Problem-solving tasks span an enormous range, from a child stacking blocks to an engineer troubleshooting a bridge design, but they all share a core structure: you have a current state, a goal state, and no obvious path between them. What makes these tasks fascinating to researchers and useful in everyday life is that how well you perform on them depends less on raw intelligence than most people assume. Working memory, mood, sleep, the way a problem is framed, and even whether you’re working alone or in a group all shape your chances of finding a solution.
The Role of Working Memory
Working memory is the mental workspace where you hold and manipulate information while you’re actively thinking. It matters for problem solving because most tasks require you to juggle several pieces of information at once: the rules, the constraints, what you’ve already tried, and what options remain. Research shows that people with higher working-memory capacity solve more problems than those with lower capacity, across both straightforward algebra-style tasks and more complex, multiply-constrained puzzles that require juggling several conditions at once.1PubMed. The role of working memory capacity in analytic and multiply-constrained problem-solving in demanding situations
One of the key ways working memory helps is by supporting attentional control. When you’re working through a challenging problem, your attention wants to wander toward irrelevant details or previously tried dead ends. Higher working-memory capacity helps you resist that pull, stay focused on promising strategies, and efficiently search through the space of possible solutions.2Current Directions in Psychological Science. Working Memory Capacity, Attentional Focus, and Problem Solving This doesn’t mean people with lower working-memory capacity are stuck. The same research suggests they can compensate by offloading demands onto the environment, such as writing down intermediate steps, using diagrams, or physically arranging materials.3PubMed. The role of working memory capacity in analytic and multiply-constrained problem-solving in demanding situations
An interesting wrinkle: when problems got harder in one study, people with high working-memory capacity were actually more affected by the increase in difficulty than low-capacity participants. The likely reason is that high-capacity individuals rely heavily on internal strategies, so when a problem overwhelms those strategies, they feel the ceiling more sharply. People who already depend on external supports have less far to fall.
Insight Versus Analytic Solving
Not all problem solving feels the same from the inside. Sometimes you grind through a problem step by step, testing possibilities in sequence until one works. Other times the answer hits you all at once, the classic “aha” moment. These two modes, analytic and insight-based, actually look different in the brain. In a neuroimaging study, insight solutions were associated with a burst of gamma-band brain activity in the right anterior superior temporal gyrus, a region involved in making distant semantic connections. That burst appeared roughly a third of a second before participants pressed the button to report their answer. Non-insight solutions showed no comparable spike in that area.4PLOS Biology. Neural Activity When People Solve Verbal Problems with Insight
Even the brain state before seeing a problem can predict which mode a solver will use. EEG recordings taken in the two seconds before an anagram appeared on screen revealed that analytic solvers showed more beta-band activity over central-parietal regions, originating in the left mid-cingulate cortex and right postcentral gyrus. Insight solvers showed a different preparatory pattern.5PubMed. Pre-stimulus brain oscillations predict insight versus analytic problem-solving in an anagram task In practical terms, this means the conditions you set up before tackling a problem, whether you’re relaxed or focused, ruminating or mind-wandering, may nudge your brain toward one solving style or the other.
Mental Traps That Stall You
One of the most well-documented obstacles in problem solving is functional fixedness: the tendency to see an object only in terms of its conventional use, which prevents you from repurposing it creatively. The classic demonstration is Duncker’s candle problem from the 1940s, where participants are given a candle, a box of tacks, and a book of matches and asked to attach the candle to the wall. Most people initially fail to see the box as a potential shelf because they’re fixated on its role as a tack container.
Modern research shows that functional fixedness is especially powerful when you’ve recently seen or been reminded of an object’s standard function. Priming someone with the typical use of an object before asking them to generate creative alternatives measurably reduces how original their suggestions are.6PubMed. Neural basis of functional fixedness during creative idea generation: An EEG study Pictorial examples are particularly sticky. When people are shown visual examples during a creative task, they tend to reproduce elements of those examples even when doing so leads to worse solutions.7PubMed Central. Functional Fixedness in Creative Thinking Tasks Depends on Stimulus Modality
That said, functional fixedness may be less of a universal problem-solving villain than textbooks suggest. One study found that while misleading knowledge about an object’s function does create mental dead ends in simple puzzles, it plays a much smaller role in harder, more complex problems. And hands-on experience and good physical intuition acted as protective factors against getting stuck.8PubMed. Functional fixedness in tool use: Learning modality, limitations and individual differences So if you’ve spent a lot of time physically building or tinkering with things, you may be somewhat inoculated against fixedness.
