• Home   /  
  • Archive by category "1"

Fregoli Syndrome Case Study

Original Article

The Fregoli Delusion: A Disorder of Person Identification and Tracking


  • Robyn Langdon,

    Corresponding author
    1. ARC Centre for Excellence in Cognition and Its Disorders (CCD), Department of Cognitive Science, Macquarie University
    • Correspondence should be sent to Robyn Langdon, Department of Cognitive Science, Macquarie University, NSW 2019, Australia. E-mail: robyn.langdon@mq.edu.au

    Search for more papers by this author
  • Emily Connaughton,

    1. ARC Centre for Excellence in Cognition and Its Disorders (CCD), Department of Cognitive Science, Macquarie University
    Search for more papers by this author
  • Max Coltheart

    1. ARC Centre for Excellence in Cognition and Its Disorders (CCD), Department of Cognitive Science, Macquarie University
    Search for more papers by this author


Fregoli delusion is the mistaken belief that some person currently present in the deluded person's environment (typically a stranger) is a familiar person in disguise. The stranger is believed to be psychologically identical to this known person (who is not present) even though the deluded person perceives the physical appearance of the stranger as being different from the known person's typical appearance. To gain a deeper understanding of this contradictory error in the normal system for tracking and identifying known persons, we conducted a detailed survey of all the Fregoli cases reported in the literature since the seminal Courbon and Fail (1927) paper. Our preliminary reading of these cases revealed a notable lack of definitional clarity. So, we first formulated a classification scheme of different person misidentification delusions so as to identify those cases that qualified as instances of Fregoli according to the above characterization: the mistaken belief that a known person is present in the environment in a different guise to his or her typical appearance. We identified 38 clear cases of this type and set out to answer a series of questions motivated by current hypotheses about the origin of the Fregoli delusion. We asked whether the patients misidentified particular strangers, made reference to the misidentified known persons using wigs or plastic surgery (or other techniques to disguise their appearance), misidentified many different strangers or only one, showed other symptoms (in particular, other misidentification delusions), and made inferences about the motives of the known persons in disguise. We conclude by discussing the implications of our findings for current hypotheses concerning the origin of the Fregoli delusion.

1 Introduction

Humans have evolved different capacities to help navigate their social world: distinguishing between the agents and the objects in their environment, identifying conscious agents, recognizing conspecifics, identifying particular known persons, and tracking their persistence across time. Any of these capacities might be disrupted temporarily, causing, for example, transient failures to recognize a familiar person as he or she walks by in the street—something we all do from time to time. Brain damage or psychiatric illness might cause a more enduring, even permanent, disruption to any one, or combination, of these capacities. Studying these more enduring cases offers a unique opportunity to learn more about errors of agent tracking and identification, and thereby to better understand the normal systems for agent tracking. Our aim in this paper was to adopt this approach to advance understanding of the normal processes for person identification and tracking by reviewing the literature on a particular type of clinical error of person identification: the Fregoli delusion.

2 The Fregoli delusion

In 1927, Courbon and Fail described a case of delusion, in which a young Parisian woman believed that two Parisian actresses of the time (Sarah Bernhardt and Robine) pursued her closely. The two actresses were unrecognizable because they were “taking the form of people she knows or meets” (Ellis, Whitley, & Luauté, 1994, p. 134), including the strangers she saw in the street, doctors, friends, and previous employers. Courbon and Fail named this syndrome after the Italian actor and mimic Leopoldo Fregoli because of his extraordinary ability to impersonate people on stage. de Pauw, Szulecka, and Poltock (1987) described a similar case, an elderly lady who believed that her male cousin and his lady friend were disguising themselves with wigs and other methods and following her around.

These two cases involve the delusional misidentification of familiar persons disguised as others. The others in the environment (a stranger, an acquaintance, or a hospital staff member, for example) are perceived correctly as being physically unrecognizable strangers or relatively unfamiliar casual acquaintances—and yet are incorrectly identified as a different known person. For example, Courbon and Fail's patient believed that the Sarah Bernhardt whom she saw on the stage on some Sunday night was the very same person that she saw on the street (in the disguise of a stranger) on the following Monday.

In these cases, there is some distinctive error(s) in the normal cognitive system for identifying and tracking known persons from the perception of their physical form. These cases are thus very unlike those other, more generic, persecutory, and referential delusions that are common in schizophrenia, and which involve delusional beliefs about being monitored by generic others, such as the Central Intelligence Agency (CIA). The others in these latter cases are personally anonymous—some personally unknown group entity rather than a specific known person, as occurs in Fregoli (see, e.g., Langdon, McKay, & Coltheart, 2008, for further discussion of persecutory delusions).

3 Aim and approach

To advance understanding of this distinctive failure of person identification and tracking, we conducted a detailed survey of all Fregoli cases reported in the literature since the original 1927 paper. Our broader theoretical framework was the “two-factor” approach (Langdon & Coltheart, 2000). This approach distinguishes between the factors that explain the generation of specific delusional themes (and vary from delusion to delusion) and the factors that explain the failure to reject a delusional belief (and may be common to all delusions). Our specific objective was thus to gain insights about the former—the factor 1 disruption(s) to normal person identification and tracking that explain the generation of the Fregoli delusional content.

Our approach was to ask a series of questions motivated by current hypotheses about the origin of the Fregoli delusional content. First, though, we needed to check that the cases provided sufficient detail of the specific known person(s) being misidentified. Without such detail we could not be sure that a so-called Fregoli case was anything more than a generic persecutory delusion. Second, we needed to check that the cases provided some clear indication that the processing of the real person's physical form was reasonably intact. Without this sort of detail we could not be sure that a so-called Fregoli case was anything more than a generic misidentification delusion—specifically, one that also involves impairment of the processing of the real person's physical form. In the latter case, for example, the patient would believe that the real person in the environment was, and looked like, the misidentified known person. Therefore, we needed to see some indication(s) that the patient believed that the known person(s) appeared in a form unlike his or her typical appearance; for example, the patient may have referred to “disguise” or “plastic surgery.”

We also wanted to know more about the specific type(s) of mismatching of unfamiliar real persons (present in the environment) to misidentified known persons (not present in the environment) that occur. For example, we wondered: Do Fregoli patients misidentify only one stranger in the crowd (or one casual acquaintance) as one other known person in disguise—that is, do they show a “one-to-one mapping” of real stranger to misidentified known person? Or, instead, do they typically believe that one particular known person is duplicating him or herself in many different physical guises—that is, do they typically show a “many-to-one mapping”? Or even “many-to-many mappings”? The answers to these questions, we hoped, would help to evaluate current hypotheses concerning the error of normal person identification and tracking that causes the generation of the Fregoli delusional content.

4 Current hypotheses concerning the origin of the Fregoli delusion

Two of the current hypotheses concerning the origin of the Fregoli delusion posit some disruption in the normal cognitive system for identifying known persons from the perception of their faces. These hypotheses differ, however, with regard to the proposed sites of disruption.

Ellis and Young's (1990) account: Ellis and Young were the first to use a cognitive model of normal face processing—Bruce and Young's (1986) cognitive model—to offer an explanation not only of Fregoli delusion but also Capgras delusion, the delusional belief that someone emotionally close—often a spouse—has been replaced by a visually similar impostor.

Fig. 1 illustrates a recent modification of their model, developed primarily to explain both Capgras delusion and prosopagnosia. People with prosopagnosia are unable to explicitly identify known people from the perception of their faces. Yet some prosopagnosic patients (not all) implicitly discriminate between the faces of people they do and do not know. Specifically, they show higher skin conductance responses (SCRs—an index of autonomic arousal) when viewing the faces of people they know than when viewing the faces of people they do not know (see Ellis & Young, 1990; for further discussion). Their problem is attributed to a breakdown at point (1) in Fig. 1. This breakdown leaves these individuals unable to access the “person identity nodes” (or PINs) that would normally allow access to known people's names and related personal information; hence, the patients cannot identify the faces he or she sees. In contrast, Capgras patients are proposed to have a breakdown at point (2) in Fig. 1. This breakdown results in the “face recognition units” (i.e., the FRUs, or stored representations of known people's faces) failing to trigger normal autonomic arousal to a familiar face. This isolated breakdown explains why Capgras patients can access the appropriate information to name whose face the “stranger's” face looks like, yet deny that the person in front of them, with the look-alike face, but generating no autonomic arousal, is the loved one.

