The nature of the defence

The automatism defence has no statutory basis; no Act of Parliament defines it, sets its limits and governs its application. It is a common law defence, depending on judicial decisions made over the years and in different jurisdictions. Different judges, dealing with quite dissimilar cases, have produced differing definitions (Box 2). There are two particular features: the lack, or an altered state, of consciousness and the absence of volition.

The internal/external distinction

The internal/external distinction in automatism is particularly troublesome (Table 1). Where automatism results from an ‘external’ cause such as head injury, it is a ‘non-insane (sane) automatism’ (hereinafter ‘sane automatism’). It is a complete defence. It results in acquittal. Where it is an ‘internal’ cause such as epilepsy, it is an ‘insane automatism’, it amounts to (legal) insanity and disposal is under section 24 of the Domestic Violence, Crime and Victims Act 2004 (the DVCA), which has amended the disposal options under the Criminal Procedure (Insanity and Unfitness to Plead) Act 1991 (the CP(IUP)A).

TABLE 1 The current classification of automatisms

Various problems result. It is quite inconsistent with medical, if not public, thinking to categorise someone with diabetes, for example, as insane. The stigma of ‘insanity’ can be a disincentive to raising the defence. Outcomes can be illogical, if not nonsensical. Someone with diabetes who fails to take their insulin and, in a hyperglycaemic state, carries out what is alleged to be a criminal act is treated as insane. If they adhere to their insulin regime but fail to eat and, in a state of hypoglycaemia, carry out the same act they are acquitted. Internal and external factors can operate together. The jury have not only to consider their relative contributions, but to do so with reference to differences in relation to burden and standard of proof.

The evidential burden, the burden of proof and the standard of proof

Where the defence of automatism is raised, the judge has to decide first whether it has a proper evidential foundation (‘the evidential burden’). If not, it cannot proceed.

If there is a proper evidential foundation, the judge has to decide whether it is insane or sane automatism. If it is insane automatism, the burden of proof (‘the legal burden’) rests on the defence to prove its case on a balance of probabilities (what is known by convention as ‘the civil standard’). If it is sane automatism, the prosecution bears the legal burden of proving all the elements in the offence necessary to establish guilt and it has to do so to ‘the criminal standard’ of proof, which means ‘beyond reasonable doubt’ ( Woolmington v DPP [1942]). This means that the prosecution must prove beyond reasonable doubt that there was a voluntary act that satisfied the actus reus requirement for the offence. This is a high hurdle for the prosecution. It is sufficient for the defence to raise no more than a doubt as to the defendant’s guilt for the prosecution to be unable to prove its case beyond reasonable doubt.

Where there are issues of both insane and sane automatism, the judge has to distinguish them in summing up to the jury and remind them of the different burdens and standards of proof. It is a tall order for many jurors to make sense of this.

Prior fault

The defence may be difficult or rendered invalid where the accused was at fault in getting into such a condition.

It was withheld from a driver who fell asleep at the wheel, because he should have pulled over when he began to feel drowsy ( Kay v Butterworth (1945)), and from another who drove while sleepwalking after drinking alcohol, because he had previously experienced sleepwalking episodes after drinking ( Finegan v Heywood (2000)). It failed where a driver suffered an acute hypoglycaemic episode: he should not have been driving because 3 months previously he had been involved in an accident when hypoglycaemic ( R v Marison (1997)). Sentencing a driver with diabetes for causing death by dangerous driving, the judge said that ‘monitoring of his condition and his blood sugar levels in particular were to put it mildly grossly mismanaged and grossly inadequate’ (Law Commission 2013: p. 236, para. B.54).

Even a driver who has swerved in response to a wasp that has flown into their car may be at fault for not stopping to deal with the distraction.

There are, however, examples of successful reliance on states amounting to automatism notwithstanding prior fault. A driver with an insulinoma was stopped by police for a driving offence when hypoglycaemic but successfully relied on hypoglycaemia in the civil action that was brought after he crashed into a house later the same day ( Mansfield v Weetabix [1998]). When the judge heard evidence about a woman’s insulin-dependent diabetes, he directed her acquittal of causing death by dangerous driving on the basis that she was not at fault ( R v Gilbert [2006]). However, she had previously suffered three hypoglycaemic episodes which had come without warning and, although recommended to check her glucose level before driving, had not done so.