How Mood and Stress Affect Your Approach
Your emotional state when you sit down with a problem matters more than you might expect, and the effects aren’t symmetrical. A mild positive mood consistently helps with creative and flexible problem solving. In a set of four experiments, people who had watched a short comedy clip or received a small bag of candy outperformed control groups on both the candle problem and the Remote Associates Test, a task where you have to find a word that links three seemingly unrelated words. Negative mood and physical exercise (used as an arousal control) did not produce the same benefits.9PubMed. Positive affect facilitates creative problem solving The effect extends to younger people too. Adolescents in a positive mood generated more category examples and more unusual examples, and solved the candle problem at higher rates than those in a neutral state.10Psychological Reports. Influence of Positive Affect upon Creative Thinking and Problem Solving in Children
Stress presents a more nuanced picture. In one neuroimaging study, participants solved compound remote associate problems under both stress and no-stress conditions, and the number of problems they got right was virtually identical across conditions, with accuracy averaging around nine out of thirteen either way. Solution speed also did not differ significantly.11NeuroImage. Effects of stress on functional connectivity during problem solving That finding might seem surprising if you’ve ever felt your mind go blank under pressure, but the key distinction is between the kind of stress and the kind of task. Acute lab-induced stress may not replicate the chronic, identity-threatening stress of a real exam or workplace crisis. The brain appears to reorganize its functional networks under stress without necessarily losing accuracy, at least for tasks that are practiced and moderately difficult.
Sleep, Incubation, and Stepping Away
The folk advice to “sleep on it” has real scientific legs. During sleep, the brain doesn’t just rest; it actively replays and reorganizes recently learned information. Different sleep stages contribute differently. Non-REM sleep appears to help the brain abstract rules and patterns from what it learned during the day, while REM sleep promotes novel associations between ideas that might not normally be connected. The repeated cycling between these stages across a full night is thought to build complex knowledge structures and then restructure them, which can lay the groundwork for creative breakthroughs the next morning.12PubMed Central. How Memory Replay in Sleep Boosts Creative Problem-Solving
A clever experiment tested this more directly. Participants attempted unsolved puzzles in the evening while hearing unique sound cues associated with each puzzle. Later, during deep sleep, the sounds linked to half of the unsolved puzzles were replayed. By morning, participants solved more of the sound-cued puzzles than the uncued ones, but only when the original puzzles had included misleading information that reinforced an incorrect approach. This suggests that sleep doesn’t just help you forget dead ends; it actively reprocesses the problem, helping you overcome the very fixation that was holding you back.13PubMed Central. Sleep and incubation: Using problem reactivation during sleep to study forgetting fixation and unconscious processing during sleep incubation
How Problem Solving Changes With Age
Problem-solving ability is not a single thing that peaks in your twenties and then declines. It’s a blend of at least two broad cognitive resources, and their relative contributions shift over a lifetime. In younger adults, the primary predictor of everyday problem-solving performance is fluid ability: the capacity for on-the-fly reasoning, pattern detection, and handling novelty. As people age, fluid ability gradually becomes less dominant and accumulated knowledge and experience, what psychologists call crystallized ability, takes over as the main driver.14PubMed Central. Cognitive Predictors of Everyday Problem Solving across the Lifespan
This shift has practical implications. Problems that emphasize concrete, real-world content and reward past experience tend to show little or no age-related decline in performance. Problems that are abstract, novel, and strip away the advantage of experience do show declines with age, and for those, fluid reasoning remains the stronger predictor regardless of how old the solver is.15Journal of Gerontology. Age Differences in Relationships Between Crystallized and Fluid Intelligences and Problem Solving In other words, a sixty-year-old mechanic troubleshooting a familiar engine may outperform a twenty-five-year-old novice, even though the younger person would score higher on a timed abstract reasoning test.
