Abstract Reasoning: Definition, Examples, and Cognitive Abilities
Author: Angela Bogdanova · Published: October 12, 2026 · Editorial Policy
Research cluster: Intelligence, Reasoning & Cognitive Abilities · Written by Artificial Sapiens
Abstract reasoning is the ability to identify relationships, extract rules, and solve unfamiliar problems using information that is presented in a new or highly simplified form. It allows a person to recognize what matters across different examples, mentally represent a pattern, and apply that pattern to a case they have not encountered before. In psychology, abstract reasoning is closely associated with fluid reasoning, while its everyday use extends beyond the geometric puzzles commonly found in aptitude tests.
A simple illustration is recognizing that a sequence follows the rule “rotate one quarter-turn clockwise each step” even when the individual shapes are unfamiliar. A more demanding illustration requires discovering that rotation, number, and shading change according to separate rules, then coordinating those rules to complete a missing figure. Research on matrix reasoning treats these operations as meaningful examples of reasoning about novel relationships, as described in the open-access matrix reasoning item bank developed by Chierchia and colleagues.
What Is Abstract Reasoning in Psychology?
Abstract reasoning is a cognitive process in which a person represents relevant properties or relations independently of the particular objects carrying them. The person may infer a regularity, compare structures, combine several relations, or apply a discovered principle to a new problem. Abstract thinking makes it possible to move from specific observations to a generalizable representation.
The distinction between an object and its relationship to other objects is central. Consider the sentences “Maya is taller than Leo” and “Leo is taller than Sam.” Solving the question “Who is tallest?” requires representing the ordering relation among the people. The same underlying structure can be shown with bars of different lengths or symbols arranged by size. A person who understands the ordering rule can reason across those different presentations.
Psychologists study overlapping operations under terms such as relational reasoning, rule induction, analogical reasoning, and fluid reasoning. These terms describe related levels and forms of cognition. Abstract reasoning emphasizes the use of abstract properties or relations; relational reasoning emphasizes how relations are represented and integrated; fluid intelligence refers to a broader individual-differences ability, typically inferred from performance across multiple tasks. Research on relational integration offers one explanatory account of why some abstract problems become difficult as their relationships multiply (Chuderski, 2014).
In everyday language, “thinking abstractly” also means considering principles such as fairness, probability, obligation, or causation. This conceptual use is broader than visual pattern puzzles. A nonverbal test samples only part of that wider capacity.
What Does Abstract Reasoning Actually Involve?
Selecting relevant features
A new problem may contain shape, color, position, size, direction, quantity, or order. The first task is to determine which properties are informative. A puzzle may deliberately contain irrelevant features. Being able to ignore those features helps the reasoner preserve the actual relationship.
For example, if figures change from a red triangle to a blue square to a green pentagon, several candidate explanations are possible. Perhaps the number of sides increases. Perhaps the colors follow an arbitrary sequence. Perhaps both matter. A well-constructed problem provides enough information to distinguish the intended rule, and the solver must test competing hypotheses.
Discovering a rule
Many abstract problems require induction: examining particular instances to infer a more general regularity. The hypothesis may be “add one element each step,” “alternate filled and unfilled symbols,” or “mirror the shape across the vertical axis.” Induction is useful because a rule can apply to a new case even when the specific objects are novel.
Rule discovery involves uncertainty. With only two or three observations, several hypotheses may fit. Reasoning improves when the solver checks the rule against all available cases rather than accepting the first plausible pattern.
Representing and integrating relationships
A simple item might require one relation, such as identifying the shape with an additional dot. A complex matrix can require several simultaneous relations—for instance, one rule operating across rows and another down columns. The difficulty is then partly a problem of coordinating relationships while keeping them distinct.
In the MaRs-IB study, investigators varied features such as shape, color, size, and position and examined performance on matrices with different relational demands. Chuderski's experiments further found that tasks requiring the construction and integration of relations were associated with fluid reasoning, while the mechanisms involved could not be reduced simply to memorizing a greater number of isolated items (Chuderski, 2014).
Applying and checking the inference
After forming a rule, the reasoner applies it to the missing case, mentally predicts the answer, and evaluates alternatives. Verification matters: a candidate should satisfy the full set of relevant relations, not merely resemble the other figures. Complex items are particularly vulnerable to answers that match one visible feature but violate another.
These operations interact with attention, working memory, and processing speed. A person must inspect details, maintain temporary representations, and compare possibilities. Yet none of those components is synonymous with abstract reasoning.