Intoxication

There is a general rule that self-induced intoxication, however caused, is not a defence to offences of basic intent or recklessness. This applies to automatism. A man who took LSD and killed his girlfriend in the belief that she was a snake had his conviction for manslaughter upheld because his automatism was self-induced ( R v Lipman [1969]).

Degree of impairment

The rule is that there has to be a complete destruction of voluntary control. Impaired, reduced or partial control is not enough.

A driver with diabetes was, over the course of 5 miles, able to react to stimuli and take control of his limbs, albeit perhaps imperfectly; he drove at appropriate speeds, veered away from a vehicle, braked behind a stationary traffic queue and, although incoherent when someone opened his car door, restarted the engine, drove home and parked the car ( Broome v Perkins (1987)). This was regarded as evidence that there was not a complete destruction of voluntary control and his conviction of driving without due care and attention was upheld.

The defence also failed where, following a collision, the driver’s ability to circumvent a police road block was regarded as evidence of responding to his surroundings and being in full voluntary control ( R v Isitt (1978)).

This rule is applied inconsistently. In sleepwalking cases, the ability to start a car engine does not seem to raise an issue as to the extent to which the mind controls the body. The defence was successful where a sleepwalking defendant started the engine to drive to the police station having already driven several miles, through seven sets of traffic lights, to his in-laws, where he killed his mother-in-law and attempted to kill his father-in-law ( R v Parks (1992)).

Sleep disorders

The case in Box 3, described by Reference Ebrahim and FenwickEbrahim & Fenwick (2008), is illustrative of sleepwalking. This was held legally to be a case of insane automatism. The defendant was ordered to be detained in hospital (this being prior to the DVCA, because the offence charged was murder, a disposal of a hospital admission order with restrictions under the CP(IUP)A was the only disposal available to the court).

Box 4 illustrates a case of sexsomnia. Potential weaknesses here were that the defendant was not disoriented on waking, the complainant recalled a conversation indicative of a high level of cognitive functioning and the defendant had partial recall of events. With evidence of both internal and external factors, the court recognised difficulties. The defence evidence was that there was (a) an established history of sexsomnia (internal), (b) something to induce a sleep disorder (external) and (c) proximity (external). As the external factors were sufficient evidence for sane automatism to be left to the jury, the judge’s direction was that in order to convict they had to be sure that the defendant was not in an automatic state at the time of penetration. He was acquitted.

Hysterical fugue

In R v Isitt (1978) it was asserted that the driver of a stolen vehicle was in a hysterical fugue so his subconscious mind was in control and therefore he would not have appreciated what he was doing. However, his ability to circumvent a police road block was regarded as evidence of being in full voluntary control. His defence failed. This is probably illustrative of the recommendation that ‘the psychiatrist who deposes to dissociation in improbable circumstances should be subjected to skilled and deeply sceptical cross-examination, and […] the Crown should, where possible, call counter-evidence’ (Reference WilliamsWilliams 1983).

Personality disorder

Box 5 concerns a man convicted of wounding with intent ( R v Roach [2001]). The defence relied on two psychiatrists who were of the opinion that this was ‘insane automatism of psychogenic type’. One diagnosed an antisocial personality disorder. His evidence was:

The other listed three contributory factors:

1. (1) ‘Mixed Personality Disorder in the form of “an inclination to impulsivity, but an avoidance of violence (he tells me)”;

2. (2) “A long-standing learnt sensitivity to humiliation from aggressive figures”;

3. (3) Impairment […] of his “higher mental functions” by a mixture of Carbamazepine […] 200 mg twice daily, Paroxetine […] 20 mg […] daily, two large vodkas and fatigue. The appellant had had life stresses in his previous six months consisting of a deteriorating relationship with his partner with outbursts of verbal violence and violence to their property and the failure of his plan to become a single parent family, perceived as sabotaged by his parents, particularly his father. His ‘immediate stressor’ was feeling abused and humiliated, ongoing for two days, by his (in his view) incompetent supervisor’.

