Can Someone Fake Dementia?

People can and do attempt to fake dementia, and the motivations range from avoiding criminal prosecution to securing disability benefits. But pulling it off convincingly is harder than most people assume. Neuropsychologists have developed a battery of specialized tests designed to catch inconsistencies that even well-prepared fakers struggle to avoid, and newer brain-imaging techniques are pushing detection rates even higher. The real complexity, though, lies not in catching obvious fakers but in the murky territory where genuine cognitive decline, involuntary psychological conditions, and deliberate deception overlap.

Why Someone Would Try

The most common setting for feigned dementia is forensic. A defendant facing serious criminal charges has a powerful incentive to appear mentally incapacitated, since being found incompetent to stand trial can delay or derail prosecution entirely. Civil cases matter too: people seeking disability payments, insurance settlements, or release from contractual obligations sometimes exaggerate or fabricate cognitive problems. In elder law, questions about whether a person was “of sound mind” when signing a will or power of attorney can create incentives for exaggeration in either direction.

The impulse to fake mental incapacity has deep roots. A comprehensive study of classical texts identified roughly thirty instances of feigned madness in the Greco-Roman world, arguing that faking cognitive or mental dysfunction served as a strategy of personal and political resistance, particularly among people with limited power who wanted to avoid harm, speak freely, or dodge unwelcome obligations.¹ The contexts have changed, but the basic logic persists: appearing incapable is sometimes safer or more advantageous than appearing competent.

How Clinicians Catch It

The primary tools for detecting faked dementia are performance validity tests, or PVTs. These are cleverly designed assessments that look difficult but are actually easy enough that even people with genuine moderate cognitive impairment pass them. The logic is simple: if someone with real brain damage can manage the task, a person faking should be able to as well, unless they are deliberately underperforming.

The most widely used PVT is the Test of Memory Malingering, known as the TOMM. It asks the person to look at a series of pictures and then identify which ones they saw. The task is designed so that even people with significant memory problems score well above chance. Healthy adults almost never fail it. In validation studies, the TOMM’s recommended cutoff scores on its second trial produced overall classification accuracy of 96%, with high sensitivity and specificity for identifying people who were deliberately underperforming.² Shortened versions of the TOMM have also proven highly accurate, with one short form achieving an area under the curve of 92%, meaning it correctly sorted genuine from faked performance the vast majority of the time.³

Another approach uses even simpler tasks. The Coin-in-the-Hand test, for example, asks the person to remember which hand holds a coin after a brief delay. The task is so easy that people with genuine neurological disorders generally perform no differently from healthy controls. Suspected malingerers, by contrast, tend to perform at or below chance level, essentially doing worse than random guessing, which is a statistical red flag since getting fewer answers right than flipping a coin strongly suggests deliberate wrong answers.⁴

This below-chance pattern is one of the strongest giveaways. To consistently score below 50% on a two-choice test, you actually have to know the right answer and choose the wrong one. Genuine ignorance produces scores that hover around chance, not below it. Clinicians trained to spot this pattern can often identify faking from the test scores alone.

Does Studying Up Help Fakers?

With information about dementia symptoms and neuropsychological testing widely available online, a reasonable question is whether someone who does their homework can beat these tests. The research here is somewhat reassuring for clinicians, but not entirely.

Studies that compare different types of coached fakers typically divide them into groups: people told to fake with no guidance, people coached on what dementia symptoms look like, and people coached on both symptoms and test-taking strategies to avoid detection. The third group, unsurprisingly, is the hardest to catch. One study found that naïve simulators and symptom-coached groups were the easiest to detect, while the group coached on both symptoms and test strategies was the hardest. Some tests lost much of their detection power against this well-coached group.⁵

The TOMM, however, appears to hold up relatively well against coaching. When researchers compared symptom-coached and test-coached simulators head to head, the TOMM’s standard cutoffs still achieved 96% overall classification accuracy regardless of coaching type. The symptom-coached group actually performed more poorly on the test than the test-coached group, but the TOMM was equally sensitive to both.² Other instruments, like the CARB-97 and the Word Memory Test, showed similar robustness, with little difference in detection between naïve and coached fakers.⁶

The takeaway is that coaching helps fakers avoid the most obvious mistakes, like performing ridiculously below chance, but it does not reliably defeat a well-administered battery of tests. The strongest tests exploit cognitive principles that are difficult to game even with preparation, because the faker does not know exactly where the threshold between “plausible impairment” and “suspicious performance” falls.

