Delis-Kaplan Executive Function System (D-KEFS)

The Delis-Kaplan Executive Function System, usually called the D-KEFS, is a battery of nine individually administered tests designed to measure the higher-order thinking skills that psychologists group under the umbrella of “executive functions.” Published in 2001 by Dean Delis, Edith Kaplan, and Joel Kramer, it was the first set of executive function tests normed together on a single large, nationally representative sample, covering people from age 8 through 89.¹ That shared normative base is a big part of what made it influential: instead of comparing a person’s performance on one test to one norm group and their performance on another test to a different norm group, clinicians could compare all nine scores against the same reference population.

What Executive Functions the D-KEFS Measures

Executive functions are the mental processes that let you plan, stay flexible, resist impulses, shift strategies when something isn’t working, and think abstractly. The D-KEFS was built to assess a wide range of these abilities, including flexibility of thinking, inhibition, problem solving, planning, impulse control, concept formation, abstract thinking, and creativity.² That’s a broader slice of executive functioning than most older tests tried to capture. Rather than giving you a single “executive function score,” the D-KEFS generates a profile across multiple distinct skills, which is more useful for figuring out exactly where someone’s strengths and weaknesses lie.

The Nine Tests and What They Tap

Each of the nine D-KEFS tests can be given on its own or as part of the full battery. A clinician might use just two or three if time is limited or if a specific question needs answering, or they might run the entire set for a comprehensive picture. Here is what each one is designed to evaluate:

Trail Making Test: A modified version of the classic trail-making task. It includes conditions for visual scanning, number sequencing, letter sequencing, motor speed, and a letter-number switching condition that isolates the ability to shift between mental sets.³
Verbal Fluency Test: Measures how quickly and flexibly someone can generate words, both from letter cues and from categories, with a category-switching condition that adds a set-shifting demand.
Design Fluency Test: A nonverbal counterpart to verbal fluency, requiring the person to draw novel designs under varying rule constraints.
Color-Word Interference Test: A modern take on the Stroop test. It includes color naming, word reading, an inhibition condition where you name the ink color while ignoring the printed word, and a fourth condition that adds switching on top of the inhibition demand.⁴
Sorting Test: Asks the person to sort cards into groups based on conceptual categories they have to figure out, then describe the sorting rules they used.
Twenty Questions Test: Based on the classic guessing game, it measures the ability to form abstract yes-or-no questions and narrow down possibilities efficiently.
Word Context Test: Requires figuring out the meaning of a made-up word from clues embedded in sentences, tapping into deductive reasoning.
Tower Test: A problem-solving task that involves moving disks across pegs to reach a target arrangement in the fewest moves, following specific rules.
Proverb Test: Assesses abstract verbal reasoning by asking the person to interpret common and uncommon proverbs.

The Baseline-Condition Design

One of the cleverest aspects of the D-KEFS is how it isolates executive skills from the more basic abilities those skills depend on. Consider the Trail Making Test. The classic version has two parts: connecting numbers in order (Part A) and alternating between numbers and letters (Part B). If someone performs poorly on Part B, it could mean they struggle with set-shifting, or it could mean they’re just slow at scanning the page or writing. The D-KEFS version addresses this by including separate baseline conditions for visual scanning, number sequencing, letter sequencing, and motor speed. By comparing performance on the switching condition to these baselines, the clinician can see whether a low score genuinely reflects an executive problem or is better explained by slow processing speed or poor motor control.⁵

The Color-Word Interference Test follows the same logic. The first two conditions measure basic color naming speed and word reading speed. The third condition adds the inhibition demand of the Stroop task. The fourth condition layers switching on top of inhibition, requiring the person to alternate between naming ink colors and reading words at irregular intervals.⁶ A person who is slow across all four conditions likely has a speed problem, not an inhibition problem. But someone who handles conditions one and two fine and then struggles on three and four is showing something specific about their ability to suppress automatic responses or shift between rules. That kind of layered comparison is what makes the D-KEFS more diagnostically informative than simpler versions of the same tasks.

How the D-KEFS Is Used After Traumatic Brain Injury

Traumatic brain injury is one of the conditions where the D-KEFS has been studied most extensively, and the results generally support its clinical usefulness. People with TBI tend to score lower on several D-KEFS subtests, including the Trail Making Test, Color-Word Interference, and Verbal Fluency, and those scores discriminate between TBI patients and both orthopedic injury patients and healthy comparison groups with moderate to large effect sizes.⁷ The battery also shows a dose-response pattern: the more severe the injury, the worse the performance, which is what you’d want to see if a test is really measuring something meaningful about brain function.

