I can't remember Your Honor: Offenders Who Claim Amnesia

Reprinted from the German Journal of Psychiatry · http://www.gjpsy.uni-goettingen.de · ISSN 1433-1055


I can’t remember Your Honor: Offenders Who Claim
Amnesia

M. Cima1, 2, H. Merckelbach 2, 3, H. Nijman 2, 4, E. Knauer1, and S. Hollnack1

1 Department of Forensic Psychiatry, Rhine Clinics Düren
2 Department of Experimental Psychology, Maastricht University
3 Faculty of Law, Maastricht University
4 Department of Forensic Psychiatry, Kijvelanden Rotterdam

Corresponding author: M. Cima, M.Sc, Department of Experimental Psychology, Maastricht University,
P.O. Box 616, 6200 MD, Maastricht, The Netherlands, maaikecima@planet.nl


Abstract

Objective: Persons who claim memory loss for the crime of which they are suspected represent a problem for the
legal system. When such claims emerge, the court often relies on expert testimonies. Mental health profession-
als who act as experts in these cases often assume that the memory loss is caused by a combination of strong
emotions and abundant substance use (e.g., alcohol). It is also common to gather arguments for such an inter-
pretation by interviewing the offender or suspect. However, discriminating between different kinds of amnesia
requires the use of sophisticated tests and tasks. Methods: By reviewing clinical, experimental, and psychomet-
ric literature, this article addresses several forms of crime-related amnesia and discusses the difficulties that
arise when expert testimony about this phenomenon is given. It also emphasizes the importance of technical
knowledge about diagnostic quality parameters. Results: A review of the literature shows that dissociative
amnesia is not the only or even the most prevalent form of offender amnesia. Experts should at least take the
possibility of other forms of amnesia (i.e., organic amnesia; malingered amnesia) into account. Conclusion:
Expert testimonies about crime-related amnesia should not rely exclusively on interviews or archival data.
Rather such testimonies require tests and tasks combined with good diagnostic interviews (German J Psychia-
try 2002; 5: 24-34)

Keywords: amnesia, memory, forensic, offender
Received: 27.11.2001
Published: 25.2.2001



Introduction
plore crime-related amnesia. To begin with, amnesia raises
the issue of automatism, which refers to criminal behavior
that is executed unconsciously and without intent. The
issue of automatism is, of course, critical for legal responsi-
t is not uncommon that offenders claim memory loss
bility. We will return to the difference between “sane” and
I(i.e., amnesia) for their crime. As a matter of fact, 25- “insane” automatism later on. Secondly, amnesia raises the
45% of criminals found guilty of homicide claim amne-
issue of competency to stand trial. An accused who has no
sia for the event (e.g., Kopelman, 1995). In cases where
memory of the crime event cannot plead in their own
suspects claim crime-related amnesia, the court often asks
cause, simply because they are unable to inform their coun-
for expert testimony. Triers of fact seem to assume that an
sel (Schacter, 1986). Thus, such a person might be incom-
expert is able to differentiate between various kinds of
petent to stand trial.
genuine amnesia, as well as between genuine amnesia and
simulated amnesia. According to Schacter (1986), there are
In a study of Merckelbach, Cima, and Nijman (2002) popu-
two obvious legal reasons as to why it is important to ex-
lar beliefs about crime-related amnesia were investigated.

---

CIMA
The authors describe a typical but real crime case in which
Crime-related Amnesia
an offender, after a night out with heavy use of alcohol and
cocaine, stabbed someone to death. The offender claimed
amnesia for the homicide. The court appointed a mental
health professional in order to give an expert testimony
Offenders who claim amnesia for their crime are by no
about the amnesia of the offender. Based on two interviews,
means rare. In an older study of Leitch (1948), 51 murder-
this expert concluded that the crime-related amnesia was a
ers were interviewed and 14 (27%) of them claimed amne-
result of “the high level of alcohol and drug use”. This case
sia for their crime. Taylor and Kopelman (1984) replicated
was administered to 54 lay persons. They were asked
this study by interviewing 34 murderers: 9 (26%) of them
whether they thought that genuine amnesia would be a
claimed amnesia. In a more recent study, Gudjonsson,
plausible scenario in this type of crime. The majority of
Petursson, Skulason, and Sigurdardottir (1989) investigated
them (82%) indicated that it was a plausible scenario. A
64 convicted criminals. This time, 21 (32%) claimed amne-
majority (76%) also felt that the court was very wise to ap-
sia for there crime. As a rule of thumb, 20 to 30% of of-
point a forensic expert. As well, respondents had very
fenders of violent crimes claim amnesia for their crime.
strong opinions about the origin of the amnesia. A large
While these claims are often raised in the context of mur-
majority of the respondents (70%) felt that alcohol, co-
der or manslaughter cases, there are other crime categories
caine, strong emotions or a combination of these are re-
in which claims of amnesia do occur. For example, claims
sponsible for offenders’ amnesia. Apparently, offenders
of amnesia regularly occur in sexual crime cases (Bourget &
who claim crime-related amnesia do not need to worry that
Bradford, 1995), domestic violence cases (Swihart, Yuille &
their claim meets widespread disbelief. Merckelbach and
Porter, 1999), and fraud cases (Kopelman, Green, Guinan,
associates (2002) wondered whether this has to do with the
Lewis & Stanhope, 1994).
way in which they presented this case. They construed a
In the literature, all kinds of taxonomies have been pro-
version of the case in which the offender claimed to re-
posed to distinguish between different types of amnesia
member “every second of the stabbing” extremely well.
(e.g., Schacter, 1986; Loewenstein, 1991; Kopelman, 1995;
Having read this version, a new sample of respondents (N =
Kihlstrom & Schacter, 1995). Most taxonomies agree that,
54) were asked whether they could also imagine the oppo-
at the very least, three amnesia types should be considered:
site, namely an offender developing amnesia for the stab-
dissociative amnesia (formerly termed psychogenic or func-
bing incident. Again, the large majority (74%) thought that
tional amnesia), organic amnesia, and feigned amnesia
it is perfectly possible that an offender develops complete
(“malingered amnesia”). Dissociative amnesia for criminal
amnesia for his crime. The idea that strong emotions, alco-
behavior is thought to originate from extreme emotions
hol and/or drugs may affect offenders in such way that they
that accompany such behavior. Several authors have argued
fully forget what they have done is apparently widespread.
that dissociative amnesia is characteristic for crimes that are
The ideas of lay people about crime-related amnesia would
unplanned, involve a significant other, and are committed
be of little or no concern to us, were it not that judges and
in a state of strong agitation (e.g., Loewenstein, 1991; Ko-
lay judges are lay people who have to decide about the plau-
pelman, 1995). The idea behind this is that extreme levels
sibility of claims of crime-related amnesia. One would argue
of arousal during the crime may hamper memory retrieval
that these triers of fact rely on expert testimony precisely to
at a later point in time. Thus, a failure in so-called retrieval
overcome their lack of technical knowledge. On the other
processes would underlie dissociative amnesia such that the
hand, in cases of claimed amnesia, mental health profes-
offender who subsequently has come to his senses, finds it
sionals who act as expert witnesses often rely on notions
impossible to access memories stored during a moment of
and assumptions that come dangerously close to those of
turbulence. A term often used in the Anglo-Saxon literature
lay people. In the sections that follow, we first discuss the
to describe amnesia as a consequence of strong emotions
prevalence rates and different manifestations of crime-
(e.g., rage) is “red-out”. In the words of Swihart et al.
related amnesia. We then move on to the legal significance
(1999): “ Apparently, an individual can get so angry with
of amnesia claims. Next, we focus of on several crucial
his/her intimate partner that s/he can severely beat or kill
discriminations that experts have to make in cases where
that partner and then not remember doing so: that is, they
such claims emerge: Thus, the expert has to differentiate
can experience a red-out resulting in circumscribed amne-
between dissociative amnesia, feigned (i.e., simulated amne-
sia” (p. 200). We will discuss the merits of this interpreta-
sia), and organic amnesia. He/she also has to consider
tion later on.
critically the precise role that alcohol plays in this category
Organic amnesia is always caused by a neurological defect.
of cases. Following a detailed examination of these issues,
This defect may be structural (e.g., epilepsy, brain trauma),
we turn to diagnostic and psychometric topics. We will
but it may also be momentary such as in the case of alcohol
argue that powerful tools are available to discriminate be-
or drug intoxication. Kopelman (1995) assumes that mem-
tween various forms of crime-related amnesia. We conclude
ory loss in organic amnesia has to do with storage problems
with a consideration of the specialized knowledge that these
rather than retrieval problems: due to an epileptic seizure,
seemingly simple tools require.
brain damage or intoxication, offenders would not be able
25

