Executive Function

Abstract

Executive function, a type of mental process, is the umbrella term for the top-down control processes that organize thought and action toward a goal. Cognitive psychology has converged on three separable components, inhibition, working-memory updating, and cognitive flexibility, that are correlated yet distinct, a structure known as unity and diversity. This article traces the construct from the central executive and the Supervisory Attentional System to prefrontal accounts of cognitive control, surveys how the components are measured and how they develop across childhood, and reviews their near-total heritability and their disruption in ADHD and the dysexecutive syndrome. Three interactive demonstrations let the reader perform the canonical task for each component, a Stroop task for inhibition, a 2-back for updating, and a cued task-switching task for shifting, and read off their own Stroop effect, updating accuracy, and switch cost.

Keywords: executive function, cognitive control, inhibition

Executive function names the set of general-purpose control processes that regulate more basic cognitive operations in the service of a goal, and it is invoked precisely when automatic or habitual responding will not suffice (Diamond, 2013). The construct grew out of the clinical observation that damage to the frontal lobes could spare memory, language, and perception while devastating the capacity to plan, to resist distraction, and to adapt when circumstances change (Stuss & Alexander, 2000). Modern cognitive psychology has given that clinical intuition a precise structure, a set of measurable components, and a developmental and genetic account.

The Three Core Components

The dominant taxonomy divides executive function into three moderately correlated components (Miyake et al., 2000). Inhibition is the deliberate suppression of a dominant, automatic, or prepotent response when it conflicts with the current goal. Updating is the continuous monitoring and revision of the contents of working memory, adding relevant information and discarding what is no longer needed. Shifting, also called cognitive flexibility or set shifting, is the ability to move between tasks, rules, or mental sets. Diamond frames these as the three core executive functions from which higher-order abilities such as reasoning, problem solving, and planning are built (Diamond, 2013).

Inhibition is most cleanly exposed by the Stroop task, in which naming the ink color of a printed word is slowed when the word spells a conflicting color, because the practiced act of reading produces a competing response that must be held back (Stroop, 1935). The demonstration below is a short Stroop task the reader can run directly: naming the ink color of each word by clicking, and because reading is automatic, the incongruent trials in which the word fights the correct response come out slower. The gap between the reader's incongruent and congruent times is a personal Stroop effect, a direct index of inhibitory control.

Updating is exposed by the n-back task, which forces the contents of working memory to be continuously revised rather than merely held (Miyake et al., 2000). In the demonstration below, letters appear one at a time and the reader decides whether each matches the letter shown two positions back, so the pair being tracked has to be updated on every step; the resulting accuracy is a measure of how well working memory is monitored and revised, not merely stored.

Shifting carries its own signature, the switch cost: alternating between two tasks is reliably slower than repeating either one, because the mental task set must be reconfigured (Miyake et al., 2000). The demonstration below cues one of two rules for each digit, whether it is odd or even, or whether it is below or above five, and lets the rule repeat or switch from trial to trial. The extra time the switch trials take over the repeat trials is the switch cost, which is one reason shifting is treated as a component in its own right rather than a by-product of the other two.

The three components are summarized in Table 1, each with its defining operation and a canonical laboratory task.

Unity and Diversity

The single most influential empirical claim about executive function is that its components are neither fully unified nor fully separate. Using confirmatory factor analysis across nine tasks, Miyake and colleagues found that inhibition, updating, and shifting were clearly distinguishable as three latent factors, yet those factors were moderately correlated with one another, evidence of some shared underlying ability (Miyake et al., 2000). This unity-and-diversity structure has been replicated and elaborated many times, and a later synthesis distilled it into four general conclusions: the components are correlated but separable, they are highly heritable, they relate to a range of important outcomes, and a common factor plus component-specific factors best captures the data (Miyake & Friedman, 2012). A subsequent review recast the same structure as a window onto cognitive architecture, arguing that the common factor reflects the ability to maintain and actively use goal-relevant information (Friedman & Miyake, 2017).