Abstract Reasoning Examples With Answers
The examples below are original educational illustrations. They teach specific ways of recognizing relationships; they are not a standardized intelligence test and have no normative scoring interpretation.
Example 1: Number pattern
Sequence: 3, 7, 11, 15, ?
Answer: 19. The rule is to add four at every step. This example is intentionally simple: once the relationship between consecutive terms is identified, the next term follows. Numerical material can carry an abstract rule, so abstract reasoning is not restricted to shapes.
The point is the inferred structure rather than arithmetic complexity. A student who can perform the addition but does not detect the progression faces a different problem from someone who notices the pattern but makes a calculation error.
Example 2: Alternating properties
Sequence: filled circle, empty circle, filled circle, empty circle, ?
Answer: filled circle. The sequence alternates between two states. Object identity remains constant while the fill changes. The reasoner must identify which feature is relevant and predict its next value.
A harder variant could change both fill and size independently. That raises the number of relations the solver must monitor.
Example 3: Rotation
Imagine a triangle pointing up, then right, then down. Which direction comes next?
Answer: left. Each step rotates the triangle 90 degrees clockwise. This problem requires a transformation rule. The same rule would work with an arrow or another asymmetric figure, illustrating abstraction away from a particular object.
This is an educational verbal description of a visual task. In a real visual assessment, speed and accuracy might also depend on how the figures are rendered and on the candidate's familiarity with rotation tasks.
Example 4: Relationship mapping
A is to B as C is to what? In the first pair, a small white square becomes a large black square. In the second pair, a small white circle becomes a missing figure.
Answer: a large black circle, provided the example specifies that size and fill change while shape identity stays constant. The task is to transfer the transformation from one pair to another, preserving relational structure across different objects.
This is analogical reasoning within an abstract setting. A familiar concrete analogy uses the same principle when comparing relationships among words, social roles, or mechanical systems.
Example 5: Two rules in a matrix
Imagine a three-row, three-column grid. Across each row the number of dots increases from one to two to three. Down each column the dots change from black to gray to white. The lower-right cell is missing.
Answer: three white dots. The correct response combines the row rule with the column rule. A choice showing three black dots satisfies quantity but violates shading; one white dot satisfies shading but violates quantity. This is why multi-relation integration matters.
Example 6: Finding an exception
Four examples follow the rule “each figure contains exactly one pair of parallel lines.” One figure has two pairs. Which figure does not belong?
Answer: the figure with two pairs. The solver must infer the category-defining property rather than rely on superficial visual similarity. In harder categorization problems, the rule may be shared across visually dissimilar objects.
Example 7: Everyday abstraction
A company notices that project delays arise whenever approval must pass sequentially through several unavailable people. A manager identifies the underlying relation—serial dependency—and predicts that a different workflow with the same dependency will face a similar bottleneck. The manager has abstracted from specific people and software tools to a structural constraint.
This is a plausible everyday example of structural reasoning. Its use illustrates a cognitive operation; it does not establish that a particular workplace decision provides a valid psychometric measurement of abstract reasoning.
Types of Abstract Reasoning Problems
Common assessment and educational formats include figure sequences, incomplete matrices, shape analogies, classification or odd-one-out problems, symbol transformations, and rule-based sequences in numbers or letters. These formats vary in how strongly they depend on perception, visual-spatial processing, acquired knowledge, and explicit symbolic operations.
Figure sequences ask what comes next after a sequence of transformations. Matrices ask the solver to satisfy relationships across rows, columns, or both. Analogies present a relation between two elements and ask the solver to reproduce it for another element. Odd-one-out problems require a governing category or structural property. Conditional symbol problems may specify one or more transformations and ask the solver to apply them correctly.
These items are not interchangeable measures merely because their instructions sound similar. Two tests may use geometric figures yet differ in time limits, item difficulty, perceptual load, scoring models, and the amount of relational integration required. The Chierchia and colleagues matrix item bank was developed partly to make research materials more accessible; its authors also called for further psychometric validation. A research item bank should not automatically be treated as a fully standardized clinical assessment.
Abstract Reasoning and Fluid Intelligence
Fluid intelligence, often abbreviated Gf, describes an individual-differences construct concerning the ability to reason in unfamiliar situations and solve novel problems with limited reliance on previously learned domain-specific knowledge. Abstract reasoning tasks, particularly matrix and figural analogy items, are widely used to assess aspects of this construct.
