Problem-Solving

Last reviewed 5 October 2026

What it is

"Problem-solving" names at least three different things, and the research on psychedelics depends on which one is meant.

The first is a process. A person notices a problem, defines it, comes up with options, chooses one, carries it out and checks the result. The leading model of this process, from social problem-solving theory, puts a fifth element in front of those steps: a person's general stance toward problems, whether they tend to approach them or avoid them. A structured therapy built on the model, problem-solving therapy, trains the stance and the four skills together [24]. The questionnaire most used to measure the process, the Social Problem-Solving Inventory-Revised (SPSI-R), is a commercial product. The validation study read for this page tested 499 adult men held in prison for sexual offences. It supported the five-part structure, and it also found that the subscales tracked a measure of socially desirable responding [23]. So the process model is well established as theory, while the search behind this page turned up no check of its questionnaire in ordinary healthy adults.

The second meaning is task performance, and it is the one almost every psychedelic study actually measures. Convergent thinking means finding the single correct answer. In the Picture Concept Task a person looks for the link between pictures. In compound Remote Associates problems a person gets three words and must find a fourth that forms a compound word or phrase with each of them [40]. Functional fixedness tasks ask whether someone can use a familiar object in an unfamiliar way. Divergent thinking, listing as many uses for an object as possible, is a different skill that belongs more to creativity research than to problem-solving.

The third meaning is cognitive flexibility, the ability to drop a rule that has stopped working and adopt a new one. It is usually measured with set-shifting tasks that count how often a person keeps applying the old rule [16]. Flexibility is one ingredient of problem-solving. It is not the whole of it.

One more distinction runs through everything below. Researchers sort solutions into analytic and insight. An analytic solution is reached step by step, and the solver can say how. An insight solution arrives suddenly, often after being stuck, as a reinterpretation of the problem, and it comes with an "Aha" feeling [20] [21]. The same puzzle can be solved either way, so studies usually ask the solver which it felt like. In four experiments with students on time-limited laboratory puzzles, solutions that people judged to be insights were on average more accurate than their analytic ones [22]. The split matters here because the popular claim about psychedelics, that they help someone see a problem differently, is really a claim about insight. None of the drug trials described below sorted solutions this way.

Why would an organization that studies natural psychedelics for personal growth care about this dimension? Because the claim is common and the measured evidence is thinner than the stories. Saying so plainly is part of the job.

What is known

The short version. In controlled studies of healthy adults, no psychedelic has been shown to improve convergent or analytic problem-solving. At higher doses, convergent thinking went down while the drug was active, and in the one placebo-controlled psilocybin trial it was still lower than placebo a week later. Repeated very low doses, known as microdosing, produced no reliable change on problem-solving tasks once a placebo group was included. Gains show up mainly in uncontrolled studies, which cannot separate the drug from expectation and practice. The best-supported positive signal in the field, better cognitive flexibility, comes mainly from clinical samples, such as patients with depression. It has not been shown in healthy adults, and it has not been shown to mean better problem-solving.

Where the claim began

The claim has a source. In 1966, Harman, McKim, Mogar, Fadiman and Stolaroff published "Psychedelic Agents in Creative Problem-Solving: A Pilot Study" [1]. Twenty-seven professionally employed people, described as men although the test data include eight women, each had a single session in one of seven small groups, after extensive selection and preparation. The agent in the reported sessions was mescaline, a natural psychedelic found in the peyote cactus. LSD-25, a semi-synthetic compound, appears in the abstract and in earlier trial sessions. The authors wrote that the session was structured on purpose to establish the participants' expectations and a setting friendly to creative work. They called their findings tentative, drawn from creativity tests, self-reports, and the utility of the solutions. Their reading was that the agents seemed to help creative problem-solving, especially in the "illumination" phase, the stage in which the answer comes into view, and that the benefit might last some weeks. The paper has no control group by design and reports no test-retest reliability data, and its authors did not try to separate the drug from expectation and preparation, so the study cannot say how much came from the drug. This is a historical pilot study, graded 1 here as a primary report, and it is best read as the hypothesis that later work set out to test. A recent psilocybin paper describes the early studies from the 1950s and 1960s as largely inconclusive and limited by the methods of their time [4].