The evidence of the prosecution’s psychiatrist was that the amnesia, even if genuine, did not necessarily denote inability to form the required intent. He found no evidence that this was a psychogenic automatism and pointed out that, if the defendant had been in such a state, the man would have been bewildered following arrest. His opinion was that the behaviour was entirely consistent with someone who had become increasingly anxious and lost his temper. The defendant was convicted.

It was a ground for appeal that, in his summing up, the judge had failed to draw attention to the evidence of two witnesses which provided some ambivalent support for the case that the appellant had not realised what he was doing. This was accepted and, also for other reasons, the Court of Appeal concluded that the conviction was unsafe and quashed it. This conclusion was reached with some reluctance, as it seemed to the Court that the case against the appellant was strong and that the jury had almost certainly rejected the defence on the basis of a strong preference for the prosecution psychiatrist’s assessment and evidence.

Psychosis

In the case of R v Coley [2013], where Mr Coley was convicted of attempted murder carried out after consuming a great deal of cannabis and suffering a ‘brief psychotic episode’, the Court of Appeal upheld the trial judge’s decision not to allow the defence of automatism. However, it thought that there was an issue as to whether the appellant had been acting consciously. It accepted that his mind may well have been affected by delusions or hallucinations ‘and in that sense his detachment from reality might be described by some as an absence of conscious action’, but it went on to hold that such a condition ‘clearly falls short of involuntary’ and observed:

Stress and post-traumatic stress disorder

Although ‘stress’ is obviously an external factor, the courts have not allowed it to found a defence of sane automatism because ‘the ordinary stresses and disappointments of life which are the common lot of mankind do not constitute an external cause’ ( R v Rabey [1980]). As they do not make everyone behave like an automaton, the fact that someone has done so is regarded as evidence that the real cause of the behaviour lies in their make-up, it is internal, an insane automatism.

Box 6 is a case in which post-traumatic stress disorder (PTSD) resulting from rape was relied on as the basis for an automatism defence, but the defendant was found guilty ( R v Tate [1990]). This is not surprising. Following her victims, opening the penknife blade and demanding the second victim’s handbag suggest that her mind went with her actions. However, it is an important case as it has established that because rape cannot be regarded as one of the ‘ordinary stresses of life’ and ‘such an incident could have an appalling effect on any young woman, however well-balanced normally’, the defence raised is not one of insane automatism resulting from a disease of the mind, i.e. PTSD, but sane automatism resulting from a malfunctioning of the mind due to an external factor, i.e. rape.

History

Enquiry may reveal a family history of a relevant condition, for example a family history of sleepwalking or, in cases of sexsomnia, of arousal disorders. The judge will assist the jury as to the weight to be attached to such a history.

Enquire as to previous psychiatric history. Major depressive disorder, obsessive–compulsive disorder (Reference Ohayon, Mahowald and DauvilliersOhayon 2012) and alcohol misuse/dependence (Reference Moldofsky, Gilbet and LueMoldofsky 1995; Reference Ebrahim and FenwickEbrahim 2012; Reference Ohayon, Mahowald and DauvilliersOhayon 2012) increase the risk of sleepwalking. There is an association between sleep apnoea and depression (Reference OhayonOhayon 2003).

A history of mild sleep disordered breathing (snoring) or sleep apnoea is important, as these are the most commonly identified triggers for sleepwalking. There is an association between sleep apnoea syndrome and sexsomnia (Reference Ohayon, Mahowald and DauvilliersOhayon 2012). Excessive daytime sleeping may be a pointer to sleep apnoea/hypopnoea syndrome.