In contrast, Ellis and Young attributed the generation of the Fregoli delusional content to a problem with the PINs and the associated “cognitive system” (shaded in gray in Fig. 1).1 Their general idea was that the cognitive system (including who a person is currently thinking about) overexcites certain PINs resulting in a particular PIN being sufficiently activated to identify a known person as present, even when the person is not. To illustrate, Courbon and Fail's patient had a passion for the theater, even preferring to go without food to attend a performance. So this patient's cognitive system was continually activating the PINs of her favorite actresses, Sarah Bernhardt and Robine. In that case, any additional input from the processing of a real person present in the environment—say from perceiving a stranger wave her arms just like Sarah Bernhardt2 or partial activation of Sarah Bernhardt's FRU because of some visual similarity with a stranger's face—might have been sufficient to tip the PIN over the threshold to identify Sarah Bernhardt as present, even when a matching face was not also present.

Ellis and Young were less specific about why one particular stranger in the crowd was misidentified as the known person. They did suggest, however, that assessing similarities (e.g., between Sarah Bernhardt and the real stranger) would be of theoretical value.

Ramachandran and Blakeslee's (1998) account: Ramachandran and Blakeslee proposed that a different sort of disruption to face processing causes the generation of the Fregoli delusional content. They hypothesized that Fregoli patients have hyperactive temporo-limbic connections between the facial recognition systems and the amygdala system for affective arousal. In other words, Fregoli patients, like Capgras patients, have some problem at point (2) in Fig. 1—say, inappropriately overexcited inputs to the affective response system, or, perhaps, an overactive affective responsiveness system itself. The authors were not specific about how this overexcitation occurs, focusing more on the neuropathology they believed to be involved. Nevertheless, their general idea was that, whereas in Capgras the problem is underactivation of normal autonomic arousal, the problem in Fregoli is excess inappropriate arousal to viewing unfamiliar faces. As with Ellis and Young's account, this excessive autonomic arousal is experienced in the absence of conscious recognition of any matching known faces in the environment, resulting in the belief that some known person is appearing in disguise. Why one (or more) particular known person is believed to be present is not explained, however.

Thus far, these two hypotheses suggest that some disruption in the normal face-processing system causes the generation of the Fregoli delusional content. That disruption is somewhat circumscribed, however, since Fregoli patients are aware of the mismatch between the perceived form of the stranger and the known person's usual appearance. So we wondered: Do Fregoli studies report the results of any cognitive testing, particularly testing of patients' face processing?

As for the predictions that these two accounts make about the numbers of real persons and numbers of mistaken known identities, any Fregoli cases involving one-to-one mappings (i.e., one real person misidentified as one different known person), if found, would be less consistent with Ramachandran and Blakeslee's account, because hyperactive connections between the facial recognition systems and the amygdala would be expected to continually trigger the misidentifications of strangers in the environment (i.e., multiple real persons being misidentified as one or more known others). Ramachandran and Blakeslee's account also predicts that Capgras and Fregoli delusions should not co-occur, because underactive and overactive autonomic arousal to faces cannot occur simultaneously. So we also recorded any details of co-occurring symptoms, in particular, other misidentification delusions.

Other questions were motivated by a different account of Fregoli, from Hirstein (2005).

Hirstein (2005): Hirstein proposed that misidentification delusions are caused by emotional “mind-reading” deficits that cause mismatches between the “internal representations” of other people (who they are “on the inside,” including their characteristic beliefs, desires, attitudes, traits, and emotional states) and the “external representations” of others' physical appearances. In Capgras patients, for example, the patients' external representations are intact but their internal representations are damaged or inaccessible. So Hirstein proposed, they pair up an impoverished, generic internal representation with the external representation of a known person. For example, Hirstein and Ramachandran (1997) reported a Capgras patient who believed that his father was an impostor, referring to him generically as “that nice Jewish gentleman.” In contrast, in Fregoli patients, the internal representations of known people are inappropriately paired with external representations of strangers in the street or of casual acquaintances.

Mind-reading deficits are traditionally demonstrated empirically by difficulties with going beyond people's observed behavior to infer their inner lives, specifically the beliefs, feelings, and motivations that explain their observed behavior (see, e.g., the literatures on “theory of mind” impairment in autism and schizophrenia). So we wondered: Are Fregoli patients able to use their mind-reading skills to generate mental-state explanations of why the known person(s) is adopting a disguise?

In sum, our objectives were as follows: (a) To first check whether the reported Fregoli cases provided sufficient detail of specific unfamiliar persons being misidentified as specific known persons. This was necessary to rule out a generic persecutory delusion. (b) Next, we needed to check that the case descriptions provided sufficient indication that the processing of the physical forms of the real persons was intact, or intact enough. This was necessary to rule out more generic misidentification delusions. In addition, we sought to answer a number of questions: (c) Do Fregoli patients show face-processing impairment on cognitive testing? (d) How many strangers and known persons are misidentified? (e) Does Fregoli delusion ever co-occur with other misidentification delusions, particularly Capgras delusion? And (e), what inferences, if any, do Fregoli patients generate to explain the motives of the known persons in disguise?

5 Undertaking the survey

We searched the Psycinfo, Pubmed, and Web of Knowledge databases using the terms “Fregoli” and “Fregoli-like,” excluding articles not written in English and papers that did not provide actual case details (e.g., reviews).3 In this way, we identified 63 papers reporting 72 different cases. Our preliminary reading of these cases revealed a notable lack of definitional clarity. Thus, we needed to develop a classification scheme of different person misidentification delusions to identify those cases that qualified as instances of Fregoli according to de Pauw et al.'s (1987) characterization: “the delusional misidentification of familiar persons disguised as others” (p. 433). To that end, we focused on distinguishing between different combinations of familiar versus unknown persons being misidentified (psychologically and/or physically) as other familiar versus unknown persons. Table 1 summarizes our scheme: column 1 lists whether the real person(s) are either familiar, including oneself, or strangers; column 2 describes the mistaken beliefs about the real persons; column 3 lists the term commonly used in the literature to describe the specific delusional content; and column 4 provides a relevant case.4

1Person emotionally close to you (e.g., a spouse)13It's a stranger.14 Acknowledged as physically identical to, or very similar to, the real person but believed to be psychologically different and unfamiliar; hence, an impostorCapgras delusionCapgras and Reboul-Lachaux (1923), translated by Ellis et al. (1994)
2Yourself (reflected in a mirror)It's some strangerMirrored-self misidentificationBreen, Caine, and Coltheart (2001) report two cases of this type
3YourselfTransforming into another known personReverse intermetamorphosisBreen, Caine, Coltheart, Hendy, & Roberts (2000) report case RZ
 Stranger (sometimes known to some degree)
4Stranger(s) or person(s) who are recent or casual acquaintances (e.g., nurses in the hospital)It's a person you know well (could be more recent or long-standing acquaintance, or known but not personally familiar—i.e., a famous person). The stranger appears physically different from the misidentified known persons, but is believed to be psychologically identical; hence, a known person in disguiseFregoli delusion

Strangers being misidentified: de Pauw et al. (1987)

Both strangers and casual or recent acquaintances being misidentified: Courbon and Fail (1927)

5Stranger (or known) personDifferent known person, both physically and psychologicallyA generic variant of misidentification delusion (likely involves more basic face-processing difficulties)Young, Flude, and Ellis (1991)15
6Stranger (or known) personAs in 5 above but with reference to transformation into the other known person.16IntermetamorphosisCourbon and Tusques (1932), translated by Ellis et al. (1994)
7StrangerPhysical duplicate of yourselfSubjective doubleChristodoulou (1978)

We leave readers to consider these different types of person misidentification delusion. For our own purposes, we focused on those cases that meet criteria for a Type 4 misidentification delusion: a real person, who is a stranger (or a more recent or casual acquaintance, e.g., a nurse in a hospital), is misidentified as a different person, who is known to the patient. The known person may be personally familiar (a more recent or long-standing acquaintance) or known, but not personally familiar (i.e., a famous person). The stranger appears physically different to the misidentified known person but is believed to be psychologically identical to that known person.5

