Scientific studies on chess and the brain: the annotated bibliography

Dartigues et al., 2013 · BMJ Open · DOI 10.1136/bmjopen-2013-002998

Playing board games, cognitive decline and dementia: a French population-based cohort study

Who

3,675 people aged 65 and over (PAQUID cohort, Gironde and Dordogne)

Type

Observational cohort, 20-year follow-up · level of evidence: Moderate (association)

Result

Regular board-game players (32% of the sample) have a 15% lower risk of dementia (HR 0.85) after adjustment for age, sex, level of education and health status. The gap is no longer significant once baseline cognitive score and depression are taken into account (HR 0.96).

What the study does not say

Practice reported only once, with no frequency or duration; chess is not separated from cards, lotto or draughts; reverse causation is possible.

Verghese et al., 2003 · New England Journal of Medicine

Leisure activities and the risk of dementia in the elderly

Who

469 people over 75 living at home (Bronx Aging Study, New York)

Type

Observational cohort, median follow-up 5.1 years · level of evidence: Moderate (association)

Result

124 cases of dementia during follow-up. Reading, board games, playing a musical instrument and dancing are associated with a lower risk; each point on the cognitive activity score corresponds to a 7% lower risk (HR 0.93).

What the study does not say

Modest sample; activities self-reported; no figure specific to chess.

Wilson et al., 2002 · JAMA

Participation in cognitively stimulating activities and risk of incident Alzheimer disease

Who

801 elderly Catholic nuns, priests and brothers (Religious Orders Study, United States)

Type

Observational cohort, mean follow-up 4.5 years · level of evidence: Moderate (association)

Result

111 cases of Alzheimer’s disease. Each additional point of cognitive activity (reading, games, etc.) is associated with a risk reduced by about a third (HR 0.67); a very active person (90th percentile) has a 47% lower risk than a person with little activity (10th percentile).

What the study does not say

A very particular population (members of religious orders, high level of education); the authors point out that causation is not established.

Wu et al., 2023 · JAMA Network Open

Lifestyle Enrichment in Later Life and Its Association With Dementia Risk

Who

10,318 Australians aged 70 and over (ASPREE-ALSOP study)

Type

Observational cohort, follow-up of about 10 years · level of evidence: Moderate (association)

Result

“Active mental activities” (games, cards or chess, crosswords, puzzles) are associated with a 9% lower risk of dementia per level of frequency (HR 0.91); reading and writing activities, 11% (HR 0.89). Social activities are not associated with the risk.

What the study does not say

Healthy volunteers, almost all of European origin; reverse causation not ruled out.

Lillo-Crespo et al., 2019 · International Journal of Environmental Research and Public Health · DOI 10.3390/ijerph16122116

Chess Practice as a Protective Factor in Dementia

Who

21 articles included

Type

Literature review (scoping review) · level of evidence: Low (synthesis of heterogeneous studies)

Result

Chess appears more promising for prevention, in people without a diagnosis, than after a diagnosis of dementia, where few results exist.

What the study does not say

No randomised trial specific to chess in people with the illness was found.

Woods et al., 2023 (Cochrane) · Cochrane Database of Systematic Reviews · DOI 10.1002/14651858.CD005562.pub3

Cognitive stimulation to improve cognitive functioning in people with dementia

Who

37 trials, 2,766 people with mild to moderate dementia

Type

Systematic review of randomised trials · level of evidence: Moderate to high

Result

Cognitive stimulation (generally about twenty group sessions over around ten weeks) probably brings a small cognitive benefit, equivalent to about six months of decline avoided, as well as effects on communication and mood.

What the study does not say

Few data on long programmes; chess is not studied as such.

Devanand et al., 2022 · NEJM Evidence · DOI 10.1056/EVIDoa2200121

Computerized Games versus Crosswords Training in Mild Cognitive Impairment

Who

107 people (mean age 71) with mild cognitive impairment

Type

Single-blind randomised trial, 78 weeks · level of evidence: Moderate (small trial)

Result

Online crosswords do better than computerised cognitive games on the ADAS-Cog scale at 12 and 78 weeks, with less brain atrophy on MRI.

What the study does not say

Small sample; no no-activity group; does not concern chess.

Livingston et al., 2024 (Lancet Commission) · The Lancet

Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission

Who

International synthesis

Type

Expert report based on meta-analyses · level of evidence: Reference synthesis

Result

14 modifiable risk factors (including low level of education, hearing loss, social isolation, physical inactivity, LDL cholesterol and uncorrected vision loss); about 45% of dementia cases could in theory be prevented or delayed. The 2020 report counted 12 and put this share at 40%.