---

DEFENDANTS WHO CLAIM AMNESIA

to store their memories in the first place, which would
Still, according to some leading Anglo-Saxon commenta-
eventually lead to a total “blackout” for their crime.
tors, claims of amnesia may have far-reaching legal implica-
tions (e.g., Bradford & Smith, 1979). Meanwhile, it is not
A number of authors have emphasized that excessive alco-
amnesia per se, but what amnesia reveals about the state of
hol use often contributes to dissociative amnesia for crime
the defendant at the moment he committed his crime.
(Kopelman, 1995; Bourget & Bradford, 1995; Swihart et
More specifically, amnesia may indicate a state of “automa-
al., 1999). Bower’s (1981) “state-dependent memory theory”
tism”, which refers to unconscious, non-intentional, and
is often invoked to account for the combination of dissocia-
therefore uncontrollable behavior. In the literature, eccen-
tive amnesia and alcohol (Swihart et al., 1999). In short,
tric examples of “automatic” crimes committed during
this theory states that when memories are stored in an
sleepwalking, epileptic seizures, or hypoglycaemia states,
exceptional context (strong emotion and/ or under the
abound.
influence of alcohol), subsequent retrieval of these memo-
ries is facilitated when a similar context is reinstated. How-
In countries like Canada and Australia, automatism has
ever, in a different context (e.g., when one is relaxed and
been divided into “sane” and “insane” automatism. For
sober), the pertinent memories would be inaccessible and
both types of automatism, it is assumed that the mens rea
so dissociative amnesia would occur. The case of Sirhan
(i.e., “wicked mind”) aspect of the crime, is at stake
Sirhan, who was held responsible for the murder on Robert
(Hermann, 1986). “Sane” automatism refers to a crime
Kennedy, is often presented as an example of state depend-
committed by someone who is essentially healthy, but who
ent memory. Sirhan claimed that he was unable to remem-
is in a temporary state of madness due to some external
ber the murder. However, when he was hypnotized and
agent (e.g., insulin). Such a scenario may lead to acquittal.
brought into an agitated state, he suddenly said he remem-
“Insane” automatism refers to a crime originated from
bered the details of the murder.
structural brain dysfunction. The prototypical example is
the man who killed his neighbor during an epileptic seizure
As for feigned amnesia, offenders may malinger amnesia for
(Fenwick, 1993). Such a scenario can lead to a “not guilty
a crime in an attempt to obstruct police investigation
by reason of insanity” verdict. Kopelman (1995) notes that
and/or to avoid responsibility for their acts. In an older
“automatic” crimes always fall prey to amnesia. However,
study, Hopwood and Snell (1933) found that 20% of the
the opposite does not hold: when a defendant claims crime-
offenders who claimed amnesia were malingerers. However,
related amnesia, it does not necessarily imply that the crime
there are good reasons to believe that the rate of malinger-
came about automatically (Kalant, 1996). Curiously
ing is actually much higher (see below). The literature pro-
enough, this is nevertheless what the German Bundes-
vides strong examples of defendants who feigned amnesia
gerichtshof seems to think, because it argued that “a veri-
in order to gain tactical advantage in legal procedures. Our
fied amnesia for the criminal act - alone or in combination
favorite example is that of Rudolf Hess, who at the start of
with other factors – is a sign of an emotionally based disor-
the “Nuremberg” trials, claimed to be amnestic for his
der of consciousness” (BGH 4 Str 207/87; see also Barbey,
Third Reich period. A group of prominent psychiatrists
1990).
examined Hess and concluded that his amnesia was genu-
ine. When it became clear to Hess that the amnestic role
Another point that is important to stress is that the distinc-
confers a disadvantage in the sense that one cannot respond
tion between controlled behavior, “sane” automatism, and
to allegations, he suddenly announced during one of the
“insane” automatism is more problematic than it at first
trial sessions that he had fooled the psychiatrists and
glance may appear. Referring to several case examples,
feigned his amnesia (Gilbert, 1971; for a somewhat differ-
McSherry (1998) showed that this distinction is highly
ent interpretation see Picknett, Prince & Prioir, 2001).
dependent on arbitrary judgments made by expert wit-
nesses: it just depends on what a forensic expert is willing to
consider as a structural brain dysfunction. She described
The Legal Context
two similar cases of domestic murder. Due to expert witness
testimonies, one case was classified as “sane” automatism
leading to full acquittal, whereas the other case, was classi-
fied as an example of “insane” automatism, leading to a
Does it matter legally when defendants claim crime-related
“not criminally responsible” verdict and admission to a
amnesia? For the Anglo-Saxon situation Parwatikar, Hol-
psychiatric hospital.
comb and Menninger (1985) noted that: “no court has
found a defendant incompetent to stand trial solely because
of amnesia” (p.202) (see also Hermann, 1986). And, in-
deed, in Anglo-Saxon countries amnesia does not figure in
What science says
the listing of disorders that regularly lead to “not guilty by
reason of insanity” trial outcomes. Typically, such outcomes
involve schizophrenia and mood disorders, especially in
Lay persons’ ideas about what is possible, do not by defini-
combination with alcohol or drug use.
tion have to be wrong. However, for a trial judge who is
confronted with a defendant claiming amnesia, the issue of