Research has found that specific aspects of TBI severity relate to specific D-KEFS subtests. In one study of outpatients with mild to severe TBI, the length of time a patient was in a coma predicted their performance on the Color-Word Inhibition/Switching condition, while the presence of diffuse brain lesions on neuroimaging was more related to Verbal Fluency Category Switching scores.⁸ Patients with moderate-to-severe TBI performed significantly worse on both switching subtests than those with mild uncomplicated TBI or healthy controls, and the mild group didn’t differ from controls. The combined classification accuracy of these subtests was around 66%, with the researchers describing the clinical utility as modest but real.

A study of Iranian TBI patients with frontal contusions found especially dramatic results: compared with a matched normative sample, the patients showed substantial impairments on all executive measures from the D-KEFS, with very large effect sizes. Even after accounting for performance on the lower-level baseline conditions, the differences on the highest-level switching conditions remained significant.⁹ This reinforces the point that the D-KEFS switching conditions are picking up genuine executive difficulties, not just generalized slowness.

Distinguishing Types of Dementia

Different dementias damage different brain regions and produce different cognitive profiles, and some of the most interesting D-KEFS research involves teasing those profiles apart. On the Tower Test, both frontotemporal dementia (FTD) and Alzheimer’s disease (AD) groups show poorer overall scores than healthy controls. But the FTD group commits significantly more rule violations than both healthy controls and the AD group, suggesting worse performance monitoring.¹⁰ That distinction matters clinically because it points toward frontal-lobe dysfunction, which is the hallmark of FTD, rather than the memory-system breakdown that characterizes early Alzheimer’s.

A confirmatory factor analysis of the Tower Test itself found that it captures three distinct dimensions: first-move time (how long someone pauses before starting), excess moves (how many unnecessary moves they make), and rule violations.¹¹ Someone who rushes in quickly but makes lots of excess moves is showing a different problem than someone who takes a long time to start but then executes efficiently. And someone who repeatedly breaks the rules is showing yet another kind of difficulty. That three-way split is diagnostically useful because different conditions tend to produce different patterns.

The Proverb Test also helps distinguish FTD from Alzheimer’s. Patients with the behavioral variant of FTD performed significantly worse than AD patients at interpreting proverbs, producing more concrete, literal interpretations. Interestingly, this difference was most pronounced for common, familiar proverbs rather than obscure ones, suggesting that the FTD patients weren’t failing because the proverbs were unfamiliar but because their ability to extract abstract meaning was compromised.¹²

Schizophrenia and Executive Function Profiles

Executive function deficits in schizophrenia are well documented, and researchers have used the D-KEFS to get a more fine-grained picture of what those deficits look like. Factor analyses of D-KEFS performance in people with schizophrenia have consistently yielded two main components. One study found factors labeled “inhibition/set shifting” and “mental flexibility,” with the important finding that different symptom dimensions of schizophrenia mapped onto different executive weaknesses. Both negative symptoms (like flat affect and social withdrawal) and cognitive symptoms were linked to the inhibition/set shifting factor, while cognitive symptoms alone predicted the mental flexibility factor. Positive symptoms like hallucinations and delusions weren’t related to either executive component.¹³

Another study using the D-KEFS in schizophrenia identified two factors as well, which they labeled “cognitive flexibility/timed tests” and “abstraction.” After statistically removing the influence of processing speed, the cognitive flexibility factor was still distinct from abstraction, meaning the two aren’t just reflections of how fast someone thinks. Better cognitive flexibility was negatively correlated with overall psychopathology, while better abstraction was associated with higher education, shorter illness duration, and better real-world functional capacity.¹⁴ This sort of detailed breakdown is useful for treatment planning because it suggests that targeting different executive skills might matter depending on a patient’s symptom profile.

ADHD, Autism, and Developmental Conditions

The D-KEFS is widely used in evaluating children and adults with ADHD and autism spectrum disorder, both of which involve executive function difficulties. A broad meta-analysis of studies directly comparing executive function in ASD and ADHD found that both groups perform worse than typically developing individuals on measures of attention, flexibility, working memory, processing speed, and response inhibition, with no consistent difference between the two diagnostic groups overall.¹⁵ That overlap is part of what makes differential diagnosis difficult.

Preliminary research with the D-KEFS Advanced Trail Making Test, a newer version of the original, suggests the test may be sensitive to some performance differences between these groups that broader measures miss. In one study, individuals with ADHD scored significantly higher than those with ASD on both the Number Sequencing and Switching-Distraction conditions of the advanced trail-making task.¹⁶ These are early findings from a small study, so they need replication, but they hint at the potential for more refined executive-function tests to help differentiate conditions that look similar on broader cognitive screens.