Thirty-four of the original 72 cases did not qualify as clear instances of Fregoli according to this Type 4 definition. Some cases were discussed as Fregoli like but involved other types of misidentification delusion—for example, Christodoulou (1978) described a case of subjective doubles as being Fregoli like with regard to there being some doubling-up of persons, in this case, a doubling-up of the self. Other cases were excluded because there was no clear indication that the patient believed that the known person appeared in a physical form unlike his or her usual appearance (e.g., the report made no mention of “disguise,” “plastic surgery,” “physical transformation,” etc.). In other cases, there was no clear indication that the delusional belief concerned other specific people known to the patient. For example, we excluded a case where the patient believed that tradesmen who had come to the door (the real persons) were impostors trying to gain entry to his house to kill him and his wife (Wolff & McKenzie, 1994). Here, there is no known person involved at all, so this appears to be a generic sort of persecutory delusion. We likewise excluded a case where the patient believed that “freemasons from Athens” were appearing as his neighbor to get close to him to try to harm him (Margariti & Kontaxakis, 2006). In this case, there was no clear indication that the delusion involved any particular freemasons who were known to the patient. So the belief might simply have been yet another persecutory delusion about an abstract group entity of nefarious freemasons pretending to be neighbors to get close to people for persecutory reasons.6 In contrast, we allowed cases where the known persons were well-known fictional entities (e.g., TV characters). In these cases, we take the view that there still exists some specific known person, albeit, a fictional entity, about whom the patient has a store of personal identity information (e.g., the known fictional character's typical physical appearance, their characteristics, etc.). This is just the sort of information that we all store and update when we track fictional characters in books and movies. We provide an example of this type of Fregoli case later.

The remaining 38 clear cases of Fregoli varied markedly with regard to the numbers of real persons and numbers of mistaken known identities involved. We provide examples below. We also selected these particular examples to illustrate some of the answers to our questions.

6 Subtypes of Fregoli delusion

If we consider the logical alternatives, there are four possible combinations of numbers of real persons by numbers of mistaken known identities:

  1. One real person (e.g., a particular casual acquaintance) is misidentified as a single known person;
  2. Two or more real persons (e.g., different strangers on different occasions) are misidentified as a single known person appearing in different physical guises;
  3. One real person is misidentified as two or more known persons (on different occasions, one assumes—for example, a particular shop keeper is misidentified as Robine in disguise on 1 day and as Sarah Bernhardt appearing in the same guise on another day); and
  4. Two or more real persons are misidentified as two or more known persons.

Although category (c) is logically possible, we found no clear case of this type. In contrast, we found clear examples of subtypes (a), (b), and (d) and provide examples of each.

6.1 Examples of subtype (a) (a single real person misidentified as a single known person)

There were seven clear cases of this subtype, including a 68-year-old woman with schizophrenia who suffered a stroke (Moriyama et al., 2007). Cognitive testing revealed face recognition impairments similar to those seen in prosopagnosia.7 Her face recognition impairments were unstable, however, and, during a period when her prosopagnosia disappeared, she developed Fregoli delusion. She misidentified a male fellow patient (the single real person—a relative stranger) as her husband (the single known person). This was despite the lack of physical similarity between the two and that her husband had died 4 years earlier. She offered no explanation of why her husband might have been in the hospital.

In contrast, McEvedy, Hendry, and Barnes (1996) reported a subtype (a) case with marked paranoid elaborations about the “pursuer's” motives. This was a 38-year-old female who believed that a stranger, whom she had encountered on the Thames Embankment in London (the single real person), was a particular cab driver, whom she knew from the United States, in disguise. She had been arrested for assaulting this cab driver 2 weeks earlier in Los Angeles, before traveling to the United Kingdom. She also believed that a Russian Jewish musician in the United States, whom she also knew, had hired the cab driver to run her down. She believed that the cab driver was now in the United Kingdom and had used plastic surgery to alter his appearance. This patient showed no abnormalities on cognitive testing (including of verbal and visuospatial processing and memory, and frontal function).8

One other example of this subtype is of particular interest because the patient also had a Capgras delusion. This was a 62-year-old woman with paranoid schizophrenia who developed a Fregoli delusion during her inpatient stay (Turkiewicz, Zanetti, Zung, & Cordeiro, 2009). She believed that her female doctor (the single real person) was her husband “disguised with a woman's mask” (p. 245). Clearly, there were no visual similarities here. Moreover, the husband had died 2 years earlier. Her admission was precipitated by other delusional beliefs at the time, including that her 20-year-old son had been replaced by an impostor—a Capgras delusion. It was not clear, however, whether the patient's Fregoli and Capgras delusions were ever present at the same time. Another example is more clear-cut in this regard (see below).

6.2 Examples of subtype (b) (multiple real persons, typically strangers, misidentified as a single known person)

There were 13 clear cases were of this subtype, including a 36-year-old woman with paranoid schizophrenia (Papageorgiou, Lykouras, Ventouras, Uzunoglu, & Christodoulou, 2002). This patient's misidentification delusions began after she had separated from her fiancé, who had refused to marry her, even though she was pregnant. After this she had an abortion. When admitted to hospital, she had prominent Capgras delusional beliefs about her father, mother, and brother, whom she believed had all been replaced by doubles, causing her to leave her family home. She also experienced depersonalization and derealization and developed a Fregoli delusion. Specifically, she misidentified a series of strangers (the multiple real persons) as a single ex-friend, who had asked her to marry him, but whom she had turned down.9 During the psychiatric interview, she also misidentified the psychiatrist as this same ex-friend, stating that “although there was not any physical similarity between the two persons, she recognized the psychiatrist as her ex-friend by the way he looked at her” (p. 806).

In this case, the simultaneous presence of Capgras and Fregoli delusions was more clear-cut. This is at odds with Ramachandran and Blakeslee's account of Fregoli being due to hyperexcitation of face processing and autonomic arousal connections, as is the co-occurring depersonalization and derealization because depersonalization and derealization have been attributed to reduced autonomic responses to all manner of personally familiar and significant events (Phillips & Sierra, 2003). In this case, mannerisms of the psychiatrist, rather than any visual similarities between the ex-friend and the psychiatrist, may have triggered the activation of the ex-friend's PIN to cause him to be identified as present. Finally, although this patient had been diagnosed with paranoid-subtype schizophrenia, the excessive activation of the ex-friend's PIN was likely to have been motivated more by a need for a caring person than paranoid fears. Cognitive testing of this same patient was undertaken by Lykouras, Typaldou, Gournellis, Vaslamatzis, and Christodoulou (2002) and revealed impairments of attention, memory, visual-spatial processing, and executive function, but there was no specific testing of face processing.

Another subtype (b) case involved a fictional known person. Mulholland and O'Hara (1999) reported a 44-year-old male with chronic schizophrenia and a web of delusions, many of which were about the place where he lived; although this place resembled a Northern Ireland city (which it was), to the patient, it was really “Middle Earth” from Tolkien's (1954) “The Lord of the Rings.” His Fregoli delusion was part of this web; he believed that he was being followed and persecuted by “Saruman,” one of the main characters in Lord of the Rings, who took on the guise of the different people he met (the multiple real persons).

This patient, like the patient who believed that she was being pursued by the hit-man cab driver, attributed harmful intentions to the known person. Not so for the next example, a 35-year-old divorced woman with chronic schizophrenia, whose Fregoli delusion arose in the context of erotomania (Wright, Young, & Hellawell, 1993). She believed that she was the secret girlfriend of the American actor Erik Estrada. She believed that Erik planned to marry her one day, and visited her often, disguised as casual acquaintances or as her current boyfriend (the multiple real persons, one of whom was personally familiar to her—her current boyfriend). Cognitive testing revealed face-processing impairments, including poor matching of unfamiliar faces, poor memory for unfamiliar faces, and difficulty with recognizing photographs of familiar faces. She showed no tendency, however, to misidentify unfamiliar faces as familiar (e.g., she did not claim that any of the photographs were Erik, in disguise).

This is the second Fregoli example involving face-processing impairment and another case where the patient made elaborate inferences about the known person's motives, in this case positive motives.

We turn now to illustrate some subtype (d) cases.