Enquire as to drug use, which is a recognised trigger for sleepwalking (Reference Moldofsky, Gilbet and LueMoldofsky 1995; Reference Ohayon, Mahowald and DauvilliersOhayon 2012). Enquire as to alcohol use. The finding that a minority of patients report an association between alcohol and sleepwalking (Reference Lopez, Jaussent and ScholzLopez 2013) confirms the generally held view of some experts (e.g. Reference Ebrahim and FenwickEbrahim 2009, Reference Ebrahim and Fenwick2012) that there is a positive association between alcohol and sleepwalking, but other experts disagree strongly (e.g. Reference Pressman, Mahowald and SchenckPressman 2009). Furthermore, it should be noted that the International Classification of Sleep Disorders (American Academy of Sleep Medicine 2014: p. 237) states that:

However, this prohibition as to diagnosis is not universally accepted. For example, Reference Rumbold, Riha and MorrisonRumbold et al (2014) state:

That this is such a controversial area is an indication that in cases of what appears to be sleep disorder, most psychiatrists are likely, indeed advised, to recommend the further instruction of one of the UK’s very few sleep disorder experts.

Stress, strong emotions, sleep deprivation, alcohol and intense physical activity have been identified as triggers for increasing the frequency and severity of sleepwalking episodes in non-rapid eye movement (non-REM) sleep arousal disorders (Reference Lopez, Jaussent and ScholzLopez 2013).

Enquire whether the accused has previously experienced symptoms the same as, or similar to, those experienced at the material time. In a sleepwalking case, a childhood history of sleepwalking will be highly relevant. This is common in adult sleepwalkers. Onset is less common in adolescence and uncommon in adulthood. Reference Lopez, Jaussent and ScholzLopez et al (2013) found that 17% of adult sleepwalkers had onset as an adult. However, it can occur following a head injury or febrile illness or as a side-effect of medication, particularly antidepressants, major tranquillisers and minor tranquillisers. In sexsomnia there is usually a history of sleep disorders. Such a history needs to be brought to the attention of the court so that it can decide whether this amounts to prior fault.

Obtain an account of the accused’s history just before and at the material time. The accused may report nothing unusual prior to a period of apparent amnesia that includes the material events. Their account may include a report of potentially significant symptoms prior to a period of apparent amnesia for the material events. Their account may include a recollection of events that occurred at the material time.

This account may usefully start with what had happened from the beginning of the day. Sometimes there is a longer relevant history, as stressful life events, fatigue and sleep deprivation can trigger sleepwalking. Excess alcohol, too little sleep and stress are recognised triggers for sexsomnia (Reference Moldofsky, Gilbet and LueMoldofsky 1995). An incident in the early hours of the morning or early afternoon can suggest obstructive sleep apnoea/hypopnoea.

In someone with a history of diabetes, find out whether they checked their blood glucose and, if it was low, whether they ate some food. Enquire as to physical exertion. Someone with established diabetes might be expected to know that a particular level of exercise would make them vulnerable to hypoglycaemia.

Where an accused person reports such symptoms as profuse sweating, rapid heart beat and tremor and has a history of, or has been subsequently diagnosed with, diabetes, this may be consistent with acute hypoglycaemia. Failure to respond appropriately to premonitory symptoms can rule out the defence. Likewise, for a driver, failure to pull over in response to a partial loss of control can raise an issue of prior fault.

The last recollection prior to the incident can be critical. Box 7 is the case of a lorry driver charged with causing death by dangerous driving. His last recollection of experiencing something ‘like a welder’s arc flash’ was relied on as evidence that he had suffered an occipital lobe seizure.

An accused person may recall nothing abnormal prior to a period of apparent amnesia for the material time but have a history of diabetes, including a history of what may be subacute neuroglycopaenia or hypoglycaemic unawareness, and therefore a history of episodes in which there is no awareness of the usual premonitory warning signs of acute neuroglycopaenia. Previously unrecognised hypoglycaemic unawareness is very rare so, where it is raised as a defence, it is highly likely that the defence will fall foul of the prior fault principle.

What does the accused recall of the material events? Complete amnesia is usual. This is to be expected if there is unconsciousness. Some partial recollection may be consistent with automatism if there has been an impairment of voluntary control of actions. There may be some fragments of distorted memory in cases of sleepwalking.

What are the first memories following the episode? Usually, mentation on waking from sleep walking is of non-narrative and non-dreamlike experiences with only a vague visual content and consisting mostly of thoughts and feelings (Reference Ebrahim and FenwickEbrahim 2008). There may be disorientation or bewilderment.

Examination

The nature and extent of physical examination will depend on the medical condition at issue and any physical examination already carried out by more appropriately qualified experts.