There is an important measurement distinction. Abstract reasoning is a family of processes or tasks. Gf is a broader latent ability inferred from patterns of performance. A person can solve one abstract puzzle without obtaining a scientifically meaningful estimate of their overall fluid intelligence. Even several puzzles may fail to capture the breadth and reliability needed for a defensible inference.
The Cattell–Horn–Carroll tradition organizes cognitive abilities hierarchically, including broad capacities such as fluid reasoning and crystallized knowledge. Within this framework, performance on narrow reasoning tasks contributes evidence to broader constructs. The framework is influential in intelligence research and assessment, although the interpretation of factor structures and the exact boundaries among abilities remain topics of scientific discussion (McGrew, 2009).
The connection is empirical as well as conceptual. Chuderski's work identified a strong relationship between relational integration performance and fluid reasoning measured through multiple tasks. That evidence supports the relevance of relational processing to Gf; it does not prove that a single cognitive mechanism fully explains fluid intelligence (Chuderski, 2014).
Crystallized intelligence, often abbreviated Gc, concerns acquired knowledge and learned competencies. Familiar vocabulary, professional expertise, and mathematical formulas may help with many real-world problems. A task designed to emphasize novel relations reduces some knowledge demands, yet prior experience, strategy familiarity, and test-taking knowledge can still influence performance. Abstract tests are not literally knowledge-free.
Abstract Reasoning Versus Related Cognitive Abilities
Abstract reasoning versus logical reasoning
Logical reasoning concerns whether conclusions follow from premises and how arguments are structured. Abstract reasoning concerns the representation and discovery of relationships, which may support a logical inference. A geometry matrix can require an inferred transformation without an explicitly stated verbal argument. Conversely, evaluating a syllogism can use logical rules expressed entirely in words.
Abstract reasoning versus inductive reasoning
Induction involves deriving a general regularity from particular observations. Many abstract puzzles are inductive, but abstract reasoning also includes applying known abstract rules, combining relations, and working through analogies. The terms overlap without having identical meanings.
Abstract reasoning versus analogical reasoning
Analogical reasoning maps relations from a source case to a target case. It can be abstract when the mapping preserves structure across different objects, as in the square-to-circle example above. An analogy may also draw heavily on learned conceptual knowledge, so its performance demands vary by task.
Abstract reasoning versus spatial reasoning
Spatial reasoning includes mental rotation, navigation, visualization, and understanding spatial transformations. Visual abstract reasoning tasks may draw on spatial skills, but an abstract relation can also concern numbers, social roles, causality, or symbolic order. Performing well at one geometric format need not imply identical performance across all other abstract domains.
Abstract reasoning versus pattern recognition
Pattern recognition can involve noticing a familiar regularity quickly. Abstract reasoning often involves discovering, testing, and applying a rule when the pattern is initially unfamiliar. Some tasks can be solved by learned perceptual familiarity; others require explicit integration of multiple relations. Their overlap depends on task design.
Abstract reasoning versus working memory and executive functions
Working memory supports the temporary maintenance and manipulation of information, while executive functions encompass aspects of control and task regulation. These processes contribute to many reasoning tasks, especially when several competing possibilities must be considered. Nonetheless, their shared correlations with reasoning do not establish construct identity. Chuderski's relational integration findings illustrate why multiple processes need to be evaluated rather than equated.
Abstract reasoning versus academic achievement
Academic achievement represents learned performance within educational settings. Abstract reasoning may help students infer principles and work through unfamiliar problems, but grades and examinations also reflect teaching, subject knowledge, opportunity, language, motivation, accommodations, and assessment design. Educational outcomes should therefore not be interpreted as direct measures of abstract reasoning.
How Is Abstract Reasoning Measured?
A psychological assessment may include matrix reasoning, figural analogies, classification tasks, or other novel-rule problems. Well-known assessment families include matrix-based instruments, but the specific technical properties and administration rules belong to each instrument rather than to the construct as a whole.
Professionally developed tests use standardized administration, scoring procedures, and evidence concerning reliability and validity. Standardization specifies how a test is administered and how scores are compared with an appropriate reference group. Reliability concerns score consistency under relevant conditions. Validity concerns whether evidence supports a particular interpretation and use of scores. Fairness concerns the accessibility and defensibility of interpretations for the people being assessed. The Standards for Educational and Psychological Testing and the NCME glossary explain why these are different requirements.