6.3 Examples of subtype (d) (multiple real persons misidentified as multiple known persons)

There were nine cases of this subtype.10 Sometimes these cases involved two or more one-to-one mappings (i.e., each particular real person was misidentified as a single known person); and sometimes multiple different real persons were misidentified as a single known person, and this happened for two or more known people. And then there were cases that were a mix.

Duggal (2004) reported a case that involved only one-to-one mappings: a 30-year-old woman with “interictal psychosis”11 who had been in a motor vehicle accident 8 years earlier. At admission she misidentified multiple strangers as multiple known people, but always in one-to-one mappings. She misidentified another patient on the ward (one of the real people involved) as her current boyfriend (the misidentified known person), a nurse as her mother, and the social worker as her sister. In no case was there any physical resemblance. Nevertheless, the patient believed that her family members and her boyfriend had been transformed physically into these unfamiliar physical forms. No other details were provided.

de Pauw et al.'s (1987) classic case was of the other subtype (d) variant: multiple real persons being misidentified as a single known person with this occurring for two distinct known people. This was a 66-year-old widow who misidentified multiple different strangers as her cousin and his lady friend. The patient went into great detail, describing the wigs that the couple used, their false beards, make-up, etc., and complaining that the lady friend would sometimes dress like a man and that her cousin could make himself out to look like an old man. Clearly, there were marked visual differences between the real persons and the known persons. Cognitive testing revealed impaired naming of famous faces, while discrimination of unfamiliar faces and performance on the Benton Test of Facial Recognition remained within normal limits (Young, Ellis, Szulecka, & de Pauw, 1990). This patient did, however, actively engage in making inferences about the known persons' motives. She feared her “pursuers” (the cousin and his lady friend) and took complicated routes to get from one place to the next, trying to shake them off. She believed that they thought that she was the only one in the family who knew that they were having an affair and also knew that they were engaged in criminal activities.

We turn now to consider the nature of aberrant person tracking and identification in Fregoli.

7 What has been disrupted and what remains intact in the person tracking of Fregoli cases?

Focusing first on what remains intact; this includes, first and foremost, the processing of the perceived forms of the real persons. This is why the patients refer to disguises and plastic surgery or otherwise comment on the physical differences. That we found no clear cases of subtype (c) suggests that most Fregoli patients can also update their known person identity information with new information, including their use of a disguise.12 This is why, if ever that same disguise is seen again, the same known person is misidentified as present, as occurs in the several type (a) cases of one-to-one mapping that we identified.

Many Fregoli cases also made inferences about the known person's motives, suggesting that the normal mind-reading capacities that sustain mental-state inferences of this type are intact in cases of Fregoli. Hirstein (2005) proposes some very specific sorts of mismatching “mind-reading” error, rather than a general reduction of mind reading (as traditionally tested on “theory of mind” tasks) as the underlying cause of Fregoli, however. He also refers to the brain regions involved in misidentification delusions and related conditions to support his mind-reading account (see Hirstein, 2005, for further detail). Nevertheless, if one wants to develop and test a cognitive model of the role of disordered mind reading in misidentification delusions, one would want to provide some behavioral evidence of the specific sorts of mind-reading deficits that are hypothesized to be involved. To the best of our knowledge, no detailed testing of Fregoli patients' mind-reading abilities has been undertaken.

While the previous section illustrated the heterogeneity of Fregoli, there is always a homogeneous core: the misidentification of a real person as a different known person appearing in the physical guise of the real person. The error of person tracking involved here is a “false positive:” The patient believes that the known person is present when in fact he or she is not. Matters are slightly different with respect to the real person. Often this is a stranger. Thus, the notion of tracking does not apply here because to track a person is to keep track of him or her over time, which obviously cannot be done if this person has never been encountered before. But sometimes the real person is someone the patient knows. When a patient misidentifies her current boyfriend as Erik Estrada, for example, that is also a failure of tracking: a “false negative”—the patient fails to identify the boyfriend as present, when in fact he is. In these latter cases, the tracking error could be made more explicit by asking the patient “And where is your boyfriend?” As far as we know, this kind of question has not been put to Fregoli cases.

As for identifying the cause(s) of these false-positive and false-negative tracking errors, three of the examples reported in the previous section displayed a concurrent face-processing impairment. Across all 38 Fregoli cases included in this review, face processing was assessed in nine cases, with six of these nine showing some impairment on one or more face-processing tasks. Whether face-processing impairment is causally implicated in these cases of Fregoli, or whether it is just an associated symptom with no causal connection to the delusion, remains unknown. Neither the Ellis and Young (1990) nor the Ramachandran and Blakeslee (1998) accounts implicate any disruption to the first two components of the Ellis-Young model in Fig. 1. Both explanations conceive the primary cause of Fregoli as hyperactivity. For Ellis and Young, particular PINs become hyperexcitable and readily engaged; for Ramachandran and Blakeslee, the neural connections between the facial recognition systems and the amygdala system for affective arousal are hyperactive. The seven cases of one-to-one mapping (of real person to misidentified known other person) are more consistent with Ellis and Young's account than the account of Ramachandran and Blakeslee because the latter predicts many-to-one or many-to-many mappings.

Even more compelling, we found one clear case of concurrent Fregoli and Capgras delusions and another three cases that were suggestive of the same co-occurrence (e.g., Turkiewicz et al., 2009). Any co-occurrence of Fregoli and Capgras is inconsistent with Ramachandran and Blakeslee's hypothesis because these authors attribute the Fregoli delusional content to excessive inappropriate arousal to faces, while Capgras is attributed to reduced arousal to faces; both cannot occur at the same time. And even if the additional suggestive cases involved only sequential rather than concurrent Fregoli and Capgras delusions, Ramachandran and Blakeslee's account would still need to explain how excitation of the face processing and the amygdala connections in these cases can fluctuate so dramatically between hyper- and hypoexcitation.

8 Summary and conclusions

We present the first detailed review of the error(s) of person identification and tracking that characterize Fregoli delusion. Having identified a notable lack of definitional clarity in the literature, we developed a novel classification scheme of person misidentification delusions to identify those cases that qualify as genuine Fregoli cases which enabled us to identify 38 cases as providing sufficient detail to qualify as a genuine Fregoli case. Those 38 cases clearly showed, however, that Fregoli involves a quite distinctive error of person identification and tracking. For example, we showed that Fregoli delusion differs from other, more generic, misidentification delusions that involve more basic errors of visual and/or face processing. We similarly discussed how Fregoli is different to those everyday misidentification error(s) that we all make from time to time when we mistakenly think that we see someone we know in the crowd. This is because Fregoli patients believe that they see someone they know in disguise. We were also able to show that Fregoli differs from more generic persecutory delusional beliefs about being monitored by impersonal group entities, such as the CIA. This is because Fregoli patients nominate specific known person(s) as appearing in disguise.

Most of the 38 Fregoli cases engaged in mental-state inferencing about the known person's motives. There were several cases involving some degree of face-processing impairment but also some cases where investigations failed to detect such impairment. There were seven one-to-one cases (of one real person misidentified as one known person), a pattern which is more consistent with Ellis and Young's (1990) account than that account of Ramachandran and Blakeslee (1998). The one clear-cut case of concurrent Capgras and Fregoli delusions also runs counter to Ramachandran and Blakeslee's account. Another three cases of patients with both Capgras and Fregoli delusions also cast doubt on their account.

Only Ellis and Young's hypothesis was generally consistent with our findings. These authors had proposed that the Fregoli delusional content is generated when hyperexcitation from the cognitive system to the PINs causes a known person to be identified as present, even when no matching face is also present. This account remains underspecified, however, with regard to why a particular stranger in the crowd is picked out as the known person in disguise. This cannot be because of partial activation of a known person's face recognition unit, at least in the many cases that involved no visual similarity between the real person and the misidentified known person (e.g., several cases involved different genders). Therefore, to make further advances—either to refine or to refute this account or even to discover that no single account adequately explains all cases—we will need more detailed single-case studies of Fregoli. These single-case studies need to provide not only detailed clinical case descriptions but also the results of comprehensive neuropsychological testing (e.g., including assessment of face processing) and psychophysiological testing (e.g., to record SCRs across multiple modalities, such as face and voice). Almost no detailed work of this type has been done with Fregoli cases.


We thank the CCD Belief Formation Group for their comments on earlier drafts of this paper.