Mental state examination may be normal. Abnormal mental state findings may be relied on in support of psychiatric diagnosis, but are unlikely to assist as to mental state at the time of the alleged automatism.

Investigations

Obtain the accused’s medical records. They will assist in verifying, establishing or perhaps even refuting the accused’s reported medical history. Of particular assistance are medical records relating to the time of the alleged offence.

There may be blood glucose measurements made soon after the incident. In a case of causing death by careless driving ( HM Advocate v Aitken (2012)), the defence was put on the basis that the accused had started a very low calorie diet the previous day. However, not only was there evidence that hypoglycaemia does not occur within 2 days of starting a low calorie diet, or even starvation, but the accused had a normal blood glucose when subsequently admitted to hospital.

Investigations can also include study of witness statements, particularly eyewitnesses, CCTV recordings, police custody and interview records (especially where arrest has taken place soon after the material events), the police accident report and any independent accident investigation report.

Look for eyewitness accounts of the person’s conversation and behaviour. A defence of automatism will be hard to advance if the person speaks and behaves as would be expected in someone who is not in an abnormal mental state.

In the 20–30 min following a concussive head injury there may be thrashing or flailing of limbs in a non-directed manner, uncooperativeness, wild swinging of the limbs in response to confrontation or attempted restraint, and aimless wandering (Reference McCroryMcCrory 2001). The accused may refuse to get on a stretcher or try to get out of an ambulance.

Eyewitnesses may report that the accused said that they were hungry, appeared anxious or pallid, were sweating or tremulous. They may describe the accused as looking and acting as though intoxicated, appearing bewildered, seeming to be in shock, acting like a robot or automaton, wandering aimlessly or lacking appropriate facial expressions. They may use terms such as ‘paranoid’, ‘fidgety’ and ‘out of it’ or say that the accused’s eyes were not working normally. Or they may describe the accused responding appropriately to what they see or hear.

In the case of sudden arousal disorders or confusional arousal, only the evidence of eyewitnesses may establish the four diagnostic features:

• asleep long enough to reach deep sleep

• a stimulus sufficient to induce waking

• onset of abnormal behaviour immediately on waking

• the episode lasting for just a few minutes, although it can be longer in sleepwalkers and if there is alcohol intoxication.

There may be evidence supporting a post-traumatic amnesia. A statement from a witness may include details of a conversation with the person, whereas evidence from later witnesses indicates that the person has no recollection of the matters discussed.

Pay attention to the reported behaviour of the accused after the alleged offence. Is it that of someone with little or no recollection of what has just happened?

The medicolegal report

Where a non-psychiatric condition has been identified as a potential cause of automatism, the defendant’s solicitors will usually have obtained a report from an appropriate specialist. They will then seek a report from a section 12 approved doctor. Although they ride ‘piggy-back’ on the other medical practitioner, the psychiatrist should not just ‘rubber stamp’ the first medical practitioner’s opinion. They should look carefully at their opinion and how it has been reached.

Readers are referred elsewhere for a more detailed account of the law relating to, and the investigation by the appropriate specialists of, non-psychiatric disorders capable of founding a defence of automatism (Reference Rix, Gall and Payne-JamesRix 2015). For example, in order to consider hypoglycaemic unawareness, investigations should include continuous glucose monitoring to examine the relationship between hypoglycaemia and reported symptoms. There was evidence in R v Clarke (unreported) (see Reference Rumbold and WasikRumbold 2011) that Mr Clarke had experienced 26 episodes of biochemical hypoglycaemia but was aware of only two.

In the case of a possible sleep disorder, the role for most psychiatrists will be to provide a report that is sufficiently clear and detailed for the instructing solicitors to make out a case for a second report from a sleep disorder expert.

If the first expert appears to have overlooked evidence that calls their opinion into question, this should be raised in a preliminary draft report. In any event, the psychiatrist’s report should include a statement to the effect that, insofar as there is evidence that falls outside their psychiatric expertise, they defer to the appropriate non-psychiatric experts.

Where the defence proceeds, the prosecution may obtain its own psychiatric evidence. The prosecution psychiatrist’s role is no different from that of the psychiatrist instructed by the defence.