Sleep consolidates learning: French neuroscientist urges phone bans, phonemic literacy in schools

Stanislas Dehaene arrived in Porto Alegre last week with a simple argument: if we want to fix education in Brazil, we need to listen to what neuroscience has learned about how children's brains actually work. The French neuroscientist, who holds a chair at the Collège de France and directs the Cognitive Neuroimaging Unit at France's National Institute of Health and Medical Research, spent decades mapping the neural foundations of human learning. He came to speak at Brain Week, an event focused on mental health and neuroscience, and what he had to say was direct: education can improve dramatically if schools and parents take cognitive science seriously.

Dehaene's first target was the smartphone. In a press conference, he acknowledged that educational apps exist and can be valuable. But for most children, he said, a screen means TikTok—shallow content, endless scrolling, the replacement of real human interaction with the illusion of connection. The damage, he argued, happens in the earliest years, when children are supposed to be learning language and social skills through actual conversation with peers and teachers. Every hour spent on social media is an hour lost to the kind of face-to-face engagement that builds the neural architecture for communication. He called for phones to be banned from classrooms, not as punishment but as protection.

When asked about the 2024 Rio Grande do Sul floods, which kept thousands of children out of school for months, Dehaene's response pointed to something deeper than lost classroom time. The brain, he explained, learns according to principles that seem almost mechanical in their precision. One of the most important: learning must be distributed across days and weeks, with sleep in between. During sleep, the brain does not absorb new information. Instead, it consolidates what was learned during waking hours, deepening and strengthening the mental representations built through experience. Each night of sleep is another layer added to understanding. When a child stops attending school, they lose not just instruction but the rhythm that makes learning stick. There is a window for this kind of learning—neuroplasticity, the brain's ability to rewire itself, declines as children move into adolescence. Every day matters. Parents, he insisted, must understand that sending their children to school every single day is not a preference; it is the foundation of learning itself.

Dehaene also came to Brazil to advocate for a specific method of teaching reading. Called Kalulu—named after a figure from African folklore—the approach focuses on phonemes, teaching children the relationship between letters and sounds, and how those sounds combine into syllables and words. This is not how many Brazilian schools teach reading. Instead, they show children whole words and ask them to memorize. Decades of research, Dehaene said, have proven this does not work. The Kalulu method has been tested in France, Colombia, Argentina, and Brazil, in collaboration with researchers from the Federal University of ABC and the Collège de France. The results were clear: children using phoneme-based instruction learned to read significantly faster than those taught through traditional Brazilian methods. Porto Alegre, according to recent data from the Ministry of Education, has the lowest rate of age-appropriate literacy among Brazil's state capitals—children should be reading fluently by the end of second grade, and many are not. Dehaene suggested that changing how reading is taught could be part of the answer.

Underlying all of this was a observation about the infant brain itself. Through his research, Dehaene had discovered that even at birth, the human brain is far more structured than artificial intelligence systems. A newborn's brain already contains the neural connections present in an adult brain. It arrives in the world not as a blank slate but as an intricately organized machine, primed with basic concepts that make learning possible. Babies are born ready to acquire language, to recognize symbols, to build understanding from experience. This is fundamentally different from how AI learns—artificial systems require millions of examples to develop even basic competence, while human infants begin life already equipped for extraordinary learning. Understanding this, Dehaene suggested, changes how we should think about education. The brain is not a vessel to be filled. It is a living system with its own logic, its own needs, its own rhythm. Education works best when it aligns with that logic rather than fighting against it.