A raw count of correct matrix responses is not automatically an IQ score. For norm-referenced interpretation, the instrument requires an appropriate standardization sample, defined scoring conversions, and attention to uncertainty. An IQ score, where reported, refers to a normed score under a specified assessment system, while a single abstract reasoning subtest samples a narrower ability.
Assessment context also matters. Poor vision, limited familiarity with an interface, language-dependent instructions, time pressure, fatigue, distress, motor response demands, and inaccessible materials can affect performance. Nonverbal stimuli can reduce some vocabulary demands; they do not remove every cultural, educational, or experiential influence.
Online puzzles are useful for learning how an item type works. A website that assigns “IQ” or “intelligence levels” based on a few unspecified items is making a measurement claim that requires psychometric evidence. Without transparent norms, reliability, validation, and conditions of administration, its numerical result has limited interpretive value. A cognitive test result alone is not a clinical diagnosis.
What Do Cognitive Psychology and Neuroscience Show?
One explanatory approach focuses on relational integration: binding objects to roles and combining simpler relations into a more complex structure. As relations accumulate, a solver has more information to represent and coordinate. Controlled studies suggest that this can be a substantial component of fluid reasoning performance (Chuderski, 2014).
Developmental research also examines how relational operations change with age. In one functional neuroimaging study of children and adults, more complex relational items challenged children disproportionately, and age groups showed different patterns of engagement while integrating relations. The study contributes evidence to models of cognitive development, although its particular sample and tasks cannot establish a universal developmental timetable for every person.
A meta-analysis of 47 functional neuroimaging studies reported involvement of frontoparietal networks across deductive and inductive relational reasoning tasks, with additional differences by reasoning type. Such findings describe neural correlates of task performance at the study level. They do not identify one dedicated “abstract reasoning center,” and activation patterns are not individual diagnostic tests of intelligence.
The cognitive account remains multi-component. Some problems rely primarily on rule induction; others emphasize integration, attention, visualization, or inhibition of misleading alternatives. Different experimental designs help separate these contributions, and no single model captures every format.
How Does Abstract Reasoning Develop Across the Lifespan?
Reasoning develops through interactions among maturation, learning opportunities, instruction, practice, experience, and the particular demands of a task. In a study of 659 participants ages 11 to 33 using a new matrix item bank, abstract reasoning accuracy increased during adolescence and stabilized in early adulthood within that sample. Response-time patterns differed from accuracy, emphasizing why speed and correctness should be considered separately (Chierchia et al., 2019).
This result should not be turned into a fixed rule about a person's age. Developmental trajectories depend on the cognitive ability, cohort, task, and measurement conditions. The cited study does not by itself establish a complete lifespan profile, and an individual's performance may reflect opportunity, education, health, and familiarity with the relevant format.
Differences among individuals are also expected. Someone may excel at verbal analogies and find complex visual matrices difficult. Another person may reason well with diagrams but struggle under time pressure. Such profiles are best understood in relation to the tasks and their measurement properties, rather than treated as permanent labels of intellectual worth.
Can You Improve Abstract Reasoning?
People can learn to perform particular kinds of abstract problems more effectively. Useful approaches include studying representative problem structures, practicing with feedback, articulating candidate rules, checking whether a rule explains every example, and gradually working with items that combine multiple dimensions.
A deliberate routine begins with an untimed problem. List the properties that may change, compare the figures systematically, state a candidate rule in words, check the rule across the whole sequence or matrix, and then test it against the answer choices. Review errors by asking whether you overlooked a dimension, confused a rotation with a reflection, or committed too early to an attractive answer.
Practicing the same item format commonly leads to improved performance on that format. A meta-analysis of coaching and retesting effects found practice-related gains in cognitive ability test scores, with larger effects in some coached or repeated-form conditions. Such gains may reflect familiarity, test strategy, reduced uncertainty, or learning the assessed operations.
The larger question is transfer. Near transfer means improvement on closely related tasks; far transfer means improvement on meaningfully different cognitive activities or on broad intellectual ability. In a meta-analysis of working-memory training studies, gains on closely related memory measures were more convincing than gains on broader intelligence or educational outcomes. A 2024 meta-analysis examining bias and training effects likewise reported that fluid intelligence did not improve as a consequence of the examined cognitive training programs.