  1. 1

    Ellis and Young also attribute the uncritical acceptance and maintenance of the Fregoli delusion to inappropriate evaluation of evidence by the cognitive system. As noted earlier, however, our focus in this paper is to better understand the errors that generate the Fregoli content in the first place. Hence, we make no further reference to the role of a disordered cognitive system in defective belief evaluation.

  2. 2

    Ellis and Young made no reference to the processing of mannerisms as playing such a role; we merely offer this as a possible example.

  3. 3

    We were very inclusive at this stage and included all cases labeled as Fregoli or discussed as Fregoli-like.

  4. 4

    We make no claim that Table 1 distinguishes between all possible combinations of this type. That level of complexity was not necessary for our purpose, which was to identify those cases that qualify as Fregoli for the purposes of answering our questions.

  5. 5

    Note that our definition does not require that the patient makes reference to the known person “following” or “being in pursuit.” This is to allow that one particular real stranger might be misidentified as one other known person, in which case any mention of following is less likely.

  6. 6

    We acknowledge that cases involving a “group agent” might be worth pursuing further, for example, to compare to classic Fregoli cases. This was beyond the scope of this first detailed survey of Fregoli, however.

  7. 7

    The authors provided no further details.

  8. 8

    No more specific details were provided.

  9. 9

    This ex-friend was not the ex-fiancé.

  10. 10

    There were nine clear cases of Fregoli that did not provide sufficient detail to enable classification according to subtypes (a), (b), (c), or (d).

  11. 11

    In this condition, psychotic symptoms appear in the intervals between epileptic seizures.

  12. 12

    We acknowledge some need for caution here since sufficient details were missing from 9 of our 38 cases to classify them as (a), (b), (c), or (d). Nevertheless, we think it worthy of comment that we found no identifiable subtype (c) cases and many subtype (a) cases.

  13. 13

    Misidentifications of non-human animals have also been reported (e.g., Somerfield, 1999).

  14. 14

    Description of the stranger may differ (e.g., patient may describe the person as a robot, or an impostor).

  15. 15

    We acknowledge that this case is not a clear example. Such clear examples are difficult to identify. In this paper, Young and colleagues report a man who had suffered a right hemisphere stroke and who mistook a postgraduate student who came to test him for his daughter. The patient pointed out photographs of his daughter to the student, suggesting that he was pointing out what he saw as visual similarities.

  16. 16

    Sometimes this transformation is witnessed. Sometimes the transformation is not witnessed.


Article Information

Format Available

Full text: HTML | PDF

Copyright © 2014 Cognitive Science Society, Inc.

Request Permissions

Publication History

  • Issue online:
  • Version of record online:
  • Manuscript Accepted:
  • Manuscript Revised:
  • Manuscript Received:


  • Fregoli delusion;
  • Misidentification delusions;
  • Person tracking;
  • Person perception;
  • Person identity processing;
  • Face processing


  • Breen, N., Caine, D., & Coltheart, M. (2000). Models of face recognition and delusional misidentification: A critical review. Cognitive Neuropsychology, 17(1), 55–71.
  • Breen, N., Caine, D., & Coltheart, M. (2001). Delusional misidentification: Two cases of focal onset dementia. Neurocase, 7, 239–254.
  • Breen, N., Caine, D., Coltheart, M., Hendy, J., & Roberts, C. (2000). Towards an understanding of delusions of misidentification: Four case studies. Mind & Language, 15(1), 74–110.
  • Bruce, V., & Young, A. (1986). Understanding face recognition. British Journal of Psychology, 77(3), 305–327.
  • Capgras, J., & Reboul-Lachaux, J. (1923). L'illusion des “sosies” dans un delire systematisé. Bulletin de la Société de Médecine Mentale, 11, 6–16. Translated by Ellis, H. D., Whitley, J., & Luauté, J-P. History of Psychiatry, 1994, 5, 134–137.
  • Christodoulou, G. (1978). Syndrome of subjective doubles. American Journal of Psychiatry, 135(2), 249–251.
  • Courbon, P., & Fail, G. (1927). Syndrome d' “Ilusion de Frégoli” et schizophrénie. Bulletin de la Société Clinique de Médecine Mentale. Translated by Ellis, H. D., Whitley, J., & Luauté, J.-P. History of Psychiatry, 1994, 5, 134–137.
  • Courbon, P., & Tusques, J. (1932). Illusion d'intermétamorphose et de charme. Annales Médico- Psychologiques, 90, 401–405.Translated by Ellis, H. D., Whitley, J., & Luauté, J.-P. History of Psychiatry, 1994, 5, 134–137.
  • Duggal, H. S. (2004). Interictal psychosis presenting with Fregoli syndrome. Journal of Neuropsychiatry & Clinical Neurosciences, 16(4), 543–544.
  • Ellis, H. D., & Lewis, M. B. (2001). Capgras delusion: A window on face recognition. Trends in Cognitive Sciences, 5(4), 149–156.


Delusional misidentification syndromes (DMSs) are complex psychotic phenomena that may be present in a variety of ways within the context of several neurological and psychiatric disorders. Since the first case of Capgras syndrome was described in 1923, various other syndromes have been identified, including Fregoli syndrome, intermetamorphosis, subjective doubles, reduplicative paramnesia, mirrored self, delusional companions, and clonal pluralization of the self. In this article, we review each of the different syndromes in definition and presentation, as well as the field's attempts at classifying them. We then describe their role in forensic psychiatry, particularly in regard to their potential as a marker of a particular subpopulation or of illness severity and their consideration in risk assessments of violence. A review of the literature was conducted for this purpose, and, although it was extended to include publications from over four decades, it revealed a paucity of research on DMSs.

Without wearing any mask we are conscious of, we have a special face for each friend.—Oliver Wendell Holmes

Few concepts in psychiatry can be as confusing as the delusional misidentification syndromes (DMSs). One goal in psychiatry is to achieve a better understanding of the self: who it is, how it is organized, and how it develops and reacts to others. DMSs introduce a multiplicity of aspects into this understanding of identities and relationships, adding to its inherent multifaceted complexity. They are fascinating because they are disruptions in what we consider the normal integrity of the self, and for forensic experts, interesting for how they help us understand entanglements between mental health and unlawful behavior.

Delusional misidentification syndromes are rare psychopathologic phenomena that may occur within the context of schizophrenia or affective or organic illnesses. They include Capgras syndrome, Fregoli syndrome, intermetamorphosis syndrome, syndrome of subjective doubles, mirrored self, delusional companions, and clonal pluralization of the self. Misidentification syndromes show a great degree of overlap and do not represent distinctive syndromes, nor can they be regarded as an expression of a particular disorder. Evidence suggests that one type of misidentification delusion may evolve into another type.1 However, these syndromes merit distinct identification and therapeutic approaches because of their possible underlying disorders and their potential for dangerous behavior.2 Furthermore, for forensic experts, they may be instrumental in assessments of risk and criminal responsibility.

In this article, we review the available literature regarding these syndromes. We also analyze DMSs and the forensic population, seeking any correlations between diagnosis of DMSs and other key concepts such as dangerousness or legal underpinnings. Finally, we attempt to describe guidelines for the clinical management of these patients, or for the incorporation of this psychopathology into forensic assessments.

Definitions and Classifications

DMSs all carry a common classic theme of one person being an exact likeness of another: the sosie or double. They can be distinguished as hypoidentifications (Capgras syndrome) and hyperidentifications (the other syndromes).3 Different terminologies and classifications have led to confusion in the past. Roessner4 presented a classification that included two categories: one in which the object of the delusion is physically altered (or replaced) in the patient's mind, and a second in which the object is doubled, rather than replaced or transformed. Another proposed classification system using developmental and regressive understanding has also been proposed,5 in which DMSs are divided into two main subgroups: relational misidentifications (misidentification of human relationships) and identical misidentifications (misidentifications of identity itself, which includes Capgras and Fregoli). Identical misidentifications are further subdivided into divided-identity type, unionized-identity type, and transformed-identity type. This typologic approach allows for developmental understanding and the application of regressive theories. Signer6 proposed an extension of reverse types of misidentification syndromes, distinguished by alteration of the self rather than of others. Beyond definitions, the phenomena have sparked human interest throughout time, as reflected in stories and myths about doppelgangers, imposters, and clones. Some of these artistic examples emphasize the importance to human psychology of the identification of self and others and the potential ramifications of wrongful identification. We briefly review the concepts and psychopathology and offer a few illustrative examples.