The logic is straightforward. If each task draws on one shared executive factor plus an independent task-specific factor, then the correlation between any two tasks equals the proportion of variance they share through that common factor. A positive but sub-unity correlation, which is what the data show, is exactly the unity-and-diversity signature: enough shared variance to make the components correlate, not so much as to collapse them into one.

Models of Executive Control

The construct has deep roots in models of attention and working memory. Norman and Shallice proposed that routine action is governed by automatic schemas competing through contention scheduling, but that non-routine situations, those requiring planning, error correction, or the overriding of a strong habit, recruit a separate Supervisory Attentional System that biases the competition from above (Norman & Shallice, 1986). Baddeley imported essentially the same idea into his model of working memory as the central executive, the attentional-control component that coordinates the subsidiary storage buffers, and he acknowledged that it had long been the least understood and most casually invoked part of the model (Baddeley, 1996). Much subsequent work can be read as an effort to decompose that central executive into the specific, measurable functions surveyed above.

A related tradition ties executive control tightly to working memory capacity. Engle argued that what capacity-limited working-memory measures predict is not storage as such but executive attention, the ability to maintain goal-relevant information in the face of interference, which is why working-memory span predicts performance on the Stroop task and other conflict tasks (Engle, 2002).

The Prefrontal Cortex and Cognitive Control

The neural home of executive function is chiefly the prefrontal cortex, though not it alone. In the integrative theory of Miller and Cohen, the prefrontal cortex holds representations of goals and the means to achieve them, and it exerts control by sending top-down bias signals that guide activity along the neural pathways needed to link inputs to goal-appropriate outputs (Miller & Cohen, 2001). On this account control is not a separate homunculus but a pattern of biased competition: the prefrontal cortex tilts processing elsewhere in the brain toward what the current goal requires. A complementary line of evidence identifies a multiple-demand system, a specific frontoparietal network that is recruited across an unusually wide range of cognitive challenges and is closely tied to fluid intelligence and to the assembly of complex behavior from simpler mental steps (Duncan, 2010). Frontal injury correspondingly produces the dysexecutive syndrome, a disorganization of goal-directed behavior that can leave more elementary abilities intact (Stuss & Alexander, 2000).

Development

Executive function has a protracted developmental course that tracks the slow maturation of the prefrontal cortex. The rudiments are visible in infancy, but inhibition, working memory, and flexibility continue to improve markedly through childhood and into adolescence and early adulthood (Best & Miller, 2010). A standard measure for young children is the Dimensional Change Card Sort, in which a child first sorts cards by one dimension, such as color, and is then asked to switch to another, such as shape; three-year-olds typically persevere on the old rule while five-year-olds switch successfully, a sharp developmental transition in cognitive flexibility (Zelazo, 2006). The stakes of these early differences are considerable. A longitudinal study following a birth cohort to age thirty-two found that childhood self-control, an executive-function-laden trait, predicted adult physical health, personal finances, and criminal-offending outcomes, even after adjusting for intelligence and social class (Moffitt et al., 2011). Crucially, executive function is trainable: a review of school-based and other interventions concluded that activities as varied as certain curricula, computerized training, and martial arts can improve executive function in children, with the largest gains for those who start out weakest (Diamond & Lee, 2011).

Assessment

Because executive function is defined by control rather than by any single content, it is measured through tasks engineered to pit a goal against a habit or to demand rapid reconfiguration. Inhibition is probed by the Stroop task and by stop-signal and flanker paradigms (Stroop, 1935). Updating is probed by the n-back and running-span tasks that force continuous revision of working-memory contents (Miyake et al., 2000). Shifting is probed by explicit task-switching paradigms, in which a cue specifies which of two rules to apply to each stimulus so that switch trials can be compared with repeat trials, and by the plus-minus and number-letter tasks Miyake and colleagues used to define the shifting factor. A persistent methodological difficulty, the task-impurity problem, is that every executive task also recruits non-executive processes specific to its materials, so any single task is an imperfect index of the underlying function. The latent-variable approach was devised precisely to extract what the tasks share and set aside what is idiosyncratic to each (Miyake et al., 2000).