Controlled trials at higher doses

Three controlled studies measured convergent thinking in healthy adults under the influence of a psychedelic. The compounds differ, and only one of them, psilocybin, is a natural psychedelic in the usual sense. The DMT and harmine combination below is a laboratory formulation inspired by ayahuasca.

In a randomized crossover trial, 24 healthy volunteers received LSD or an inactive placebo, double-blind, and did a battery of creativity tasks near the drug's peak. Novelty, originality and symbolic thinking rose. Utility and convergent thinking fell, and elaboration fell slightly. The authors described a "pattern break" with less organization [2].

In a within-subject trial, 30 healthy men took part in three sessions with a DMT and harmine combination, harmine alone, and placebo. The combination significantly impaired convergent thinking, most of all in men who started with stronger reasoning scores. Divergent thinking showed no overall effect. In a real painting task, both the combination and harmine alone reduced transitions related to an "incubation" stage (stepping back from the problem), and the combination also reduced moves from incubation to "illumination". So not every process effect can be put down to DMT. Participants' feelings of insight predicted their divergent scores and not their convergent ones [3].

The only placebo-controlled psilocybin trial found for this page is a double-blind parallel study in Maastricht of 60 healthy volunteers, 30 per arm, all with some previous psychedelic experience. Psilocybin lowered convergent thinking during the session. Seven days later it was still lower than in the placebo group. The authors urge caution about that second result. Scores in the placebo group rose with repeated testing, which looks like a practice effect. Psilocybin blocked that rise during the session, and at seven days its group had risen by less, so part of the gap may be practice in one arm. Two other findings point in different directions. Volunteers on psilocybin rated themselves, on a three-item scale, as having more insights during the session, yet they produced fewer ideas on the tasks. And a week later they generated more ideas that were new to them on an uses-of-an-object task, a borderline result (p = 0.05, with no correction for the number of comparisons), though not a difference in convergent problem-solving. The authors concluded that psilocybin did not enhance creative thinking as such. It changed particular parts of it at particular times [4].

A field study of ayahuasca (26 people at two spiritual workshops, no placebo, everyone already experienced with the tea) found that convergent thinking fell after drinking. The divergent result was thinner. One measure, originality relative to the number of ideas, rose on one of two tasks, while fluency and originality did not change overall and the other divergent task showed no effect [14].

Microdosing

Microdosing means taking a small amount of a psychedelic repeatedly, below the level that produces a full psychedelic experience. It is often said to sharpen focus and problem-solving. The controlled evidence does not back that up.

Two double-blind, placebo-controlled trials of psilocybin truffles in semi-naturalistic settings tested cognitive control, memory, social cognition and well-being, the areas where microdosing is said to sharpen focus [5]. The abstract does not mention problem-solving tasks, so the result bears on the claim only indirectly. Compared with placebo, microdosing did not significantly affect any behavioural or subjective measure. Early signals in social cognition, mood and self-reported cognitive flexibility did not survive correction for multiple comparisons, a statistical step that guards against chance findings when many tests are run. Participants in the active condition reported more negative bodily feelings, and they stayed effectively blinded throughout. Only the abstract of this paper was available.