26

---

CIMA
what is possible is less interesting. In as much as the judge
this context, hysterical traits refers to manipulative behav-
appreciates expert witness testimonies, he will mostly bene-
ior, including the tendency to feign symptoms (O’Connel,
fit from an expert who educates him about the type of am-
1960).
nesia that the defendant is suffering from and how this
While the concept of hysteria has largely disappeared from
diagnostic judgment was reached. For the expert witness, it
psychiatric vocabulary, the findings of more recent studies
is very difficult to differentiate between dissociative, organic
in this domain point in the same direction. For example,
or feigned amnesia on the basis of interviews with the de-
Lynch and Bradford (1980) reported that claims of amnesia
fendant. This has to do with the fact that simulators can
are often raised by defendants with an antisocial personality
give a compelling imitation of someone with a dissociative
disorder. A hallmark feature of this disorder is, of course,
or organic amnesia. It is only on the basis of psychological
manipulative behavior (Porter et al., 2001). Similarly, Cima,
tests and tasks, that an expert will be able to identify simu-
Merckelbach, Hollnack, and Knauer (2002a) noted that
lators. Nevertheless, our impression is that mental health
criminals who claimed amnesia had low intelligence and
professionals acting as experts in cases of amnesia often use
displayed antisocial personality features. The picture emerg-
interviews with the defendant as the sole source for making
ing from these studies is that defendants who claim disso-
their diagnostic judgments. This state of affairs is hardly
ciative amnesia often rely on a simple form of denial in an
surprising: the average mental health professional is not a
attempt to minimize their responsibility.
memory specialist and has no specialized knowledge about
tests that might be helpful in discriminating between disso-
A second reason to critically consider claims of dissociative
ciative, organic, and feigned amnesia. We will take a closer
amnesia is that the idea of amnesia for crime is based on
look at this distinction below.
the dubious assumption that “the majority of crimes which
are followed by amnesia are those accompanied by strong
emotional reactions” (Hopwood & Snell, 1933, p.32). Ac-
Dissociative versus feigned amnesia
cording to this interpretation, strong emotions lead to
repression or, to use a more recent notion, to dissociation
Dissociative amnesia is defined as “an inability to recall
and this would produce “retrieval” problems. If this line of
important personal information, usually of a traumatic or
reasoning were correct, one would expect that, for example,
stressful nature, that is too extensive to be explained by
victims of concentration camps would also display amnesia
ordinary forgetfulness” (American Psychiatric Association,
for the horrifying events they have experienced. However,
1994: p. 477). A number of authors have argued that the
this is not the case (e.g., Kuch & Cox, 1992; Yehuda, Elkin,
term dissociative amnesia is quite confusing (Pope, Hud-
Binder-Brynes, Kahana, Southwick, Schmeidler & Giller,
son, Bodkin & Oliva, 1998; Kopelman, 1995). It not only
1996). On a related note, eyewitnesses of extreme violence
suggests that the cause of the memory loss is a dissociation
seldom develop amnesia for the events they witnessed (Por-
between consciousness and memory, it also assumes that
ter et al., 2001). In short, the notion that people may de-
memory loss for emotional trauma does exist. Although lay
velop amnesia for events that are accompanied by strong
people find it very plausible that an emotional provocative
emotions, is very controversial (Pope et al., 1998) and,
event, like murder, can lead to memory loss (cf. supra),
therefore, it is wise to consider dissociative amnesia as a
specialists have not yet reached a consensus about this. For
rare phenomenon.
example, some authors (e.g., Kopelman,1995) opine that
A third reason to be skeptical about claims of dissociative
dissociative amnesia does exist and their most important
amnesia is that recent psychiatric literature shows that a not
argument is that defendants who claim amnesia often in-
insignificant minority of people tend to feign a variety of
form the police about the crime: “This makes an account of
symptoms and tend to confabulate stories if this serves their
amnesia as simulation to avoid punishment less plausible”
interests. For example, it is estimated that as many as 20%
(p.435; for a similar reasoning, see Porter, Birt, Yuille &
of closed head injury patients pursuing financial compensa-
Herve, 2001). However, this argument is not very convinc-
tion exaggerate their symptoms (Binder & Rohling, 1996).
ing. A defendant who knows that there is a vast amount of
Likewise, in the United States hundreds of Vietnam veter-
forensic evidence against him, may argue that he will make
ans with Post Traumatic Stress complaints have never even
a more sympathetic impression on triers of fact if he simu-
served in Vietnam (Burkett & Whitley, 1998). When motor
lates amnesia than if he provides them with a lucid descrip-
vehicle accident victims feign neurological complaints and
tion of the crime details. Thus, Sadoff (1974) argues that
when military personnel invent a complete autobiography
most cases of dissociative amnesia are actually feigned.
to qualify for disability payments, why should the criminal
There are three good reasons to take his point seriously.
who simulates amnesia be a rarity? Let’s put it in a different
The first is that there is quite some knowledge about the
way: lay persons as well as many expert witnesses tend to
psychological characteristics of those who claim dissociative
view dissociative amnesia as the rule and feigned amnesia as
amnesia for a crime. Older studies reported that criminals
the exception. Given the considerations discussed above,
who make such claims can be distinguished from other
we think that it would be wise to reverse these probability
criminals by their relatively low intelligence and their hys-
estimates. This conclusion is further supported by studies in
terical traits (O’Connel, 1960; Parwatiker et al., 1985). In
which normal subjects were instructed to play the role of a
27