Capgras syndrome was first described in 1923 by psychiatrists Joseph Capgras and Jean Reboul-Lachaux. It is the most prevalent of the delusional misidentification syndromes and is described as a disorder in which a person holds a delusion that an identical-looking impostor has replaced a friend, spouse, parent, or other close family member. One can imagine this syndrome in the performance by Donald Sutherland in the movie, Invasion of the Body Snatchers,7 a remake of the original 1956 science fiction film in which humans are replaced by emotionless alien clones. Reverse Capgras syndrome refers to the psychological change of the self as opposed to others. A study8 found that, of 30 subjects with this syndrome, most of the new identities were famous figures or others who were admired by the affected person. Most of the individuals also experienced a sudden belief or realization of having a new identity or of having rediscovered a pre-existing one.

Leopoldo Fregoli, an entertainer from the late 19th century, perfected a style of performance known as protean or quick-change. He could switch costumes and characters during his stage shows so rapidly that it was thought that several other Fregolis must have existed for his act to be possible. Fregoli syndrome is the delusional belief that one or more familiar persons, usually persecutors following the patient, repeatedly change their appearances (i.e., the same person assumes numerous different disguises).

Intermetamorphosis is a misidentification syndrome in which an individual has the erroneous belief that familiar persons have exchanged identities. In the syndrome of subjective doubles, patients believe that there are other persons who look like them, but that they have different traits and live different lives. This situation has been commonly depicted in movies, such as The Sixth Day,9 where Arnold Schwarzenegger's character is cloned without his knowledge or consent, and in TV shows, such as Battlestar Galactica10 and Star Trek,11 where clones represent the main rivals to the shows' heroes. Mirrored-self misidentification involves the misperception that one's reflection in the mirror is a stranger. Individuals affected with the syndrome of delusional companions believe nonliving objects possess consciousness, can think independently, and feel emotion. The movie Night at the Museum12 features objects exhibited at a museum that appear alive to the protagonist. Clonal pluralization of the self differs from the syndrome of subjective doubles, in that the patient believes that there are multiple copies of himself who are physically and psychologically similar to themselves. As an example, Ranjan and colleagues13 reported a patient with schizophrenia who thought that there were triplicate copies of herself and others.

Further extensions of these core DMSs have also been described. Somatoparaphrenia is a subtype of asomatosognosia, in which patients also display delusional misidentification and confabulation. It also involves orbitofrontal dysfunction, which distinguishes it from asomatosognosia.14 Reduplicative paramnesia is the belief that a place or location has been duplicated or relocated. This is the scenario in the movie The Truman Show,15 where the protagonist finds that his world is actually a reality TV show set. Similarly, the concept of the physical world as an illusion has been depicted in Vanilla Sky,16The Thirteenth Floor,17 and The Matrix.18 Other extensions of DMS have been postulated to include lycanthropy,19 Ekbom syndrome, delusional hermaphroditism,20 delusion of sexual transformation,21 and the antichrist delusion.22 DMSs have even been documented in the context of a folie à deux shared delusion of doubles.23


Feinberg24 recently described a comprehensive, multimodal, hierarchical model for understanding the neuropathologies of the self, which includes negative factors (defects or absence of neurological functions) and positive factors (productive, defensive, and motivational brain features). Neurobiological findings strongly support a structural basis for DMSs. Facial processing involves right ventromedial occipitotemporal regions and areas of prefrontal cortex via the uncinate fasciculus and limbothalamic pathways.25 Other researchers26 have suggested that impairment of facial recognition plays a role in the pathogenesis of DMSs. Lesions found in DMSs are usually bifrontal, right hemispheric, or both. A disconnection is observed between the frontal lobes and the right temporolimbic regions (hippocampus), which are necessary for reconciling information about self-identification of the person and his associated emotions.

Neuropsychological testing further suggests that misidentification delusions are associated with subtle abnormalities in facial recognition abilities and with nondominant cerebral compromise.1 Underactivity in the perirhinal cortex seems to be responsible for loss of familiarity in Capgras syndrome, whereas overactivity seems to account for hyperfamiliarity, seen in the Fregoli, intermetamorphosis, and subjective doubles syndromes. Impaired connectivity between the right fusiform and right parahippocampal areas has also been implicated in deficits in visual memory recall, face recognition, and identification processes in these patients.

According to cognitive models, the dysfunction extends beyond facial recognition, whereby the person cannot be globally considered. The feeling of familiarity is absent because of the inability to integrate successive memories about a person along with episodic experiences, thus generating delusional doubles in accordance with the patient's needs or drives.27

Devinsky28 postulated a dual mechanism: on the one hand, negative effects from the right hemisphere and frontal lobe dysfunction impair self-monitoring, ego boundaries, and attached emotional valence to familiar stimuli; on the other, the preserved left hemisphere areas exert a positive effect from release or overactivity, providing a narrator from the monitoring of self, memory, and reality. This effect leads to excessive and false explanations, or, because of the dual-category style of cognitive categorization, it leads to invention of a duplicate or impostor to resolve conflicting information. Politis and Loane29 echoed this theory and highlighted a consensus that right and bifrontal lesions, as well as the cognitive dissonance associated with impairments in memory, visuospatial abilities and conceptual integration, are common factors in DMSs (reduplicative paramnesia in particular).

Psychodynamic models may also be used to understand the phenomenology and subjective experience of DMSs. The central theme revolves around defining identity and its multifaceted dimensions. Consideration must be given to the introjected sense of self and the capacity to hold a reflective and observing ego. Conversely, the pathological expression that would give way to a DMS is the experience of depersonalization. Other ego functions may be compromised, including the loss of motivation and self-initiated drives. DMSs may also be conceptualized as defensive structures, whereby negative aspects of the self are split off and projected into an external (and targetable) other.

From a developmental perspective, an arrested early development or regression process may be identified. Misidentification phenomena are manifestations of defense mechanisms of splitting and projection.30 An aspect of the internalized self or object representation to whom negative emotions are attached are split off from the self and projected externally, onto a different identity. These mechanisms, being primitive in nature, can also be explained by regression theory.31 When higher cerebral functioning is affected, its compromise results in reactivation of primitive modes of thinking characterized by the theme of doubles and dualisms, also found in myths, primitive religion, and literature.

Other variants of this theory propose that deep regression reactivates a developmental stage before the establishment of object constancy, where there is a splitting of objects into all good or all bad and an absence of self–object differentiation. This effect is consistent with other elements of relations theory, in which a primitive self is unable to establish a trusting relationship with a cohesive other, described as the Kleinian paranoid–schizoid position. Along these lines, others highlight the inability to attribute uniqueness to the self and surrounding people,32 despite positive and negative attributes.

Altered affective response toward others may also be at play. Intolerable affective ambivalence toward others may be neutralized through the imagined existence of doubles.33 For example, individuals with Capgras syndrome may harbor anger or envy toward a close relative. Denial is then used to make this emotion tolerable and free of guilt for the delusional person. If this mechanism becomes insufficient, the person may then split the object, attributing only positive feelings to the original object and only negative feelings to the delusionally altered identity. Projection is involved in directing the negative emotions at the object without experiencing significant internal conflicts. Projective identification becomes possible when the patient in turn experiences positive feelings coming from the delusionally altered identity. The affected individual fears others as hostile and may even strike preemptively in response to it.34

Others have placed emphasis on the role of language and narrative in the manifestation of DMSs. They propose difficulties in the self-reflexive property of the human mental functioning and the first-person linguistic expression of human experience,35 with an aberrant semantic processing of identity.

Diagnosis, Treatment, and Prognosis

Ongoing discussion persists within the field pertaining to whether DMSs are neurologic or psychiatric syndromes, whether they are a feature of schizophrenia or delusional disorder, or whether they constitute a distinct phenomenon described in the Diagnostic and Statistical Manual of Mental Disorders.