A New Zealand trial gave 80 healthy men LSD or placebo every third day for six weeks. It found no drug-by-time effect on the Remote Associates Task, the uses task, an expert-rated art task, or an everyday problem-solving questionnaire. That questionnaire is a home-made, unvalidated self-report in which people nominate a problem from their own life and rate how hard it was to think of solutions, how useful they were and how satisfying. Reported difficulty fell in both groups, which the authors read as a possible placebo or participation effect. The tasks were done about four hours after dosing, not at peak effect. Participants felt more creative on dosing days, and before the trial they expected creativity to be the domain microdosing would help most [6]. A laboratory study of single very low LSD doses in 20 healthy young adults, all with earlier psychedelic experience, likewise left cognitive task scores largely unaffected, apart from a borderline rise in word-association items attempted [7]. A preregistered crossover trial of psilocybin microdosing, which looked at emotion rather than problem-solving, found no effect on its confirmatory measures, and participants broke the blind in its second block [10]. A preprint pooling three double-blind, placebo-controlled microdosing trials (171 people, not peer reviewed) found no change in convergent thinking. One divergent measure, originality relative to the number of ideas, did rise. The authors call the findings preliminary [41]. In a self-blinded citizen-science study of 191 people, using substances the participants sourced themselves and that were never verified, the microdose group improved on all the psychological measures and the placebo group on some of them, and the small acute differences, creativity among them, could be explained by participants guessing which they had taken. The blind was partly broken, with 72 percent of guesses right against 63 percent expected by chance [8]. A systematic review of 24 microdosing studies, 3,681 people in all, including observational studies, concluded in narrative form that randomized trials showed minimal impact on cognition or creativity. Its pooled statistics covered depression, anxiety and stress rather than creativity, and the randomized part rested on two trials [9].

The positive findings come from designs without a placebo. At a microdosing event, 38 volunteers did tasks before and after a non-blinded dose of truffles. Twenty-seven of them had usable Picture Concept Task data. Convergent and divergent performance improved, and fluid intelligence did not change. The authors called this preliminary and asked for placebo-controlled tests. They judged a practice effect unlikely [11]. At a psilocybin retreat, convergent thinking was unchanged the morning after and higher a week later, while idea fluency and originality rose the morning after and were back at baseline by the week. The sample shrank from 55 people to 22 and there was no comparison group [12]. The placebo-controlled trial above shows why such results are hard to trust. Its placebo group improved on the same kind of task, so repeating a task tends to raise scores on its own [4].

The cognitive flexibility signal

The one consistent positive thread concerns cognitive flexibility, and it needs a careful reading.

In an open-label study at Johns Hopkins, 24 adults with major depressive disorder received psilocybin with extensive psychotherapy, and 22 completed the set-shifting task. Their perseverative errors (continuing to apply a rule that no longer works) on a set-shifting task fell and stayed lower at four weeks. A delayed-treatment group of ten people showed no such change before treatment, which argues against simple practice, though only with moderate statistical support. The gain did not track the improvement in depression, and larger gains in a brain-imaging measure of neural flexibility went with smaller gains on the task. Tasks of response inhibition, selective attention and abstract reasoning did not change. With no placebo arm, expectation and the therapy itself cannot be ruled out [16].

A systematic review of 67 studies, 47 in people and 20 in animals, reported that psilocybin had the most consistent evidence of lasting gains on cognitive tests, especially flexibility and attention, with no consistent changes in the days after use for LSD or DMT. The authors add two cautions. Improvements appeared more often in psychiatric than in healthy samples, which may reflect repair of an existing deficit more than a rise above normal. And the included studies often reported outcomes selectively or in too little detail [17]. That is why the finding does not overturn the conclusion above. A tool that restores flexibility in people with depression is not shown to sharpen problem-solving in people who are well. A study of 48 ayahuasca drinkers who completed a flexibility scale and a card-sorting task before and within 24 hours after a ceremony saw improvement, but it had no comparison group and covers a single day [13]. In the microdosing trials, self-reported flexibility did not hold up after correction [5].

One survey, handled with care

A 2016 survey of 68 people is the lone naturalistic result that points the right way. Eleven participants reported classic psychedelic use together with a mystical experience, and they solved a functional-fixedness puzzle faster than the others, a large difference (Cohen's d = -0.87, an effect size in which about 0.8 counts as large, with the minus sign reflecting shorter solving times). But lifetime mystical experience was unrelated to solving time (standardized beta = -0.06, a coefficient where values near zero mean no relationship), so it did not explain the difference. The result is a retrospective association in eleven people, with no cause established [15].