---

DEFENDANTS WHO CLAIM AMNESIA

murderer who during interrogation is confronted with
The role of expectations
abundant evidence. The most frequently chosen strategy of
these subjects is to claim amnesia for the criminal act and
to attribute it to an internal force that they cannot control
Experiments show that the behavioral effects of alcohol are
(Spanos, Weekes & Bertrand, 1986; Merckelbach, Devilly
to a great extent guided by expectations that people have
& Rassin, 2001a).
about these effects: “Things that are believed real are real in
their consequences” (Thomas & Thomas, 1928: p. 572). A
straightforward procedure to demonstrate this phenome-
The role of alcohol
non is the so-called “balanced placebo design”. The crux of
this design is that some subjects are given a non-alcoholic
drink that they believe to contain alcohol, whereas other
As mentioned earlier, excessive alcohol or drug use is often
persons consume an alcoholic drink that they believe is a
said to precede criminal acts for which later dissociative
non-alcoholic refreshment. Under these circumstances,
amnesia is claimed. At least, that is what criminals who
extravert behavior, tension reduction, and other positive as
claim such memory loss tell researchers and expert wit-
well as negative effects that people associate with alcohol do
nesses. The “state-dependent memory” hypothesis is often
not depend on actual alcohol intake, but on the belief that
invoked to explain the apparent link between alcohol and
one has consumed alcohol (Critchlow, 1986). This finding
amnesia (Swihart et al., 1999), but a close look at the litera-
underlines the fact that people have strong ideas about the
ture shows that this hypothesis is not based on solid evi-
effects of alcohol which in turn may affect their behavior.
dence. For example, in a study of Wolf (1980), a substantial
amount of alcohol was given to criminals who had commit-
Is it possible that a similar expectancy effect occurs in cases
ted murder under the influence of alcohol and who claim
of dissociative amnesia? In other words, is it possible that
to be amnestic. The “state-dependent memory” hypothesis
some defendants claim amnesia for their crime because they
would lead one to predict that alcohol would produce
assume that this is a probable outcome, given the traumatic
complete disappearance of the amnesia, but this is not what
character of the event and the use of alcohol? The case of
happened. Subjects maintained that they could not re-
Gudjonsson, Kopelman, and MacKeith (1999) demon-
member the crime details.
strates that such an imaginary amnesia does exist. In this
case, a defendant was convicted for the murder of a little
Admittedly, the phenomenon of “alcohol blackout” does
girl. During the police interrogations, he could not remem-
exist. However, this phenomenon is best viewed as an or-
ber anything of the crime. Experts interpreted this as a
ganic form of amnesia resulting from an excessive amount
manifestation of dissociative amnesia, a scenario in which
of alcohol consumed within a very short time span (e.g., 5
the defendant himself started to believe, subsequently, it
glasses of whisky or 20 glasses of beer within 4 hours;
became clear that he was innocent.
Goodwin, 1995). Even if subjects consume such a large
dose, only some of them will develop an “alcohol blackout”.
Germane to the issue of expectations is also an experiment
Furthermore, in most cases where dissociative amnesia is
of Christianson and Bylin (1999). These authors gave their
claimed, alcohol doses do not reach such an extreme level.
subjects a case vignette of a murder and subjects were in-
This suggests that in these cases, defendants’ reference to
structed to identify themselves with the offender. Next, one
alcohol or drug use may fulfil a different function, namely
group of subjects was told to play the role of an amnestic
that of the “partial excuse for the essentially inexcusable”
offender during a task that consisted of a series of questions
(Room, 2001: p. 194). That is, for those who have seriously
about the case. The control group was encouraged to per-
violated the law, an appeal to alcohol or drug intoxication
form this task as best they could. After a week, subjects
may give an explanation for the crime that has been com-
returned to the lab and, again, answered questions about
mitted as well as for the memory loss that is claimed. The
the case. This time, all subjects were instructed to perform
example of the Canadian Supreme Court shows that courts
as best they could. During the first testing session, subjects
are not insensitive to such “intoxication defense”. In the
who played an amnestic role gave fewer correct answers
case of a rapist who claimed not to remember the crime
than the control group. This is not remarkable: it only
because he was in a state of alcohol intoxication, this Court
shows that the “amnestic” subjects took their role seriously.
ruled that the crime had been committed in a state of
However, during the one-week follow-up test session ex-
“drunken automatism” (Kalant, 1996). Lay persons not
simulators were still performing under the level of control
only seem to have strong opinions about the plausibility of
subjects. This is remarkable: apparently, “playing” amnesia
dissociative amnesia, they also have strong opinions about
has memory-undermining effects. It is not too far-fetched to
the behavioral and memory effects of alcohol or drugs.
assume that expectations are the driving force behind this
Again, these opinions are not always in line with what is
effect. That is, people who initially play the role of an am-
actually known about these effects.
nestic person may have a strong expectation that they will
perform poorly on subsequent memory tasks. This, in turn,
may give rise to a “self-fulfilling prophecy” when the person
is given such a memory task. This phenomenon is also
known from studies on placebos. Subjects who receive a