Several treatment approaches have been described. Treatment of co-occurring psychiatric, substance use, or medical disorder is required.36,–,38 Antipsychotics39 are often used. According to case reports, DMSs have responded favorably to olanzapine,40 sulpiride and trifluoperazine,41 clorazepate,42 and pimozide.43 Antidepressants may be useful regardless of whether the primary diagnosis is a mood disorder.44 If the DMS is associated with mania, lithium is a mainstay.45,46 Group therapy has also been proposed.47

Prevalence and Relevance in the Psychiatric and Forensic Population

DMSs have been associated with both focal and diffuse neurologic conditions, such as hypothyroidism,48 right hemispheric stroke,49 multiple sclerosis,50 and dementia. They have also been associated or identified with psychiatric conditions, especially schizophrenia, Alzheimer's disease (AD),51 and PTSD.52 Research has shown that within psychotic illnesses, paranoid schizophrenia seems to be the most common diagnosis in patients with DMSs.53,54

Prevalence in all psychiatric inpatients ranges from 1.3 to 4.1 percent.55 A study in Turkey at a university hospital inpatient setting showed the 5-year prevalence rate of Capgras syndrome to be 1.3 percent (1.8% for females and 0.9% for males).56 However, DMSs occur more frequently than previously thought.57 According to Dohn and Crews,58 the prevalence of DMSs among patients identified as schizophrenic is 15 percent. They postulated an estimated prevalence of 0.12 percent in the general population for Capgras syndrome. A study in patients with AD demonstrated a prevalence between 2 and 30 percent.59 In a different study, DMS was identified in 15.8 percent of cases of AD, 16.6 percent of patients with Lewy body dementia, and 8.3 percent of individuals with semantic dementia.60 With regard to inpatient prevalence, a survey was conducted of all admissions to a locked psychiatric inpatient unit in the Boston metropolitan area from April 1983 to June 1984. Twenty-six (3.1%) of 835 patients admitted to the unit met the criteria for DMSs.61

Prevalence specifically within the forensic population is unknown. Some case reports have been published and will be discussed below.

Is There an Association Between DMS Phenomenology and Violence?

No studies have looked at the association between DMSs and legal history or type of offense. Data correlating criminal behavior and DMSs are limited mostly to case reports or, at best, to descriptive retrospective studies with a low number of study subjects. To assess this correlation accurately, one would need a reliable diagnosis of the phenomena and the legal history, which are often unavailable. To our knowledge, research gathering such data has not been conducted to date. However, some literature addressing this association is currently available.

Research has shown that DMS patients view the misidentified person with suspicion and hostility,62 which may contribute to a mounting paranoia and physical aggression in the form of pre-emptive self-defense. Silva et al.63 postulated that delusional cognition drives the affected individual to construct a narrative in which the misidentified object is conceived as a person for whom biographical history radically departs from a stable, good construct (the original identity), only to be replaced by a bad object, which is then considered authentic. In their research, they found that these patients become aggressive toward the misidentified objects, because they perceive the newly constructed object as threatening to their own welfare. In their book, Silva et al.47 explain how such disturbed thinking can result in dangerous/aggressive acting out toward individuals who are at risk of being harmed by such patients. Of their sample of six patients, all exhibited verbal aggression, and five became physically aggressive toward delusionally misidentified objects.

Another study of 82 subjects with DMSs defined violence as verbal threats or physical violence directly associated with a misidentification delusion. Fifty of the 82 patients had attacked someone else, the most common victims being parents.64 In another study by Silva et al.,65 of 29 patients with DMSs, 16 had threatened others without acting on the threats, whereas 13 became physically assaultive in connection with their misidentification syndromes. In yet another study, Silva et al.66 also found that dangerous patients with DMSs were less likely to use weapons than were their non-DMS counterparts.

Aggression or physical assault may escalate to the level of murder, as described in the cases above. Silva et al. also reported the case of a homicide associated with delusional misidentification.67 Capgras syndrome has been described in incidents of parricide.68 In one review of the literature, delusional misidentification cases leading to homicide usually involved more than one delusionally misidentified object; in four of nine cases of DMS-associated homicide, there was a prior history of serious physical violence directed at other delusionally misidentified objects.67 In a French study, the highest percentage of homicide occurred in patients with paranoid schizophrenia, and their delusions of misidentification had usually been present for an extended period before the homicidal act.69

Special consideration may be given to cases where a DMS involves a child, especially as the object of a delusion. Silva et al.70 explored these cases more than 20 years ago, finding that they may involve mechanisms of aggression that differ from paranoid self-defense. They described how DMSs have been found in some folklore and regional legends that suggest that aggression might rid the impostor and return the authentic personality. In one Swedish tale, a woman who believed that her child was an impostor was instructed to put the child in an oven to recover her original baby. Although these actions may not literally translate into reality, the concept of physically extricating the imposter to regain the authentic persona remains. Furthermore, DMS has been described in the context of puerperal depression with psychotic features, where cases of Capgras71 and Fregoli72 syndromes have been reported (a phenomenon historically called changeling). When the child is the entity that is believed to have been replaced, the risk of aggression toward the child increases, as the mother perceives that violence may be the only means of freeing the original child from the impostor.

Rather than simply investigating the relationship between two entities (DMS and violence), we must ask a more difficult question: across (or within) diagnostic groups, are patients with misidentification delusions more likely to behave violently than comparable patients without such delusions?73 De Pauw and Szulecka41 postulated that morbid suspiciousness, hostility, and discord, as well as previous aggressive behavior and delusions that focus on interpersonal relationships, rather than on locations or inanimate objects, render DMS patients particularly prone to attacking the subjects of their suspicions. Nestor and coworkers74 found a higher incidence of Capgras syndrome in forensic psychiatric patients with psychotic disorders who had committed homicide or other severely violent acts, than in less violent patients.

Risk Factors for Violence in DMS

Violence in DMSs has been determined to be multifactorial, and several of these factors have to be taken into consideration to gauge adequately the risk of violent enactment of DMSs. These patients seem to have a significant history of physical aggression, independent of their delusional misidentification.66 Increased risk of violence in DMSs may be derived from their high-risk triad of highly valued beliefs, negative affects, and identified targets (often a close relative or attachment figure). Anger due to delusion is a key factor that explains the relationship between violence and acute psychosis,75 and close relatives or attachment figures have been established as significantly higher risk targets.

Based on four case reports, De Pauw and Szulecka41 identified certain factors that may have contributed to aggression in each of their reported patients. These include structural or metabolic brain injury, subjectively experienced salient historical events (associated with negative emotions), low intelligence, poor social and occupational skills, primary psychiatric pathology, and pre-existing personality constructs. It is important to note that these factors were significant for each of these cases and are not offered as generalized risk factors for violence.

Specifically regarding Capgras syndrome, a review study76 looked at demographic and clinical features that may have contributed to an increased risk for violence in cases of Capgras syndrome involving assault. Common factors noted included men with long-standing delusions, a history of aggressive behavior, diagnosis of schizophrenia, and comorbid substance use. Persecutory delusions and anger toward the misidentified person were often present, and command hallucinations and sexual preoccupation were at times also present. Negative syndrome features, such as social withdrawal and blunted affect, were noted. The victim was usually a cohabiting family member, and the violence was usually well planned.

Silva et al.65 also found that males represent a higher proportion (70%) of patients with DMSs exhibiting aggressive behavior. Other factors that increase the likelihood of aggression include specific delusions, such as erotomania77,78 or delusional jealousy.79 In a three-patient case report,80 the authors argued that the degree of threat perceived by the patient from the delusionally misidentified object is the most important factor in determining the patient's response to the delusions. The authors of another case report of three patients agreed with that principal factor and added that impulsivity and dissociation may also play a significant role.81 Case reports suggest that alcohol and substance intoxication facilitate patients' acting on the delusions.82,83 Thompson and Swan84 described 2 subjects with Capgras syndrome who, after heavy drinking, committed serious acts of violence toward family members, who had become subjects of delusional misidentification. Although substances such as alcohol may facilitate the enactment, they do not account for the genesis of the delusional beliefs.