When Real Dementia Looks Like Faking

Here is where the picture gets uncomfortable. The same tests designed to catch fakers can also flag people with genuine, severe cognitive impairment. This is the most significant limitation in the field, and it has real consequences for patients who are genuinely sick.

A study of older adults across the cognitive spectrum found that the proportion failing two or more PVTs climbed sharply with severity: about 2% among cognitively normal participants, roughly 7% among those with general cognitive impairment, 13% among those with mild cognitive impairment, and a striking 53% among those with dementia.⁷ In other words, more than half of people with genuine dementia produced validity test scores that looked like faking under standard criteria.

The problem is especially acute in certain types of dementia. Research on patients with early-onset dysexecutive Alzheimer’s disease found that nearly half performed below recommended cutoffs on the TOMM’s first and second trials. Several of these patients had actually been misdiagnosed with other conditions partly because their TOMM scores were interpreted as evidence of unreliable effort rather than genuine impairment.⁸ That finding is a sobering reminder that a failed validity test is not proof of faking. It is one piece of information that has to be weighed against everything else clinicians know about the person.

Research across dementia clinics confirms this complexity. When PVTs validated on non-dementia samples were applied to people with varying degrees of cognitive decline, the standard cutoffs frequently needed adjustment to maintain adequate specificity. For mild cognitive impairment and mild dementia, slight modifications were enough. For moderate dementia, more significant changes were required.⁹ Clinicians who apply cookie-cutter cutoffs without considering how impaired the person already is risk labeling sick patients as fakers.

Brain Signals That Give Fakers Away

Because behavioral tests have these limitations, researchers have looked for physiological measures that are harder to fake. The most promising avenue involves brain wave patterns called event-related potentials, specifically a component known as P3b. This electrical signal shows up when the brain recognizes something meaningful or familiar. The key insight is that a person who genuinely does not recognize a stimulus will not produce a strong P3b response, while someone who secretly recognizes it but is pretending not to will.

In a study comparing people with probable Alzheimer’s disease against healthy adults instructed to simulate dementia, combining the TOMM’s second-trial score with the P3b peak amplitude achieved 100% sensitivity and 90% specificity in distinguishing the two groups. That was a noticeable improvement over using the TOMM score alone, which managed 95% sensitivity and 80% specificity.¹⁰ The brain wave data caught some of the fakers that the paper-and-pencil test missed.

Related work on dissociative amnesia, a condition where people lose access to personal memories without any detectable brain damage, found that P300 responses indicated the presence of stored memory information even when patients reported complete amnesia. The brain’s electrical signature showed recognition of autobiographical material that the patients claimed not to remember.¹¹ While dissociative amnesia is not the same as deliberately faked dementia (it is typically involuntary), the principle is the same: the brain retains information that the person’s behavior does not reveal, and electrophysiology can expose the mismatch.

These brain-based methods are not yet standard in clinical practice. They require specialized equipment and expertise, and the research samples have been small. But they represent a direction that makes faking progressively harder, because you can learn to control your answers on a test but you cannot voluntarily suppress an electrical response your brain generates in milliseconds.

Functional Cognitive Disorder and the Gray Zone

Not everyone who appears to fake dementia is doing so on purpose. Functional cognitive disorder, or FCD, is a condition in which people experience genuine, distressing cognitive symptoms like memory failures and difficulty concentrating, but without the underlying brain pathology seen in Alzheimer’s or other neurodegenerative diseases. These people are not lying. Their symptoms are real and involuntary. But on formal testing, their profiles can look confusingly similar to both early neurodegeneration and deliberate faking.

A clinic-based study comparing FCD patients with those who had neurodegenerative mild cognitive impairment found that both groups showed similar rates of sub-optimal performance on validity indicators built into personality assessments. FCD patients also scored relatively high on measures of symptom over-reporting compared with healthy controls, though their over-reporting rates were statistically indistinguishable from the neurodegenerative group.¹² This creates a diagnostic challenge: the performance validity measures that are supposed to distinguish “real” from “fake” impairment flag FCD patients at similar rates to those with genuine neurodegeneration.

The existence of FCD matters for anyone thinking about feigned dementia because it highlights that the binary of “real disease versus deliberate faking” is too simple. A person can have authentic cognitive complaints, perform poorly on validity tests, show no signs of neurodegeneration, and still not be malingering. Clinicians who treat every failed validity test as evidence of faking risk harming a vulnerable population that needs treatment, not suspicion.

