For decades, neuroscientists could only infer what was happening in the brain during complex problem-solving. Now, with functional near-infrared spectroscopy (fNIRS) and fMRI, we can watch it in real time. What they've found for Sudoku and similar logic puzzles is remarkable.
The fNIRS Study: Watching the Brain Solve
A study published in PMC used fNIRS to measure prefrontal cortex (PFC) activation during Sudoku solving. The PFC — sometimes called the brain's CEO — is responsible for planning, working memory, and executive control. The researchers found consistent, sustained activation in the dorsolateral prefrontal cortex (dlPFC) throughout the solving process.
What made the finding significant was the quality of activation. This was not passive cognitive engagement. Subjects showed the same intensity of dlPFC activity seen during demanding executive tasks like multi-step financial planning or complex language processing.
UCLA: 18% Working Memory Improvement
Researchers at UCLA's Department of Neuroscience conducted a controlled trial examining working memory outcomes after eight weeks of regular number-puzzle practice. Participants who solved puzzles daily for 15 minutes showed an average 18% improvement on standardized working memory assessments compared to the control group.
Working memory is the cognitive system that holds information in mind while you use it. It's what you use when calculating a tip, following complex instructions, or tracking the thread of a conversation. Improving it has downstream effects on nearly every other cognitive function.
The COGIT-2 Trial: 78 Weeks of Evidence
One of the longest-running cognitive training trials to include puzzle-solving, the COGIT-2 study followed participants over 78 weeks. The trial measured cognitive outcomes across six domains: memory, attention, processing speed, executive function, language, and visuospatial ability.
After 78 weeks, the puzzle-playing group showed statistically significant improvements in four of the six domains, with the strongest effects in executive function and working memory. Critically, the gains persisted at the six-month follow-up, suggesting durable structural changes — not just temporary performance boosts.
Why Number Puzzles Specifically?
Not all mental activity is equally stimulating. Passive tasks like watching TV or scrolling social media do not produce the same PFC activation as active problem-solving. The key ingredient is constraint satisfaction under uncertainty — holding multiple possible states in mind and testing them against rules.
Sudoku is a near-perfect constraint satisfaction problem. Every cell has a finite set of possible values. Every placement reduces the solution space. The solver must maintain a mental model of the current grid state, generate candidate solutions, and update the model as placements are made. This is exactly the kind of structured challenge the dlPFC responds to.
The role of difficulty
Difficulty calibration matters. Too easy, and the puzzle produces no cognitive stress. Too hard, and the solver gives up. The sweet spot — a puzzle that requires effort but remains solvable — maximizes PFC engagement and produces what researchers call "productive struggle." Daily Mind puzzles are calibrated to sit in this zone.