DMS as a Marker of Illness Severity

Given the potential for severe violence, presumed to result as the endpoint of the development of the illness, we attempted to examine whether the existence of DMS denotes a particular subgroup of mentally ill patients. To our knowledge, no studies have been conducted that analyze the relationship between DMS and markers of illness severity, such as duration of illness, age at onset, Global Assessment of Functioning (GAF) or other functionality markers, total number of hospitalizations, or length of stay. With regard to illness severity, a study by Silva and colleagues47 of 25 subjects noted that the Brief Psychiatric Rating Scale (BPRS) total score and grandiosity score were higher in DMS subjects than non-DMS subjects. A study comparing neuropsychological functions found no statistically significant differences in neuropsychological functions in schizophrenic patients with and without DMS.64 However, the study involved only 22 patients and did not include indicators of psychiatric illness severity. To our knowledge at the time of this writing, no valid rating scales or other structured clinical or forensic assessments are available for specific use in patients with DMSs.

There are no formal guidelines for delineating the standards of assessment of these syndromes. Furthermore, there are no systematic studies on treatment requirements of these patients when compared with the more general or forensic population, including the use of intensive precautions and monitoring; the number of failed medication trials; or the use of complex therapies, such as polypharmaceutical regimens, clozapine, or electroconvulsive treatments (ECT). The literature does offer some retrospective or descriptive observations regarding treatment. Zanker85 states that symptoms of DMS are very refractory to treatment despite various neuroleptic therapies.

In a meta-analysis, Silva et al.86 examined 104 misidentification syndrome cases published between 1957 and 1994, where information regarding treatment with antipsychotic medications was available: 70 patients showed improvement, whereas 34 did not. Christodoulou87 reported in his research on 20 patients with DMSs that the syndromes failed to remit in 7 of the 20 cases, and in the others, remission occurred either synchronously with or later than the remission of the underlying psychosis. Other isolated case reports suggest that patients with DMSs who have an underlying organic etiology show remission on most occasions.88,89

More accurate data gathering is rendered difficult, given that these syndromes are often underdiagnosed as distinct manifestations of psychoses. There are no guidelines currently for their assessment and treatment. Some authors have proposed manners of furthering exploration. Hillers and others26 propose that there is a common underlying neuropsychiatric mechanism to these disorders; they therefore propose a neuropsychological battery to assess Capgras syndrome. Some researchers90 have used hypnosis to recreate mirrored-self misidentification, thereby gaining insight into delusional beliefs.

More recently, event-related potentials (ERPs), especially the auditory P300 component, have constituted a useful tool for exploring brain–behavior relations. DMSs are thought to be related to dissociation between recognition and identification processes. Working memory is considered responsible for the integration and online manipulation of information, and ERPs and subsequent measured P300 wave forms are considered an index of the online updating of working memory. A study91 found P300 amplitude in prefrontal areas to be significantly reduced in patients with schizophrenia with DMS compared with those with schizophrenia without DMS and controls. They also found P300 latency in the central midline brain region to be significantly prolonged in the DMS subgroup. The methods discussed herein may offer future ways of identifying and studying these patients.

Specific Areas of Concern for Forensic Assessments

Most of the research cited above focused on the role of identifying DMS for the purposes of risk assessment of violent behavior. However, added consideration must be given to other areas in forensic assessments where DMSs may play a significant role given their impact on rationality, understanding of wrongfulness, ability to control voluntary acts, and culpable mental state. In the criminal arena, a DMS may represent the essence of incompetency, whether the DMS applies to the self or others involved in the crime or in the legal proceeding. The presence of a DMS may also raise a particular argument in criminal-responsibility disputes, in that the cognitive prong would be inherently impaired. The question of criminal responsibility could also involve the volitional prong in some jurisdictions, and the assessment of whether the patient was able to conform his conduct, even if he believed the victim (or himself) to be a different identity altogether. In states without an insanity defense, the presence of DMS perhaps should be considered for the purposes of establishing mens rea and willingness to commit a criminal act in a knowing and voluntary manner. Finally, a DMS as an element of severe illness could be considered a mitigating factor, or a relevant factor in determining sentencing needs (especially for therapeutic considerations).

In the civil arena, identification of DMS toward a child could be a center point of establishing parental capacities and custodial rights. Matuszak and Parra92 described a case in which child custody was disputed because the mother believed that her daughter had been replaced by an impostor. DMS might also be attached to a caregiver or guardian of an adult patient, directly affecting the patient's welfare. The presence of DMS could have a direct association with medical decision-making capacity should the patient believe, for example, that his doctor has been replaced by an impostor. Moreover, it becomes particularly relevant in matters of testamentary capacity, especially given the strong correlation between DMS and structural or neurological pathologies common in the elderly.

DMSs have, in fact, been a key element in several legal cases. In People v. Singer93, defendant Jerry Singer was found guilty but mentally ill in the murder of three people by stabbing. Mr. Singer perceived of three people who ran the apartment building where he lived as having been replaced by clones and operating under the influence of psychics who were intent on chasing him out of his own apartment. Both psychiatrists who provided expert witness testimony found the defendant to have schizophrenia and Capgras syndrome, and opined that he was insane at the time of the killings.

Joshua Hoge was found not guilty by reason of insanity in the stabbing and murder of his mother and stepbrother. Mr. Hoge had paranoid schizophrenia, and Capgras syndrome led him to believe that imposters had replaced them. In Re: the Estate of Pamela Kissinger v. Joshua Hoge94 the appellate court determined that his insanity acquittal did not make the state's slayer statute inapplicable to him, as was argued in civil court.


DMSs are complex psychotic phenomena that may arise in the context of neurological or psychiatric pathology and may be more prevalent in such populations than originally thought. Despite this possible underestimation, there is a paucity of research surrounding DMSs, a dearth of recent scientific literature on the subject, and very limited analysis of their impact on forensic psychiatry practice. DMSs may be a hallmark of a particular subgroup within the forensic population correlated with particular criminal behaviors or of illness severity in the general psychiatric population.

As we have explored in this article, DMSs are relevant to forensic psychiatry for several reasons. They may represent an independent risk variable and should be incorporated into dangerousness assessments. There might be a significant correlation with criminal behavior. They may also be a nuclear element in criminal insanity, civil and criminal competencies, and relationships involving the welfare of others or of the patient.

There are no formal guidelines currently delineating the standards for assessment or treatment of these syndromes. This lack of guidance becomes particularly relevant in establishing standard of care and appropriate practice. Following our findings in this review, we propose that DMSs should be directly and specifically screened for identification in the psychiatric assessment, given their correlation with neurological injury and morbidity as well as the increased risk of violence. Once identified, organic, structural, and metabolic conditions must be ruled out and adequately treated. Comorbid substance use must also be identified. Routine blood work should address metabolic parameters of hepatic and renal function (to identify causes or indicators of delirium), thyroid function, and, when pertinent, levels of drugs such as lithium. Brain imaging may help identify some neurological pathologies, such as stroke and multiple sclerosis. Neuropsychological testing can be conducted, as deemed necessary.

Finally, we emphasize the need for more rigorous research on these syndromes, to advance the understanding of their psychopathology, establish accurate incidence and prevalence rates, and correlate them with clinical indicators of severity or dangerousness. In the forensic arena, it would be interesting to examine the importance of the relationship with the persons who are the focus of abnormal beliefs or perceptions, especially in situations of odd or bizarre attachments, and the relationship with stalking. Systematic study of the variables involved in DMSs is crucial, perhaps best achieved through a case-control study that compares patients with DMS to similar patients who do not present such syndrome. In a large state hospital setting, for example, patients identified as having a DMS could be compared against those with no DMS in terms of violence, length of hospital stay, number of failed medication trials, and severity of criminal activity (if pertinent), among other areas.


  • Disclosures of financial or other potential conflicts of interest: None.

  • © 2014 American Academy of Psychiatry and the Law


  1. 1.↵
  2. 2.↵
  3. 3.↵
  4. 4.↵
  5. 5.↵
  6. 6.↵
  7. 7.↵
  8. 8.↵
  9. 9.↵
  10. 10.↵
  11. 11.↵
  12. 12.↵
  13. 13.↵
  14. 14.↵
  15. 15.↵
  16. 16.↵
  17. 17.↵
  18. 18.↵
  19. 19.↵
  20. 20.↵
  21. 21.↵
  22. 22.↵
  23. 23.↵
  24. 24.↵

One thought on “Fregoli Syndrome Case Study

Leave a comment

L'indirizzo email non verrà pubblicato. I campi obbligatori sono contrassegnati *