Insight: what is theory and what has been tested

Two papers argue from theory. One proposes that insight links ordinary problem-solving, meditation, psychotherapy and the effects of psychedelics [18]. The other proposes that breakthroughs, from small insights to deep personal change, share a mechanism of representational change, meaning that a person's mental model of a problem restructures, with three ingredients: tension, altered salience and enhanced flexibility [19]. Neither reports new data on problem-solving performance, and the authors of the second note that little evidence supports psychedelics improving performance on problem-solving or creative tasks [19].

The brain science of insight is a separate field. In EEG work with word puzzles, it links the tendency to solve problems insightfully, rather than analytically, to different patterns of resting brain activity. The finding is correlational [20]. A PubMed search run on 2 October 2026 found no controlled psychedelic study that sorted solutions into insight and analytic. So the question that matches the popular claim best has not been tested. A drug that lowers step-by-step convergent scores could in principle leave insight untouched, raise it, or lower it. That is a hypothesis with no data behind it yet.

What is not known

Most of what matters most is open.

No study found for this page follows a healthy adult through a real, hard problem in the weeks after a session. Every result above is a laboratory task done during or soon after the drug. How far such tasks predict anything outside the laboratory is itself uncertain. A 2008 meta-analysis examined that question for divergent-thinking tests and creative achievement, which is creativity rather than problem-solving, but only its bibliographic record could be checked here, so no figure from it is quoted [38].

The insight and analytic split has not been applied. Nor has anyone shown whether the flexibility gain seen in patients appears in healthy adults, or whether set-shifting is the right stand-in for how a person handles a stuck problem. The one trial that tested abstract reasoning in patients found no change [16].

Expectation is a large unknown. Creativity was the highest-rated benefit people expected from microdosing in one trial [6], and in the self-blinding study placebo and microdose matched each other [8]. A study of this dimension without a placebo and a record of expectations cannot be interpreted.

Samples are narrow. Several studies used only men [3] [6]. Others used small groups of people who already had psychedelic experience [4] [14]. The effects of a single session on the next day, month and year are almost entirely unmeasured for this dimension.

Measurement is unsettled. The process questionnaire most cited is commercial and was validated in a prison sample [23]. A published alternative, the Problem Solving Inventory [39], exists, but only its bibliographic record was checked for this page, and its terms of use have not been confirmed. Task-based measures are open, though their link to real life is uncertain.

What Metanoia seeks to learn. Metanoia's long-term plan begins with a validated measure of flourishing and a naturalistic observational cohort studied under shared reporting standards, with records kept under full legal authorization and without Metanoia supplying any substance. For this dimension, behavioural measures, such as whether a person took concrete steps on a problem they had been putting off, would be a natural thing to examine alongside laboratory tasks. Observational data of that kind could help with several questions the current literature leaves open:

1. Does anything change in how healthy adults handle real problems over months, compared with similar adults who had no such experience?
2. How much of any change survives once expectation is measured and adjusted for?
3. Do behavioural measures, such as concrete steps taken on a postponed problem, agree with laboratory tasks and questionnaires, and which is more trustworthy?
4. Does a measure of cognitive flexibility belong in such a measurement set, given that it is the one construct with a signal in patients?

None of this means Metanoia expects a particular answer. A null result would be as useful to publish as a positive one.

Compared with other paths

The comparison is unusual. Several ordinary routes have better direct evidence on insight tasks than psychedelics do. None has shown that it improves problem-solving in real life among healthy adults.

Sleep and rest. In a classic experiment, more than twice as many people found a hidden rule in a task after a night of sleep as after a comparable wake period [25]. A study from the Paris Brain Institute, with 103 healthy young adults, used a similar task. Those who reached the lightest stage of sleep for at least one 30-second scoring window during a rest period found the hidden rule far more often (83 percent of 24 people, against 30 percent of 49 who stayed awake), and the 14 who slept deeper showed no such advantage. Groups were defined by what happened during the rest, not by random assignment. The participants were selected for falling asleep easily and had been asked to sleep about 30 percent less the night before [26]. Both are laboratory experiments on one kind of task. Replication was not checked for this page.