Experts Versus Novices
The expert-novice gap in problem solving goes beyond simply knowing more facts. Experts organize their knowledge differently: they store information in richly connected clusters that are indexed by the deep structure of a problem rather than its surface features. A medical expert seeing a set of symptoms doesn’t just have a bigger mental list; they have a framework that rapidly narrows the space of possibilities. They also deploy better strategies for accessing what they know, self-monitor more effectively, and are motivated differently during the solving process.16PubMed Central. Moving from Novice to Expertise and Its Implications for Instruction
One practical bridge between novice and expert thinking is self-explanation: deliberately explaining to yourself why a worked example proceeds the way it does, rather than just passively reading it. Students who practiced self-explanation while studying math word problems outperformed peers on both the practiced type of problem and on transfer problems they hadn’t seen before. Those who generated deeper explanations rather than surface-level restatements performed the best.17Japanese Psychological Research. Effects of self‐explanation as a metacognitive strategy for solving mathematical word problems
Transfer is one of the hardest things to achieve in problem solving. Most people struggle to recognize when a problem they’ve already solved is structurally similar to a new one dressed in different surface details. One effective technique is to have people create their own version of a problem in a different domain before attempting to solve a target problem. Doing so increases the likelihood of retrieving a distant, structurally similar example from memory, and the quality of the abstract schema the person builds during invention predicts how likely they are to spot the analogy.18PubMed. Promoting interdomain analogical transfer: When creating a problem helps to solve a problem
Collaborative Problem Solving
Many real-world problems are tackled by groups, not individuals, and group dynamics add a whole new layer to the process. A group’s collective intelligence, the general ability to perform well across a range of tasks, correlates moderately with group performance on complex assignments, with correlation coefficients in the range of .40 to .47.19PubMed Central. High-performing teams: Is collective intelligence the answer? But collective intelligence isn’t just the average IQ of the room. It depends on social sensitivity, conversational turn-taking, and the proportion of women in the group, according to earlier work in the field.
In real project environments, effective team problem solving depends on practical structural factors beyond raw brainpower. A case study of shipbuilding teams found that high-performing groups relied on situation awareness (knowing the current state of the project and its environment), task-based subgroups that could form and dissolve as needed, direct lines of communication rather than hierarchical chains, and trust among members.20Project Management Journal. Team Collective Intelligence in Dynamically Complex Projects—A Shipbuilding Case These factors suggest that team structure matters as much as team talent.
Well-Structured Versus Ill-Structured Problems
Researchers distinguish between well-structured and ill-structured problems, though the boundary is less clean than it sounds. A well-structured problem has a clearly defined initial state, a well-defined goal, and a known set of legal operations you can apply, like a Sudoku puzzle or a math equation. Ill-structured problems, by contrast, have ambiguous goals, unclear constraints, or multiple defensible solutions, like designing a public park or resolving a diplomatic crisis. Herbert Simon argued decades ago that the line between the two is vague, fluid, and cannot be neatly formalized.21Artificial Intelligence. The structure of ill structured problems
This matters because many laboratory problem-solving tasks are deliberately well-structured, which makes them easy to score but potentially poor stand-ins for real-world problem solving. Tower tasks are a good example. The Tower of Hanoi and Tower of London are two of the most widely used problem-solving tasks in neuropsychological testing, both assumed to measure planning and executive function. But when both were given to the same people, the correlation between performance on the two was only .37, which is low for tasks that are supposedly measuring the same thing. The likely source of the gap was poor test-retest reliability of the Tower of London.22PubMed. Towers of Hanoi and London: Reliability and Validity of Two Executive Function Tasks Research in children found the same pattern: the two tasks are not interchangeable, and performance on them at higher complexity levels seems to depend on mental flexibility rather than pure planning skill.23PubMed. A comparison of performance on the Towers of London and Hanoi in young children The Tower of Hanoi also shows a test-retest correlation of about .5 in children, which is below accepted psychometric standards for drawing conclusions about individual development.24Journal of Child Psychology and Psychiatry. Individual Differences in Cognitive Planning on the Tower of Hanoi Task: Neuropsychological Maturity or Measurement Error?