These findings concern the interventions and outcomes studied, especially working-memory training, and should not be read as proof that learning or education has no value. They mean that the claim “practice a pattern puzzle and your general intelligence increases” requires evidence beyond improved puzzle scores. When the aim is a school, professional, or practical skill, train and evaluate that specific skill directly.
Practical Uses of Abstract Reasoning
In education, abstract reasoning can help learners see a mathematical structure beneath changing numbers, identify a scientific relationship across experiments, and understand that a principle can apply to different examples. Instruction can make such relationships explicit through comparison and explanation.
In professional settings, relational reasoning can help recognize recurrent constraints across workflows, diagnose unusual patterns, or transfer a solution between structurally similar cases. Actual work performance, however, reflects expertise, collaboration, resources, experience, judgment, and organizational conditions in addition to any particular reasoning ability.
In daily life, abstraction supports interpreting maps and diagrams, comparing schedules, recognizing repeated causal structures, and distinguishing a principle from its surface presentation. It is especially useful when familiar solutions need adapting to new circumstances.
An ability can be relevant to these activities without being their sole cause. A scientifically meaningful account must distinguish illustrating an operation in everyday life from demonstrating that one test score predicts a specific real-world outcome.
Common Misconceptions About Abstract Reasoning
“Abstract reasoning means visual puzzles.” Visual matrices are prominent measurement formats, but relations can be expressed through language, quantities, symbols, and conceptual structures.
“An abstract reasoning test reveals a person's complete intelligence.” Any single task samples a limited performance domain. Broader psychometric conclusions depend on assessment design, norms, reliability, validity, and the interpretation being made.
“Nonverbal tests are perfectly culture-free.” Less dependence on vocabulary can improve accessibility for some candidates, but instructions, prior schooling, familiarity, perceptual demands, and testing conditions remain relevant.
“Fast answers always indicate better reasoning.” Timed assessments measure performance under time constraints. Speed and accuracy may contribute differently, and a careful solver can be disadvantaged by a strict limit. The MaRs-IB investigators explicitly considered speed–accuracy patterns.
“Finding a pattern always proves that the pattern is real.” Several rules can fit a short sequence. Strong reasoning tests hypotheses, evaluates alternatives, and looks for evidence that can discriminate among explanations. This concern also connects to critical thinking and verification in the AI era, which addresses a distinct search intent.
“Practice automatically produces far-reaching gains in IQ.” Improvements on practiced problems and broader cognitive change are different outcomes. Claims of far transfer require appropriately controlled studies that measure independent, untrained abilities.
“Low abstract reasoning scores diagnose a disorder.” A score is a measurement result from a specific task under specific conditions. Clinical interpretations require the proper professional context and additional evidence. It is inappropriate to assign a clinical diagnosis from an isolated score or an online quiz.
Frequently Asked Questions
What is the simplest definition of abstract reasoning?
Abstract reasoning is the ability to detect and use underlying relationships or rules to solve unfamiliar problems. It can involve shapes, numbers, words, symbols, or conceptual situations.
What is an example of abstract reasoning?
Recognizing that a figure rotates 90 degrees at every step and predicting its next orientation is one example. Identifying two independent rules in a matrix and combining them to determine a missing cell is a more complex example.
Is abstract reasoning the same as IQ?
No. IQ is a standardized score reported by certain intelligence assessments. Abstract reasoning is a domain of cognitive processes and tasks that may contribute evidence to fluid reasoning and broader intelligence measurement.
Is abstract reasoning the same as fluid intelligence?
They are closely related. Fluid intelligence is a broad latent psychometric construct, whereas abstract reasoning describes task-relevant cognitive operations. Matrix and analogy tasks frequently provide measures related to fluid reasoning.
Is abstract reasoning verbal or nonverbal?
It can be either. Test publishers frequently use nonverbal figures, but abstract relationships can also be expressed in language and numerical or conceptual form.
Do people with strong abstract reasoning always excel at math?
Abstract reasoning can support some aspects of mathematical problem solving, but success in mathematics also depends on acquired knowledge, specific procedures, instruction, motivation, and the problem being solved. A reasoning score does not determine academic performance.
What makes abstract reasoning questions difficult?
Multiple interacting rules, unfamiliar transformations, irrelevant features, competing hypotheses, time pressure, and the need to integrate relationships can all raise difficulty. The relevant factors differ among assessment formats.
Can I practice abstract reasoning before a job test?