How the D-KEFS Compares to Older Executive Function Tests

Before the D-KEFS existed, the Wisconsin Card Sorting Test (WCST) was the go-to measure of executive function in clinical neuropsychology. The WCST remains widely used, but the D-KEFS offers some practical advantages. When researchers compared the Sorting Test from the D-KEFS with the WCST and the Trail Making Test in patients with severe TBI, all three instruments significantly differentiated patients from healthy controls. The D-KEFS Sorting Test produced significant group differences on all five of its measured variables, compared with four out of six for the WCST.¹⁷ In a receiver operating characteristic analysis, the Sorting Test’s “attempted sorts” variable and the Trail Making Test Part B completion time were the strongest predictors of group membership.

In people with multiple sclerosis, both the WCST and the D-KEFS Sorting Test showed good validity, but the availability of alternate forms on the D-KEFS was flagged as a practical advantage. Alternate forms matter because clinicians often need to retest patients over time to track progression or recovery, and using the same version twice can produce inflated scores due to practice effects.¹⁸

What Brain Regions Drive D-KEFS Performance

The D-KEFS was designed with frontal-lobe functioning in mind, and neuroimaging research generally supports a link between frontal regions and performance on several of the tests. A quantitative MRI study found that, after controlling for overall cognitive status and demographic variables, only the volume of the left frontal lobe significantly predicted performance on the Sorting Test.¹⁹ This fits the theoretical expectation that concept formation and flexible categorization depend heavily on frontal-lobe circuits.

The relationship between brain location and D-KEFS performance is not always straightforward, though. A study comparing patients with focal lesions in different parts of the prefrontal cortex found that those with dorsolateral prefrontal damage tended to score lower on several D-KEFS indices than patients with ventromedial prefrontal or non-frontal lesions. However, once full-scale IQ and processing speed were statistically controlled, none of those group differences remained significant.²⁰ The authors interpreted this as suggesting that the weaknesses seen in the dorsolateral group weren’t specific to executive functions per se but rather reflected broader cognitive efficiency. That finding is a reminder that executive function tests, including the D-KEFS, don’t purely measure frontal-lobe processes in isolation. General intelligence and processing speed always play some role, and the degree to which any test captures “pure” executive function is an ongoing debate in the field.

Cross-Cultural Considerations and Normative Limits

The original D-KEFS norms were developed on a U.S. sample, and research from other countries suggests those norms don’t always translate well. A Slovak study that tested children from different socioeconomic and ethnic backgrounds found significant differences in performance on D-KEFS primary indicators, leading the researchers to argue that country-specific and even population-specific norms are necessary.²¹ Roma children from socially disadvantaging environments scored differently from the standard population of Slovak children, highlighting that performance on executive function tests is shaped by education, language, and environment, not just underlying cognitive ability.

Researchers in Greece have developed normative data for the Color-Word Interference and Trail Making subtests for Greek adults, reflecting the same concern.²² Similar adaptation work has been done in Iran and other countries. The practical implication is that applying U.S.-based norms to someone raised in a very different educational or cultural context can produce misleading results. Clinicians working with diverse populations are increasingly aware that they need either locally validated norms or, at minimum, careful clinical judgment about how cultural and linguistic factors might affect scores.

Education level also matters within populations. Research on the D-KEFS Trail Making Test found that correcting for both education and vocabulary level improved the accuracy of normative comparisons, since people with more education tend to perform faster on the sequencing and switching conditions for reasons that have nothing to do with executive dysfunction.²³ Without those corrections, a clinician might mistake lower education for impaired executive function, or overlook genuine impairment in a highly educated person whose scores look “normal” but are actually a significant decline from their baseline.

Embedded Performance Validity

An issue that comes up in clinical and forensic settings is whether a person is putting forth genuine effort during testing. The D-KEFS Color-Word Interference Test is among the most commonly administered measures of verbally mediated processing speed and executive functioning, and researchers have explored whether its score patterns can serve as embedded indicators of effort.²⁴ The idea is that certain patterns of performance across the test’s conditions are unlikely to occur in someone genuinely trying their best, even if they have real cognitive impairment. When someone performs worse on the easy baseline conditions than on the harder executive conditions, for example, that’s a red flag. This line of research means the D-KEFS can pull double duty, measuring executive function while also flagging cases where the results might not be trustworthy.

The D-KEFS Advanced

The original D-KEFS has been the standard version for over two decades, but a newer edition, the D-KEFS Advanced, has started appearing in research and practice. The Advanced Trail Making Test, for example, adds conditions and scoring features beyond what the original offered, including distraction conditions and more granular timing measures like mean connection times for individual segments. Early clinical research using the Advanced TMT has focused on whether these additional metrics sharpen the ability to differentiate diagnostic groups, such as ADHD versus ASD.²⁵ Whether the Advanced version will replace the original in routine clinical practice depends on how the validation studies play out over the next several years and whether the added complexity translates into meaningfully better diagnostic decisions for the clinicians using it.