28

---

CIMA
placebo in combination with the story that it is a memory-
tion. This is mostly the case in civil law suits, in which, for
undermining substance later perform less well on memory
example, vehicle accident victims require financial compen-
tasks than do control subjects (Kvavilashvili & Ellis, 1999).
sation. In such cases, it is relatively easy to feign aspecific
To sum up, then, a defendant who claims memory loss is
symptoms (Youngjohn, Burrows & Erdal, 1995). Mean-
often classified by experts as someone with dissociative
while, organic amnesia is considerably more difficult to
amnesia, but there are at least two alternative interpreta-
simulate, at least for lay persons, precisely because it has
tions that should be considered: malingered amnesia and
such a typical course. In other words, organic amnesia re-
imaginary amnesia.
quires the specific sequence of trauma, loss of conscious-
ness, PTA, memory loss relating to recent rather than old
memories, and memory recovery in such a way that older
Organic versus feigned amnesia
memories come back more readily than more recent ones.
If a person claims organic amnesia for a traumatic event
Unlike dissociative amnesia, organic amnesia is a relatively
(e.g., accident) that does not fit this sequence, there is every
unproblematic phenomenon. In many cases, organic amne-
reason to be skeptical and to consider the possibility of
sia will be a persistent symptom of traumatic brain injury,
feigned amnesia.
or closed head injury. Note that in this context the word
trauma has a completely different meaning. In the psychiat-
ric literature, it refers to a major emotional event (e.g., a
Tests
crime), whereas in the neurological context it refers to the
external causes (accidents, drugs, fights) of acute brain
damage (Hacking, 1996). Even in cases of mild traumatic
If a defendant claims crime-related amnesia, how should an
brain injury, acute loss of consciousness and subsequent
expert witness determine what type of amnesia the defen-
Post Traumatic Amnesia (PTA) may occur. PTA refers to
dant is suffering from? One possibility is that the defendant
disoriented behavior and serious memory problems imme-
has sustained brain injury and consequently developed
diately after the incident that caused the brain injury.
organic amnesia. The expert may explore this possibility by
When loss of consciousness exceeds 30 minutes and PTA
examining whether and how the defendant’s amnesia dis-
duration is longer than 24 hours, the traumatic brain injury
appears over time. If the defendant’s amnesia follows Ri-
is said to be severe (Faust, 1996).
bot’s law, that information might be crucial for the defen-
Regardless of whether the brain injury is mild or severe, in
dant’s counsel. Consider the example of the defendant
the period after the PTA, the patient usually reports all
charged with murder. If the defendant has organic amnesia
kinds of complaints that vary from concentration difficul-
and it can be shown that this amnesia originates from the
ties to depressive feelings. These complaints are sometimes
victim hitting the defendant on his head before he was
referred to as the “post-concussion” syndrome, but this
murdered, then a self-defense interpretation of the murder
impressive term suggests more clarity than the neurological
case might be considered.
literature really offers. For example, Lees-Haley, Fox, and
Another possibility is that a defendant believes that he is
Courtney (2001) noted that most symptoms associated with
suffering from amnesia. As far as we know, there is no valid
this syndrome are surprisingly aspecific and are also highly
test to explore this possibility. However, with the findings
prevalent among people who never sustained a brain injury.
of Kvavilashvili and Ellis (1999) in mind, the expert might
However, an inability to recall important details of the
consider giving the defendant a placebo along with the
events surrounding the trauma (the incident, the fight) –
instruction that it is a memory-enhancing drug. In a way,
organic amnesia – is a rather specific symptom of post-
such a manipulation is deceptive, but on the other hand, is
concussion syndrome. According to the 19th century French
highly similar to forensic hypnosis, because in hypnosis
memory psychologist Theodule Ribot (Haber & Haber,
expectations also play a key role.
1998), organic amnesia follows a course that is now known
as Ribot’s law. More specifically, this law refers to the phe-
A third possibility is that a defendant feigns his amnesia.
nomenon that organic amnesia pertains to the traumatic
Again, this possibility needs to be scrutinized and this can
incident itself and events that immediately preceded and/or
be done in either of two ways.
followed it, rather than events that took place long before
the trauma. If such older memories have nevertheless be-
Symptom Validity Testing
come inaccessible, they will return sooner in the weeks
following the trauma than more recent memories that have
become inaccessible. Eventually, the amnesia will largely
The first way is to subject the defendant to a so-called
disappear, and will be limited to the traumatic event itself
Symptom Validity Test (SVT). With this method, it is pos-
and the few seconds that preceded it.
sible to identify defendants who simulate amnesia (Freder-
There are reasons to believe that the vague aspecific symp-
ick, Carter & Powel, 1995; Denney, 1996). Basically, SVT
toms of post-concussion syndrome are sensitive to simula-
procedures consist of a forced-choice recognition test. Dur-
ing SVT, the defendant is asked a series of dichotomous
29

---

DEFENDANTS WHO CLAIM AMNESIA

(true-false) questions about the crime and the circumstances
peared quite symmetrical, suggesting that much of the vari-
under which it took place. The defendant is instructed to
ability is due to random chance.
guess in case he does not know the right answers because of
Some researchers have argued that a clever defendant who
his amnesia. Typically, 15 to 100 items are presented, each
attempts to simulate amnesia will readily recognize the
followed by a two-alternative forced-choice recognition.
rationale behind SVT. By this view, defendants would
Typical questions asked in the SVT are for example “Did
quickly realize that they have to perform at chance level
the offender use a pistol or a knife?”; “Was the person
(half of the answers correct and the other half incorrect). A
known to sell ecstasy or crack?”; “Was the victim black or
recent study by Merckelbach, Hauer, and Rassin (2001b)
white?”; and “Was the street made of cobblestones or as-
tested this idea. In this study, 20 students were instructed to
phalt?”. With any number of items, chance performance
steal an envelope with some money. Next, students were
(guessing) can be determined fairly precisely. This has to do
told to simulate amnesia in a way that would convince
with the fact that purely random responding will result in
experts. To explore how well the SVT could identify this
about 50% of the answers being correctly answered. Indi-
feigned amnesia, students took 15 true-false item SVT.
viduals who perform significantly below chance avoid cor-
More than half (53%) of the student sample had less than 4
rect alternatives, which means that they have knowledge
correct answers and, thus, they were identified as malinger-
about the correct answers, and this implies that they are
ers. The other students succeeded in performing at chance
feigning memory impairment. SVT is based on binomial
level and thus, seemed to be able to simulate in a convinc-
statistics and this has the clear advantage that one can
ing way. However, post-experiment interviews with the
quantify memory performance.1 Thus, one can determine
subjects made it clear that only a minority of them were
the exact chance that someone with genuine memory loss
able to verbalize the rationale behind SVT.
will give only 3 right answers to 15 true-false questions. On
the basis of chance such a person should have 6, 7 or 8
Given the fact that a majority of students cannot beat the
correct answers. The chance that someone with memory
SVT, the efficacy of the SVT in identifying malingering
loss will produce only 3 correct answers is smaller than 5%
should be considerably better with less educated defendants
(see footnote 1; the exact chance can be calculated as fol-
who simulate amnesia. This is especially true if the number
lows: z = (3 + .5 – (15 x .5)/ √(15 x .5 x .5) = 2.1, which
of SVT questions is increased. The Merckelbach et al.
corresponds to a p-value of ≤ 0.02).
(2001b) experiment relied on a relatively small number of
test items (i.e., 15). However, with more SVT items, say 30,
Several case studies have been reported in which suspected
it becomes more and more difficult for defendants to moni-
(Frederick & Carter, 1993; Frederick, Carter & Powel,
tor whether one is still performing at chance level. And
1995) or confirmed (Denney & Wynkoop, 1995) malinger-
even when defendants perform at chance level, there is an
ers showed a response pattern in which their performance
appropriate test to examine whether the pattern of correct
fell below chance. One study (Denney, 1996) presents three
and incorrect answers is random (as it should be in the case
cases within the criminal forensic setting. To evaluate the
of genuine amnesia) or structured (as is the case in a strate-
assumption that the procedure actually conforms to prob-
gic attempt to perform at chance level) (Cliffe, 1992). A
ability standards of the binomial theorem, this author used
discussion of this “runs” test, falls beyond the scope of this
a normative sample of 60 adults who had no prior knowl-
article, but the principle behind it can be summarized as
edge of the events in question. Results of this study showed
follows. Suppose you have a coin and on 15 consecutive
that this test procedure is rather robust and can handle
trials you throw heads, while on the next 15 consecutive
variability in the probability of several items. Although the
trials you throw tails. That would be nicely at chance level.
more questions created the more reliable the test and the
However, there must be something wrong with the coin,
more test variance it will contain (Crocker & Algina, 1986),
because too few so-called “runs” have been made (namely 2)
the results of Denney (1996) suggests that even with a test
to assume that the pattern of heads and tails is random.
of only 10 items, the varied item probability had little prac-
The same is true for some forms of SVT performance at
tical effect on the test. The sample (N = 60) means for each
chance level. Those who during a SVT of 30 true-false items
case were shown to be either the same as the hypothesized
answer the first 15 items correctly and the last 15 items
mean or were in the positive direction. For example, in one
incorrectly, perform nicely at chance level, but are identi-
of the cases described, the defendant answered only 7 of 29
fied as malingerers by the “runs” test. So, even if the results
questions correctly (z = -2.6, p< 0.005) which suggests that
fall within the predictably random range and distribution,
such a performance would only occur less than 5 times out
but the response pattern shows that the first half of the test
of 1000 by chance alone. The sample results of this case
is answered correctly and the second half is not, it indicates
showed a mean of 15.83 (SD = 2.77) and a range from 8 to
a structured response pattern. To reveal malingerers one
23. Moreover, the distribution of item probabilities ap-
should not calculate the approximately equal numbers of