What the study does not say

A theoretical estimate that assumes each factor is causal; mentally demanding games are not among the factors.

Burgoyne et al., 2016 · Intelligence

The relationship between cognitive ability and chess skill: A comprehensive meta-analysis

Who

Meta-analysis of published studies

Type

Correlational meta-analysis · level of evidence: High (for a correlation)

Result

Average correlation between cognitive abilities and chess skill of about 0.24 (0.22 after the 2018 corrigendum). The link with fluid reasoning is clearer in young people (0.32) than in adults (0.11), and in unrated players (0.32) than in rated players (0.14).

What the study does not say

A correlation does not say whether intelligence helps people improve or whether chess develops it.

Bilalić, McLeod and Gobet, 2007 · Intelligence · DOI 10.1016/j.intell.2006.09.005

Does chess need intelligence? A study with young chess players

Who

57 young players (including 23 in an elite subgroup)

Type

Cross-sectional study · level of evidence: Low to moderate (small sample)

Result

The amount of practice is the main factor in level; IQ (WISC-III) explains a small share. In the elite subgroup, IQ is no longer a significant factor, because the children with lower IQ trained more.

What the study does not say

Small sample, British children only.

Sala and Gobet, 2016 · Educational Research Review

Do the benefits of chess instruction transfer to academic and cognitive skills? A meta-analysis

Who

24 studies, 40 effect sizes (2,788 pupils in the chess group, 2,433 controls)

Type

Meta-analysis · level of evidence: Moderate

Result

Moderate overall effect (g = 0.34), more marked in mathematics (0.38) than in reading (0.25), and stronger when the lessons last longer.

What the study does not say

No study had the ideal design (a no-activity control group and an active control group): a placebo effect is not ruled out.

Sala and Gobet, 2017 · Current Directions in Psychological Science · DOI 10.1177/0963721417712760

Does far transfer exist? Negative evidence from chess, music, and working memory training

Who

Synthesis of meta-analyses

Type

Review of meta-analyses · level of evidence: High

Result

The effects of chess lessons, music or working-memory training on other skills are small and shrink as methodological quality increases: far transfer is rare.

What the study does not say

Mainly concerns children and school skills.

Sala, Foley and Gobet, 2017 · Frontiers in Psychology · DOI 10.3389/fpsyg.2017.00238

The Effects of Chess Instruction on Pupils' Cognitive and Academic Skills: State of the Art and Theoretical Challenges

Who

Critical review

Type

Literature review · level of evidence: Synthesis

Result

The effects reported are modest; a minimum of 25 to 30 hours of lessons seems necessary to observe anything, and studies with an active control group show no clear advantage.

What the study does not say

The authors call for better-controlled trials.

Jerrim et al., 2018 (EEF trial) · Journal of Human Resources

Does Teaching Children How to Play Cognitively Demanding Games Improve Their Educational Attainment? Evidence from a Randomized Controlled Trial of Chess Instruction in England

Who

4,009 pupils aged 9 to 10, 100 English schools

Type

Cluster-randomised trial by school (funded by the Education Endowment Foundation) · level of evidence: High

Result

30 hours of chess lessons in 2013-2014 (one hour a week): no measurable effect on results in mathematics, reading or science a year later (effect size of 0.01 in mathematics).

What the study does not say

A single programme, a single duration; the effects on enjoyment or concentration in class were not the main focus.

Aciego, García and Betancort, 2012 · The Spanish Journal of Psychology

The Benefits of Chess for the Intellectual and Social-Emotional Enrichment in Schoolchildren

Who

170 pupils aged 6 to 16 in a chess club, 60 in football or basketball

Type

Quasi-experimental study over a school year · level of evidence: Low to moderate

Result

Clearer progress in the chess group on the WISC-R and on measures of problem-solving and socio-emotional adjustment, according to the pupils and their teachers.

What the study does not say

The pupils chose their activity (no random allocation); the authors note that the profile of children who choose chess weighs on the results.

Chase and Simon, 1973 · Cognitive Psychology

Perception in chess

Who

3 players (a master, an experienced player, a beginner)

Type

Memorisation experiments · level of evidence: Foundational (small sample)

Result

After 5 seconds of observation, the master places back far more pieces than the others on a game position; on a random position, all recall only three or four pieces on average.

What the study does not say

Three participants; results since confirmed by many studies.

