brain took 6,000 years to build culturally and requires years to build individually in each child. It can atrophy in months of disuse. Wolf proved that on herself.

The short-form video research suggests a parallel process running in the opposite direction: variable-ratio dopamine reinforcement and rapid context-switching training the brain toward high novelty-seeking, low sustained attention, and weakened executive control.

The biliterate brain Wolf proposes, capable of deep sustained cognition when needed and digital fluency when appropriate, might be the most important educational goal of this generation. But it doesn't happen by accident. It requires deciding, deliberately, what you're going to feed your brain today.

Sources: Keller & Just, Neuron 2009; Berns et al., Brain Connectivity 2013; Nguyen et al., Psychological Bulletin 2024; Yan et al., Frontiers in Human Neuroscience 2024; Wolf, Proust and the Squid (2007) and Reader, Come Home (2018); Dehaene & Cohen, Neuron 2007; Li et al., Advanced Science 2024; Huber et al., Nature Communications 2018; Gao et al., NeuroImage 2025; Sharpe & Spooner, SAGE Journals 2025; Hutton et al., JAMA Pediatrics 2020.