For anyone taking or administering these tests, the takeaway is that a single tower-task score doesn’t tell you very much about someone’s general problem-solving ability. The tasks tap into overlapping but distinct skills, and their reliability is shaky enough that a low score could easily reflect measurement noise rather than a genuine deficit.
The Body’s Role in Thinking
Problem solving feels like a purely mental activity, but the body plays a larger role than most people realize. Embodied cognition research shows that learning and reasoning improve when the brain, body, and environment interact, especially through hand gestures and physical manipulation of objects.25Educational Psychology Review. Research Avenues Supporting Embodied Cognition in Learning and Instruction This is part of why physical models, prototypes, and hands-on activities can sometimes break through mental impasses that purely verbal or visual reasoning cannot. When you physically rearrange puzzle pieces, sketch a diagram, or gesture while explaining a problem to someone, you’re not just expressing your thinking; you’re extending it. The motor system provides a kind of cognitive scaffolding that offloads some of the burden from working memory.
This also connects to the finding about functional fixedness mentioned earlier: people with stronger physical intuition and motor skills were less likely to get stuck on misleading function cues. If your experience with the physical world is rich, your brain has more routes into a problem beyond the purely conceptual.
Neurodiversity and Problem-Solving Profiles
People with ADHD and autistic individuals often approach problem-solving tasks differently, and understanding those differences is useful for educators, clinicians, and the individuals themselves. Research comparing adults with ADHD and adults with autism spectrum conditions on executive function tasks found distinct profiles. The ADHD group had difficulty withholding impulsive responses but showed relatively preserved planning and initiation abilities. The autism group showed the reverse: significant challenges with initiating action, forming strategies, and planning sequences, with better impulse control.26PubMed. Executive functioning differences between adults with attention deficit hyperactivity disorder and autistic spectrum disorder in initiation, planning and strategy formation
These profiles suggest that a one-size-fits-all approach to problem-solving instruction or assessment is likely to miss important variation. Someone with ADHD might benefit from externally imposed structure that prevents premature responses but allows their natural fluency in generating ideas. An autistic individual might benefit from explicit scaffolding at the start of a problem: clear prompts about where to begin and which strategies to try, followed by space for independent execution.
When AI Meets Problem-Solving Tasks
Large language models have gotten impressively good at many problem-solving benchmarks, but their weaknesses reveal something about what problem solving actually requires. A benchmark designed to test out-of-distribution reasoning, where problems required planning and explanation generation but were structured differently from training examples, found that humans were far more robust than AI models.27arXiv. Structured, flexible, and robust: benchmarking and improving large language models towards more human-like behavior in out-of-distribution reasoning tasks When the problems looked like what the models had been trained on, performance was strong. When the surface features changed while the underlying logic stayed the same, the models faltered. Humans handled the shift with relative ease.
This echoes the analogy-transfer findings in human cognition: recognizing deep structure through surface variation is one of the hardest and most valuable forms of problem solving. It’s also the skill most resistant to brute-force approaches, whether those approaches come from studying worked examples without understanding them or from an AI pattern-matching across billions of text samples.
Problem Solving in Other Species
Humans are not the only species that solves novel problems, and comparative research helps clarify what cognitive toolkit is actually necessary. New Caledonian crows have become a model species for studying physical problem solving. In a series of trap-tube experiments, three out of six crows initially learned to avoid a trap when pulling food through a tube. More impressively, those same crows continued to avoid the trap even when arbitrary visual cues they might have been using were removed, and they transferred their solution to a causally equivalent but visually different task, the trap-table. Their performance across multiple transfer conditions ruled out simple explanations like chance, associative learning, or visual generalization.28PubMed Central. Do new caledonian crows solve physical problems through causal reasoning?
Corvids and great apes share a cognitive toolkit that includes causal reasoning, behavioral flexibility, imagination, and prospection, despite having vastly different brain structures and being separated by hundreds of millions of years of evolution.29PubMed. The mentality of crows: convergent evolution of intelligence in corvids and apes The fact that complex problem-solving abilities evolved independently in birds and mammals suggests these capacities arise whenever the ecological pressures, particularly social complexity and variable food sources, are strong enough to reward flexible thinking.