Setting a problem aside. A meta-analysis of studies on incubation, which means leaving a problem and coming back, found a positive effect, larger for divergent tasks than for linguistic and visual insight problems. Easy, undemanding activity during the break helped more than rest for linguistic insight problems [27]. In one experiment, an undemanding task during the break improved performance on a uses task more than a demanding task, rest or no break [28].

Walking. A systematic review of 23 studies and 1,036 people, mostly students, pooled 20 of them for divergent thinking and found a large effect of walking (d = 0.93, 95% confidence interval 0.44 to 1.42, the range the data are compatible with). The authors note that many of the positive studies came from a single author. For convergent thinking, six studies of the same people walking and sitting gave d = 0.16 with a confidence interval of -0.31 to 0.63, which the authors rate as very uncertain evidence of a null effect. That interval includes a moderate benefit as well as no effect at all [29]. An earlier experiment, with 48 students and no balancing of order, found that walking raised uses-task scores for 81 percent of people and slightly lowered scores on a convergent word task, where only 23 percent improved [30].

Mindfulness training. In two studies with 157 people in all, trait mindfulness predicted better insight problem-solving and not better non-insight problem-solving. In the second study, a mindfulness training improved insight problems and left non-insight problems unchanged [31]. A large meta-analysis of 45 studies found psychological flexibility and mindfulness correlate (r = .44, with very uneven results between studies, I² = 92.6 percent), but a correlation says nothing about training or about solving problems [32].

Structured problem-solving training. Problem-solving therapy is the approach most closely tied to the process model. A meta-analysis of 21 samples found it as effective for depression as other psychosocial therapies and medication, and better than no treatment or attention controls [24]. That outcome is symptom relief in people seeking treatment, so it is not evidence about problem-solving skill in healthy adults.

Brain training. Practice on cognitive tasks improves the trained tasks, and shows less benefit on closely related ones and little on distantly related or everyday tasks [34]. A later critical review, by researchers who ran many of those meta-analyses themselves, argues that the overall far-transfer effect, meaning benefit on tasks unlike the trained ones, is null, a position other researchers dispute [35]. One meta-analysis in healthy middle-aged adults reported gains in executive function, but it drew on only eight articles, with a wide interval and widely varying results (I² = 80.5 percent), and it found no gain in attention or language [36].

Stimulant medicines. In healthy non-sleep-deprived adults given a single acute administration, methylphenidate gave small gains in recall, sustained attention and inhibitory control, d-amphetamine none, and modafinil a small overall effect. The inhibitory-control gain vanished when four studies were left out. None of the domains analysed was problem-solving as such, and the authors say user belief outruns the evidence [33].

Putting these side by side: for analytic and convergent tasks, psychedelics have so far lowered scores or left them flat, while sleep, incubation and mindfulness training have small experiments showing gains on insight tasks. Walking helps idea generation more than solving, and in one experiment it lowered convergent scores. Brain training and stimulants have little to offer here. For insight problems and for real-world problem handling in the weeks after a session, the effect of psychedelics is untested.

Safety signals relevant to this dimension

The findings that matter for safety in healthy adults are modest and mostly indirect. The best-measured effect is a drop in convergent performance while a full psychedelic state is under way, in research settings. No study read for this page tested decisions with real-world consequences during that state, so nothing here supports a conclusion either way about risk in daily life. In the microdosing trials, negative bodily feelings were more common in the active condition [5], though a review that pooled two randomized trials found adverse event risks similar between treatment and placebo [9].

Reactions vary widely. In an analysis that pooled raw data from 23 controlled laboratory studies at one site (409 psilocybin sessions, 261 healthy volunteers), unpleasant or anxious reactions were predicted by emotional excitability, younger age and a brain-imaging setting, and the amount taken was the strongest single predictor of the response overall [37]. Trials also screen people out. The depression study above excluded anyone with a personal or family history of psychotic or bipolar disorders [16]. Findings from such screened samples do not carry over to everyone.

History and tradition

The history that bears on this dimension is the research record itself, covered above. The 1966 pilot study [1] is the origin of the modern claim and is graded 1 as a primary report. No traditional or ethnographic source was used for this page, so none is graded 4.

The book on all twelve dimensions   Back to the twelve dimensions