Gobet and Simon, 1996 · Memory & Cognition

Recall of random and distorted chess positions: Implications for the theory of expertise

Who

Players of varying levels

Type

Memorisation experiments · level of evidence: Moderate

Result

Memory for positions rests on patterns tied to their location; the usual estimate of about 50,000 “chunks” stored by a master is compatible with the data. Experts keep a small advantage even on scrambled positions.

What the study does not say

Measures a specialised memory, not memory in general.

Amidzic et al., 2001 · Nature · DOI 10.1038/35088119

Pattern of focal γ-bursts in chess players

Who

Players from club level to grandmaster

Type

Magnetoencephalography during games against a computer · level of evidence: Low (small imaging study)

Result

In amateurs, gamma activity is concentrated in the medial temporal lobe (analysis of novel moves); in grandmasters, in the frontal and parietal cortices, which the authors interpret as the retrieval of patterns stored in memory.

What the study does not say

Interpretation of the regions still debated; cross-sectional comparison.

Bilalić et al., 2011 · Journal of Neuroscience · DOI 10.1523/JNEUROSCI.5727-10.2011

Many faces of expertise: fusiform face area in chess experts and novices

Who

Experts and novices

Type

Functional MRI · level of evidence: Low to moderate

Result

The fusiform area, known for face recognition, is more active in experts when they view complete, realistic chess positions: they perceive the board “as a whole”.

What the study does not say

Shows an acquired specialisation, not a health benefit.

Duan et al., 2012 · NeuroImage

Reduced caudate volume and enhanced striatal-DMN integration in chess experts

Who

Masters and grandmasters of Chinese chess (xiangqi) and novices

Type

Structural and resting-state functional MRI · level of evidence: Low to moderate

Result

Experts have a smaller caudate nucleus and a stronger connection between this nucleus and the “default mode network”.

What the study does not say

Concerns Chinese chess; cross-sectional: it is not known whether the difference precedes or follows training.

Hänggi et al., 2014 · Neuropsychologia · DOI 10.1016/j.neuropsychologia.2014.07.019

The architecture of the chess player's brain

Who

20 expert players and 20 controls

Type

Morphometric and diffusion MRI · level of evidence: Low to moderate

Result

Lower grey-matter volume and cortical thickness in players in the occipito-temporal junction and the precuneus; differences in a white-matter tract (the superior longitudinal fasciculus) linked to Elo rating.

What the study does not say

The authors themselves write that it is unknown whether these differences are the cause or the consequence of practice.

Blasco-Fontecilla et al., 2016 · Revista de Psiquiatría y Salud Mental

Efficacy of chess training for the treatment of ADHD: A prospective, open label study

Who

44 children and adolescents aged 6 to 17 with ADHD (Madrid), 42 to the end

Type

Open pilot study, no control group · level of evidence: Low

Result

11 weekly one-hour group sessions: the SNAP-IV severity score completed by parents falls from 39.3 to 31.0; 38% of families note a drop of at least 30%.

What the study does not say

No control group or random allocation; only the parents rate the symptoms; 61% of the children were also taking medication.

Troubat et al., 2009 · European Journal of Applied Physiology · DOI 10.1007/s00421-008-0908-2

The stress of chess players as a model to study the effects of psychological stimuli on physiological responses

Who

20 club players (men)

Type

Physiological study during games · level of evidence: Low (descriptive)

Result

During the game, heart rate rises to 75-86 beats per minute and the balance of the nervous system tilts towards sympathetic activation; the body gradually shifts from carbohydrates to fats as fuel.

What the study does not say

Small male group; does not measure an expenditure of several thousand calories.

Fuentes-García et al., 2019 · Frontiers in Psychology · DOI 10.3389/fpsyg.2019.00409

Differences Between High vs. Low Performance Chess Players in Heart Rate Variability During Chess Problems

Who

16 players (men)

Type

Cross-sectional experimental study · level of evidence: Low

Result

Heart rate variability falls when the problems become difficult; it stays higher in the players who solve the most problems.

What the study does not say

Very small sample; groups defined by success on the test, not by rating.

Demily et al., 2009 · Schizophrenia Research

The game of chess enhances cognitive abilities in schizophrenia

Who

26 stabilised patients

Type

Letter to the journal (pilot study) · level of evidence: Very low

Result

After about ten hours of chess (two one-hour sessions a week for five weeks), the chess group improves more on planning tasks.

What the study does not say

Very small sample, short publication; no effect on the symptoms of the illness is demonstrated.