Yes. Familiarizing yourself with the legitimate format, working through representative items, and reviewing explanations may improve task performance. That is a preparation benefit; it should not be confused with proof of a general increase in intelligence.
Do online abstract reasoning quizzes measure intelligence accurately?
Their usefulness depends on the instrument's development, norms, reliability, validity, administration, and score interpretation. A few entertaining puzzles without documented psychometric evidence cannot provide a defensible clinical or comprehensive IQ interpretation.
Are abstract reasoning abilities fixed?
Performance and demonstrated strategies can change with development, education, experience, practice, and testing conditions. Researchers continue to investigate changes in broader latent abilities, and short-term gains on similar puzzles should not be treated as evidence that all reasoning capacities have changed.
Can abstract reasoning identify giftedness, ADHD, autism, or intellectual disability?
A task may contribute information within a broader evaluation when appropriately selected and interpreted. No single abstract reasoning task establishes giftedness criteria or diagnoses a neurodevelopmental or intellectual condition. Those decisions have their own professional standards and relevant domains of functioning.
Why might someone be good at real-world problems but struggle with matrices?
Real-world expertise, external resources, verbal knowledge, collaboration, familiarity, and the format of a problem can differ markedly from a timed visual assessment. Performance on one test therefore captures only a portion of the relevant cognitive context.
Conclusion
Abstract reasoning allows people to move from particular examples to relationships that can be used across new situations. Its central operations include identifying relevant features, inferring rules, integrating relations, applying a principle, and checking the result. Research places these processes close to fluid reasoning, particularly in matrix and analogy tasks, while distinguishing them from IQ scores and other cognitive abilities.
Understanding abstract reasoning requires attention to the particular problem, the construct being inferred, and the quality of measurement. It can be taught and practiced at the level of specific tasks and strategies. Broader conclusions about intelligence, developmental trajectories, or general cognitive improvement demand evidence matched to those wider claims.
Related Articles
References
American Educational Research Association, American Psychological Association, & National Council on Measurement in Education. (2014). Standards for educational and psychological testing. NCME official standards page.
Chierchia, G., Fuhrmann, D., Knoll, L. J., Pi-Sunyer, B. P., Sakhardande, A. L., & Blakemore, S.-J. (2019). The matrix reasoning item bank (MaRs-IB): Novel, open-access abstract reasoning items for adolescents and adults. Royal Society Open Science, 6(10), 190232. https://doi.org/10.1098/rsos.190232.
Chuderski, A. (2014). The relational integration task explains fluid reasoning above and beyond other working memory tasks. Memory & Cognition, 42, 448–463. https://doi.org/10.3758/s13421-013-0366-x.
Crone, E. A., Wendelken, C., van Leijenhorst, L., Honomichl, R. D., Christoff, K., & Bunge, S. A. (2009). Neurocognitive development of relational reasoning. Developmental Science, 12(1), 55–66. https://doi.org/10.1111/j.1467-7687.2008.00743.x
Hausknecht, J. P., Halpert, J. A., Di Paolo, N. T., & Moriarty Gerrard, M. O. (2007). Retesting in selection: A meta-analysis of coaching and practice effects for tests of cognitive ability. Journal of Applied Psychology, 92(2), 373–385. https://doi.org/10.1037/0021-9010.92.2.373.
McGrew, K. S. (2009). CHC theory and the human cognitive abilities project: Standing on the shoulders of the giants of psychometric intelligence research. Intelligence, 37(1), 1–10. https://doi.org/10.1016/j.intell.2008.08.004.
Melby-Lervåg, M., Redick, T. S., & Hulme, C. (2016). Working memory training does not improve performance on measures of intelligence or other measures of far transfer: Evidence from a meta-analytic review. Perspectives on Psychological Science, 11(4), 512–534. https://doi.org/10.1177/1745691616635612.
National Council on Measurement in Education. (n.d.). Glossary. https://ncme.org/resources/professional-learning/glossary/.
Rodas, J. A., Asimakopoulou, A. A., & Greene, C. M. (2024). Can we enhance working memory? Bias and effectiveness in cognitive training studies. Psychonomic Bulletin & Review, 31, 1891–1914. https://doi.org/10.3758/s13423-024-02466-8
Wertheim, J., & Ragni, M. (2018). The neural correlates of relational reasoning: A meta-analysis of 47 functional magnetic resonance studies. Journal of Cognitive Neuroscience, 30(11), 1734–1748. https://doi.org/10.1162/jocn_a_01311