Clinical and Individual Differences

Individual differences in executive function are strikingly stable and strikingly heritable. A behavioral-genetic study of twins found that individual differences in the common executive factor were almost entirely genetic in origin, with heritability estimates near unity and little contribution of the shared family environment (Friedman et al., 2008). Executive deficits also sit at the center of several clinical accounts. Barkley proposed that attention-deficit/hyperactivity disorder is fundamentally a disorder of behavioral inhibition, whose downstream consequences impair the executive functions, including working memory and self-regulation, that depend on the capacity to delay a response (Barkley, 1997). According to MeSH, executive function is a distinct descriptor within mental processes, and its impairment features across a range of developmental and acquired conditions, from ADHD to the dysexecutive syndrome of frontal injury (Stuss & Alexander, 2000).

Worked Example

The updating demonstration can be reproduced by hand, which is the surest check that its scoring means what the prose claims. Its twelve-letter sequence is fixed: F, K, F, C, D, C, B, H, B, K, H, K. Comparison begins at the third letter, because the first two have nothing two positions back. Working forward, position three (F) matches position one (F); position six (C) matches position four (C); position nine (B) matches position seven (B); and position twelve (K) matches position ten (K). No other letter matches the one two before it. That is four 2-back targets among the ten scored letters. A reader who responds perfectly therefore records four hits, six correct rejections, no misses, and no false alarms, for an accuracy of ten out of ten. Those are exactly the totals the demonstration reports for a flawless run, so the worked count and the interactive scoring agree.

Discussion

Executive function matters because it is the part of cognition that makes behavior deliberate rather than merely reactive. Its unity-and-diversity structure resolves an old dispute between those who treated the frontal lobes as a single supervisory faculty and those who catalogued a long list of unrelated deficits: both were partly right, because the components share a common resource while retaining separable identities (Miyake & Friedman, 2012). The construct also carries unusual practical weight. It predicts academic achievement, health, and wealth; it is highly heritable yet demonstrably trainable; and it is disrupted in some of the most common conditions of childhood and later life (Diamond, 2013). Open questions remain, chief among them how cleanly the components separate in early childhood, where a single factor often fits better, and how the common factor should be characterized in mechanistic rather than statistical terms (Friedman & Miyake, 2017).

Glossary

Central Executive.

In Baddeley's model of working memory, the attentional-control component that coordinates the storage buffers and allocates processing resources.

Cognitive Control.

The general capacity to guide thought and action according to internal goals rather than habit or immediate stimulation; often used interchangeably with executive function.

Cognitive Flexibility.

The ability to shift between tasks, rules, or perspectives; the shifting component of executive function.

Dysexecutive Syndrome.

The pattern of disorganized, poorly regulated goal-directed behavior that follows frontal-lobe damage, often with elementary abilities preserved.

Executive Attention.

The maintenance of goal-relevant information against interference, proposed as the core of what working-memory capacity measures.

Fluid Intelligence.

The capacity to reason and solve novel problems independent of acquired knowledge, closely tied to the multiple-demand system and to executive function.

Goal Neglect.

The failure to apply a task requirement that has been understood and remembered, a hallmark of weak executive control and of multiple-demand dysfunction.

Inhibition.

The deliberate suppression of a dominant, automatic, or prepotent response that conflicts with the current goal.

Multiple-Demand System.

A frontoparietal network recruited across diverse cognitive challenges and linked to fluid intelligence and the assembly of complex behavior.

Prepotent Response.

A response that is strongly favored by habit, practice, or salience, and that inhibition must override when it is inappropriate.