true and false answers only, but should also consider
1 The Binominal formula is as follows: z = [(x ± .5) – Np)] / √Npq,
whether the sequence of the true and false answers is at
in which z is the test statistic (value and corresponding p can be
random.
looked up in a table), N is the number of items, x the number of
correctly answered items, p the chance for a correct answer when
one has to guess (idealy .5) and q is 1-p.

30

---

CIMA
SVT does not require much technical equipment. All one
ing in that subjects instructed to feign, say, amnesia can be
needs is a pencil, paper, and a basic knowledge of statistics.
identified with a high degree of precision. For example,
It is essential, though, that the correct and incorrect alter-
Merckelbach and Smith (2001c) reported that more than
natives are first evaluated by a panel of inexperienced sub-
90% of the subjects instructed to malinger amnesia are
jects. If this panel judges the incorrect alternatives as more
identified by the SIMS (sensitivity) and more than 90% of
plausible than the correct alternatives, it is possible that
the control subjects are classified by the SIMS as honest
someone with genuine amnesia will perform below chance
respondents (specificity).
level. In the clinical or forensic setting, this method of
In the one study that systematically looked at SIMS per-
control might prove unwieldy and extremely time consum-
formance of forensic patients (Cima et al., 2002b) a Ger-
ing. Nevertheless, with this restriction in mind, we would
man translation of the SIMS was administered to a sample
like to recommend the SVT to experts who have to examine
consisting of 266 participants of whom 204 were under-
cases in which defendants claim amnesia.
graduate students (98 men) and 62 were forensic inpatients
(all men). The German SIMS showed excellent stability
Self-reports
(0.97) as well as an acceptable consistency (Cronbach alpha
coefficient = 0.80). None of the participants in the control
sample (69 honestly responding healthy subjects) had a total
A second way of examining claims of amnesia, is provided
score exceeding the cutoff point of 16 recommended by
by self-report questionnaires that capitalize on the tendency
Rogers et al. (1996), while 86% of the instructed malinger-
of malingerers to exaggerate their memory complaints
ers (n = 135) and 27% of the forensic patients scored in this
(Smith, 1997). In this context, a promising questionnaire is
range.
the Structured Inventory of Malingered Symptomatology
(SIMS; Smith & Burger, 1997; for a German translation
Data of honestly responding controls, instructed malingers
see: Cima, Hollnack, Kremer, Knauer, Schellbach-Matties,
and patients were pooled in order to calculate sensitivity,
Klein & Merckelbach, 2002b). The SIMS consists of 75
specificity, positive and negative predictive power (PPP and
dichotomous (i.e., true-false) items that can be grouped into
NPP). Using the cutoff score of 16, 87% of the malingerers
5 subscales, each subscale containing 15 items. Subscales
were identified correctly (sensitivity), while 86.7% of the
correspond to symptoms domains that are sensitive to ma-
non-malingerers were classified correctly (specificity). The
lingering and include low intelligence (LI), affective disor-
Positive Predictive Power (PPP), which is explained in the
der (AF), neurological impairment (N), psychosis (P), and
next section, was 0.87. In other words, the probability that
amnestic disorder (AM). Items of the subscales refer to
someone with a Total SIMS score of 17 or higher was an
bizarre experiences (e.g., “At times I’ve been unable to re-
instructed malingerer was 87%. The specificity of 86.7%
member the names or faces of close relatives so that they
implies a false alarm rate of 13.3%. In other words, 13.3%
seem like complete strangers”) or to unrealistic symptoms
of the participants who did not receive an instruction to
(e.g.,”When I can’t remember something, hints do not
malinger nevertheless had Total SIMS scores exceeding the
help”). Other items explicitly allude to a certain syndrome
cutoff point. A closer look at these participants revealed
(e.g., amnesia) in such a way that specialists recognize that
that all of them were forensic psychiatric inpatients.
highly atypical symptoms are listed (e.g., “My past and im-
portant events became a blur to me almost overnight”). The
idea is that malingerers will exaggerate and so will endorse
Diagnostic accuracy
bizarre, unrealistic, and atypical symptoms. Answers indicat-
ing malingering are summed to obtain a Total SIMS score.
So far, a number of analog studies have looked at the accu-
Our discussion so far makes plain that experts who at the
racy with which the SIMS detects malingered symptomatol-
request of the court have to evaluate a case in which crime-
ogy (Rogers, Hinds & Sewell, 1996; Smith & Burger, 1997;
related amnesia is claimed can and should do more than
Edens, Otto & Dwyer, 1999). Although these studies came
just interview the defendant. The SVT as well as the SIMS
up with promising results, one should note that with the
are examples of tests that can be applied in this type of case.
exception of the study by Cima and colleagues (2002b), all
Other tests and tasks that can be of help are extensively
studies relied on laboratory set-ups. In this type of study,
discussed in Rogers (1997) and Hall and Poirier (2000).
undergraduate students are instructed to feign certain psy-
Selection and actual use of tests in amnesia cases should be
chiatric symptoms (e.g., amnesia) in a convincing way. Per-
guided by scientific literature and, more specifically, by
formance of these instructed malingering groups on the
what this literature tells the experts about diagnostic accu-
SIMS is then compared to the SIMS scores of control (i.e.,
racy. Experts usually focus on sensitivity and specificity of
uninstructed) groups responding honestly. It is evident that
diagnostic measurements. Sensitivity refers to the percent-
this approach is subject to a number of limitations, not
age of malingerers who are correctly classified by the test,
least of which is that for undergraduates instructed to ma-
while specificity refers to the percentage of honestly re-
linger symptoms there are hardly any risks or incentives.
sponding controls who are correctly identified. The idea
Nevertheless, so far the results with the SIMS are encourag-
that a diagnostic instrument is valid if research shows that
31