The Multidimensional Approach to Malingering Detection

Because no single test reliably separates genuine dementia from faking across all populations, the field has moved toward multidimensional frameworks. Updated criteria for identifying malingered neurocognitive dysfunction emphasize that clinicians should not rely on a single failed PVT. Instead, they should integrate multiple data streams: the presence of an external incentive (a lawsuit, a disability claim, a criminal charge), inconsistencies between test performance and observed daily functioning, results across multiple validity measures, and the exclusion of medical or psychiatric conditions that could explain poor performance.¹³

The updated criteria also expand beyond purely cognitive faking to include somatic, psychiatric, and mixed symptom presentations. This matters because someone faking dementia often does not fake only memory loss. They may also exaggerate physical symptoms, report psychiatric distress, or produce a blend of complaints that does not map neatly onto any known disease pattern. A multidimensional assessment is better positioned to detect these inconsistencies than any one memory test administered in isolation.

In practice, this means a thorough malingering evaluation in a forensic context might include several standalone PVTs, validity indicators embedded within standard cognitive tests, self-report symptom inventories with built-in over-reporting scales, review of medical records for consistency, collateral interviews with family or caregivers, and possibly neuroimaging or electrophysiology. The goal is convergence: a confident determination that someone is faking requires multiple independent lines of evidence pointing the same direction.

What Fakers Tend to Get Wrong

People who try to fake dementia usually base their performance on a popular stereotype of the condition rather than on how dementia actually presents. A few common mistakes stand out to experienced clinicians.

First, fakers tend to overdo it. They fail tasks that even severely impaired patients pass, like recognizing their own name or identifying common objects. Genuine dementia is typically patchy and follows recognizable patterns; memory for recent events degrades before memory for deeply encoded personal information. Someone who cannot recall whether they ate breakfast but also claims not to know their own birthday is raising red flags.

Second, fakers often produce inconsistent behavior across contexts. A person who cannot complete a simple memory test in the examiner’s office but navigated public transit to get to the appointment, or who engages in fluent conversation in the waiting room before appearing confused during testing, presents a pattern that does not fit with organic disease. Clinicians are trained to watch for these discrepancies and to note them as part of the overall assessment.

Third, there is the below-chance performance pattern mentioned earlier. People faking memory loss sometimes try so hard to appear impaired that they get fewer items right than random guessing would produce. This is mathematically telling: true ignorance leads to chance-level performance, not systematic avoidance of correct answers. A score significantly below 50% on a forced-choice recognition test is stronger evidence for faking than a score of zero on a free-recall test, because it implies the person knew the answers and deliberately avoided them.

Dementia Feigning in Everyday Life Versus Forensic Contexts

Most of the research on faked dementia comes from forensic settings, where the stakes are high and standardized assessment is routine. But the question of whether someone might fake dementia comes up in other contexts too, and the dynamics are different.

In families, concerns sometimes arise when an older relative’s cognitive complaints seem inconsistent or when they appear sharp in some situations but impaired in others. It is worth knowing that this kind of variability is actually a hallmark of many genuine cognitive conditions, including early Alzheimer’s, vascular dementia, and Lewy body dementia, where fluctuating alertness and attention are core features. Inconsistency alone is not evidence of faking.

Caregivers and family members sometimes worry about the opposite scenario: that a relative is hiding or minimizing cognitive decline. This is far more common than faking it. Anosognosia, or lack of awareness of one’s own deficits, is a recognized feature of many dementias and leads people to genuinely believe they are functioning better than they are. So while someone can certainly fake dementia, the more frequent clinical reality is people whose dementia is real but whose self-report understates the severity.

In institutional settings like nursing homes or assisted living, staff occasionally wonder whether a resident is exaggerating cognitive symptoms to receive more attention or assistance. Formal malingering assessment is rarely performed in these settings because the context does not typically involve external incentives like legal or financial gain. When a resident seems to fluctuate, clinical explanations like delirium, medication effects, depression-related cognitive impairment, or the natural variability of the underlying disease are far more likely and should be explored before anyone starts suspecting faking.

The research is clear that faking dementia is possible but difficult to sustain under scrutiny. The people most likely to attempt it are those with clear external incentives and a specific event to get through, like a competency hearing or disability evaluation. Sustained, day-to-day faking of dementia over months or years, fooling the people who live and interact with you constantly, is a qualitatively different challenge. The cognitive demands of consistently performing a plausible dementia simulation while also managing your actual life are, ironically, considerable.