Set Shifting.

Moving from one active task set or rule to another; the operation measured by task-switching and card-sorting paradigms.

Supervisory Attentional System.

In the Norman and Shallice model, the mechanism that intervenes in non-routine situations to bias the automatic selection of action schemas.

Switch Cost.

The extra response time incurred on a task-switch trial relative to a task-repeat trial, comprising a preparation-reducible part and a stubborn residual.

Task Set.

The configuration of perception, attention, and response mappings that a particular task requires, which must be reconfigured when the task changes.

Unity and Diversity.

The finding that executive components are correlated (they share a common factor) yet separable (each retains distinct variance).

Updating.

The continuous monitoring and revision of the contents of working memory, adding relevant items and dropping obsolete ones.

Working Memory.

The limited-capacity system that holds and manipulates information in the service of ongoing cognition, whose control component overlaps heavily with executive function.

Key Researchers

Akira Miyake. University of Colorado Boulder; led the latent-variable analysis that established the unity and diversity of executive functions. Scholar · Faculty

Naomi P. Friedman. University of Colorado Boulder; co-developed the unity/diversity framework and its behavioral-genetic extension. ORCID · Scholar · Faculty

Adele Diamond. University of British Columbia; a founder of developmental cognitive neuroscience who synthesized the core executive functions and their training. ORCID · Scholar · Faculty

Alan Baddeley. University of York; proposed the central executive of working memory, the attentional-control component later analyses sought to decompose. ORCID · Scholar · Wikipedia

Earl K. Miller. Massachusetts Institute of Technology; co-authored the integrative theory of prefrontal cognitive control through top-down bias. ORCID · Scholar · Faculty

John Duncan. MRC Cognition and Brain Sciences Unit, Cambridge; characterized the multiple-demand system and its link to fluid intelligence. Scholar · Faculty · Wikipedia)

Tim Shallice. University College London and SISSA; co-developed the Supervisory Attentional System model of executive control. Scholar · Wikipedia

Russell A. Barkley. Virginia Commonwealth University; advanced the theory that ADHD is fundamentally a disorder of behavioral inhibition and executive function. ORCID · Scholar · Wikipedia

Frequently Asked Questions

What is executive function in simple terms?
It is the brain's management system, the set of control processes that let a person plan, focus attention, resist distraction, hold information in mind, and adjust when a situation changes, rather than acting on habit alone (Diamond, 2013).

What are the three core executive functions?
Most researchers identify inhibition (suppressing a prepotent response), working-memory updating (monitoring and revising what is held in mind), and cognitive flexibility or shifting (moving between tasks and rules) (Miyake et al., 2000).

Is executive function one ability or several?
It is both at once: latent-variable studies show the components are correlated, pointing to a shared resource, yet remain statistically separable, a pattern called unity and diversity (Miyake & Friedman, 2012).

Which part of the brain controls executive function?
Executive control depends chiefly on the prefrontal cortex, which biases processing elsewhere in the brain toward goal-relevant information, working together with a wider frontoparietal multiple-demand network (Miller & Cohen, 2001).

How is executive function measured?
It is assessed with tasks that pit a goal against a habit or demand rapid reconfiguration, such as the Stroop task for inhibition, the n-back for updating, and cued task switching for flexibility (Stroop, 1935).

When does executive function develop?
It develops slowly, emerging in infancy and improving markedly across childhood and adolescence into early adulthood, in step with the prolonged maturation of the prefrontal cortex (Best & Miller, 2010).

Can executive function be improved?
Yes; reviews of interventions find that curricula, computerized training, physical activities, and other approaches can improve executive function in children, with the largest gains for those who begin with the weakest skills (Diamond & Lee, 2011).

How does executive function relate to ADHD?
An influential theory holds that ADHD is fundamentally a deficit of behavioral inhibition that in turn impairs the executive functions depending on it, including working memory and self-regulation (Barkley, 1997).

References