---

DEFENDANTS WHO CLAIM AMNESIA

its sensitivity and specificity circle around 90% seems
lent form of crime-related amnesia. Experts should at least
widely accepted by experts, but is misleading. An example
take the possibility of other forms of amnesia into account.
taken from Rosenfeld, Sands and van Gorp (2000) may
These other types involve simulated, imagined or organic
illustrate this point. Suppose that there is a questionnaire
amnesia. The diagnostic differentiation between these types
that intends to detect simulated amnesia and suppose that
of amnesia cannot take place solely on the basis of inter-
this instrument has been tested in a laboratory study that
views. Professionals should use objective measures as op-
involved 50 subjects simulating amnesia and 50 honestly
posed to relying on clinical judgement alone. Tests (e.g., the
responding subjects (control group). Let us further assume
SIMS) and tasks (e.g., the SVT) are required to make this
that the results of this study showed that 45 of the 50 simu-
differentiation. As with interviews, the use of only one test
lators were identified by this questionnaire (sensitivity =
is not sufficient in determining between the different types
90%), while 48 of the 50 control subjects were classified
of amnesia. No one should be diagnosed as malingering
correctly (specificity = 96%). So, sensitivity and specificity
simply from one test. Instead, proper diagnosis entails the
rates are excellent, but this does not mean that the test is
appropriate synthesis of information from multiple data
valid. The point is that these rates were obtained with a
sources (structural interviews, observation, other test re-
particular base-rate of malingering, namely 50% (50% of
sults), especially in the forensic setting. Statistically based
the subjects were simulating amnesia). If one lowers this
procedures, when used judiciously and in conjunction with
base rate to, say, 10% (10 subjects simulate versus 90 sub-
multiple data sources, should offset the inherent weak-
jects respond honestly), diagnostic accuracy of the test will
nesses of clinical judgement alone. Although these tests and
drop dramatically, although its sensitivity and specificity
tasks may seem simple, their use requires sophisticated
remain the same. Table 1 shows this effect. With a base rate
knowledge about diagnostic quality parameters. Many psy-
of 50%, 45 of the malingerers are correctly identified and 2
chologists and psychiatrists will not possess this knowledge
honestly responding subjects are misclassified as malinger-
and therefore their competence to act as expert witnesses in
ers. However, with a base rate of 10%, there are 9 correctly
cases in which claims of crime-related amnesia are raised is
classified malingerers and 4 honest subjects who are mis-
doubtful.
classified. A good measure to capture these proportions is
the Positive Predictive Power (PPP). PPP is concerned with
Table 1: Sensitivity and specificity of a diagnostic in-
the classification of the individual and this is what matters
strument and how they relate to base rate. Values in
in a legal context. PPP pertains to the probability that an
parentheses show what happens when the base rate of
malingering drops to 10%: 10 subjects malinger and 90
individual with a score that exceeds the cutoff does have the
perform honestly. TP = true positive, FP = false posi-
diagnosis (in this case, feigned amnesia). With a base rate of
tive, FN = false negative and TN = true negative. Sensi-
50%, the PPP of our fictitious instrument was 45/45 + 2 =
tivity = TP/(TP+FN); specificity = TN/(FP+TN); PPP
96%. However, with a base rate of 10%, the PPP of this
(Positive Predictive Power) = TP/(TP+FP)
instrument drops to 9/9+4 = 69%. In other words, with
Reality

such a base rate, the probability is 2 out of 3 that someone
who has an above-cutoff score does indeed simulate amne-
Simulation
Honest
sia. The take-home message of this example is that applying
Test:


tests in order to identify simulated amnesia requires techni-
Simulation
45 (TP)
2 (FP)

(9)
(4)
cal knowledge about diagnostic quality parameters (sensitiv-
Honest
5 (FN)
48 (TN)
ity, specificity and PPP) of these tests. This is essential,
(1)
(86)
because it is only when the expert possesses such technical
Sensitivity
0.90

knowledge and communicates it to the court that trial
Specificity
0.96
judges become aware of strengths and limitations of diag-
PPP
0.90
nostic evaluations in amnesia cases.

(0.69)
Conclusion
References
It is not uncommon for defendants to claim amnesia for
American Psychiatric Association. (1994). Diagnostic and
a crime of which they are accused. Mental health profes-
Statistical Manual of Mental Disorders (4th ed.). Wa-
sionals who appear as expert witnesses in such cases often
shington , DC: Author.
assume that this type of memory loss is the joint effect of
Barbey, I.(1990). Postdeliktische Erinnerungsstörungen:
strong emotions (i.e., dissociative amnesia) and excessive
Ergebnisse einer retrospektiven Erhebung. Blutalko-
drug or alcohol use. It is also common to seek arguments
hol, 27, 241-259.
for this interpretation by interviewing the defendant. In
Binder, L.M. & Rohling, M.L. (1996). Money matters: A
this article, we explained why this approach is dubious.
meta-analytic review of the effects of financial incen-
Dissociative amnesia is not the only or even the most preva-

32

---

CIMA
tives on recovery after closed-head injury. American
Frederick, R.I. & Carter, M. (1993). Detection of malin-
Journal of Psychiatry, 153, 7-10
gered amnesia in a competency evaluee. Paper pre-
Bourget, D. & Bradford, J.M.W. (1995). Sex offenders who
sented at the annual meeting of the American Psy-
claim amnesia for their alleged offense. Bulletin of
chological Association, Toronto.
the American Academy of Psychiatry and Law, 23,
Frederick, R.I., Carter, M. & Powel, J. (1995). Adapting
299-307.
symptom validity testing to evaluate suspicious com-
Bower, G.H. (1981). Mood and memory. American Psy-
plaints of amnesia in medicolegal evaluations. Bulle-
chologist, 36, 129-148.Bradford, J.W. & Smith,
tin of the
S.M. (1979). Amnesia and homicide: The Padola
American Academy of Psychiatry and the Law, 23, 227-
case and a study of thirty cases. Bulletin of the
233.Gilbert, G.M. (1971). Nürnberger Tagebuch:
American Academy of Psychiatry and Law, 7, 219-
Gespräche der Angeklagten mit dem Gerichtspsy-
231.
chologen. Frankfurt am Main: Fischer.
Burkett, B. G. & Whitley, G. (1998). Stolen valor: How the
Goodwin, D.W. (1995). Alcohol amnesia. Addiction, 90,
Vietnamgeneration was robbed of its heroes and its
315-317
history. Dallas, TX: Verity Press.
Gudjonsson, G.H., Kopelman, M.D. & MacKeith, J.A.C.
Cima, M., Merckelbach, H., Hollnack, S. & Knauer, E.
(1999). Unreliable admissions to homicide: A case
(2002a). Correlates of amnesia in psychiatric prison
of misdiagnosis of amnesia and misuse of abreaction
inmates. Submitted.
technique. British Journal of Psychiatry, 174, 455-
Cima, M., Hollnack, S., Kremer, K., Knauer, E., Schellbach-
459.
Matties, R., Klein, B. & Merckelbach, H. (2002b).
Gudjonsson, G.H., Petursson, H., Skulason, S. &
Strukturierter Fragebogen Simulierter Symptome.
Sigurdardottir, H.(1989). Psychiatric evidence: A
Die deutsche Version des Structured Inventory of
study of psychological issues. Acta Psychiatrica
Malingered Symptomatology: SIMS [Structured In-
Scandinavica, 80, 165-169.
ventory of Simulated Symptoms: The German ver-
Hacking, I. (1996). Rewriting the Soul: Multiple Personality
sion of the Structured Inventory of Malingered
and the Sciences of Memory. Princeton NJ: Prince-
Symptomatology: SIMS]. Submitted.
tom Universty Press.
Christianson, S-A. & Bylin, S. (1999). Does simulating
Haber, L. & Haber, R.N. (1998). Criteria for the admissi-
amnesia mediate genuine forgetting for a crime
bility of eyewitness testimony of long past events.
event? Applied Cognitive Psychology, 13, 495-511.
Psychology, Public Policy, and Law, 4, 1135-1159.
Cliffe, M.J. (1992). Symptom-validity testing of feigned
Hall, H.V. & Poirier, J.G. (2000). Detecting Malingering
sensory or memory deficits: A further elaboration
and Deception: Forensic Distortion Analysis. Lon-
for subjects who understand the rationale. British
don: CRC Press.
Journal of Clinical Psychology, 31, 207-209.
Hermann, D.H.J. (1986). Criminal defenses and pleas in
Critchlow, B. (1986). The powers of John Barleycorn: Be-
mitigation based on amnesia. Behavioral Sciences
liefs about the effects of alcohol on social behavior.
and the Law, 4, 5-26.
American Psychologist, 41, 751-764.
Hopwood, J.S. & Snell, H.K. (1933). Amnesia in relation to
Denney, R.L. & Wynkoop, T.F. (1995). Exaggeration of
crime. Journal of Mental Science, 79, 27-41.
neuropsychological deficits in pretrial evaluation: A
Kalant, H. (1996). Intoxicated automatism: Legal concept
case presentation. Paper presented at the 103rd an-
vs. scientific evidence. Contemporary Drug Prob-
nual meeting of the American Psychological associa-
lems, 23, 631-648.
tion, New York.
Kihlstrom, J.F. & Schacter, D.L. (1995). Functional disor-
Denney, R.L. (1996). Symptom validity testing of remote
ders of autobiographical memory. In In A.D
memory in a criminal forensic setting. Archives of
Baddeley, B.A. Wilson & F.N. Watts (Eds.). Hand-
Clinical Neuropsychology, 11, 589-603.
book of Memory Disorders (pp. 337-364). New
Edens, J.F., Otto, R.K. & Dwyer, T. (1999). Utility of the
York: Wiley.
Structured Inventory of Malingered Symptomatol-
Kopelman, M.D. (1995). The assessment of psychogenic
ogy in identifying persons motivated to malinger
amnesia. In A.D Baddeley, B.A. Wilson & F.N.
psychopathology. Journal of the American Academy
Watts (Eds.). Handbook of Memory Disorders (pp.
of Psychiatry and the Law, 127, 387-396.
427-448). New York: Wiley.
Faust, D. (1996). Assessment of brain injuries in legal cases:
Kopelman, M.D., Green, R.E.A., Guinan, E.M., Lewis,
Neuropsychological and neuropsychiatric considera-
P.D.R. & Stanhope, N. (1994). The case of the am-
tions. In B.S. Fogel, R.B. Scheffen & S.M. Rao
nestic intellicgence officer. Psychological Medicine,
(Eds.).
24, 1037-1045.
Neuropsychiatry (pp. 973-990). Pennsylvania: Williams &
Kuch, K. & Cox, B.J. (1992). Symptoms of PTSD in 124
Wilkins
survivors of the Holocaust. American Journal of Psy-
Fenwick, P. (1993). Brain, mind, and behaviour: Some
chiatry, 149, 337-340.
medico-legal aspects. British Journal of Psychiatry,
163, 565-573.
33

---