For years, Brian and Christine Thomas rarely went to bed together. They had decided to sleep in separate rooms because his chronic sleepwalking disrupted her rest. But they were still emotionally very close. "Each night, we'd have a kiss and cuddle first and then I'd go to my room, and the same in the morning," he said recently. "If she woke first, she'd come to my bed, and if I woke first, I'd go to hers. We never got up without being in each other's beds first."
Thomas, now 59, had been sleepwalking since he was a child. He would sometimes wake up with cut feet and stones in his bed. He once even swam in a nearby canal while asleep. His wife was so worried by these incidents that she took to locking the house at night and taking the keys to bed so Thomas could not sleepwalk out of the house.
His sleep disorder became so bad in recent years that he was prescribed antidepressants. He would come off the drugs, which he believed made him impotent, every two months so he could make love with his wife. "A side-effect of coming off them was hallucinations," he said. "Now, when I think about it, I realise that most of my problems came when I was off the drugs."
Last July, the couple went on an impromptu holiday to west Wales in their camper van, to cheer up Christine after a cancer scare. Brian had come off antidepressants a week before they went away, and they slept in the same bed.
On the last night of their holiday, they stopped the camper in a car park in the seaside village of Aberporth. They had dinner at a pub, saw a beautiful sunset and then went to bed. "I always slept with her back towards me and my right arm under her neck and my left arm over her," Thomas said.
Later in the night they were woken by what Thomas described as "boy racers" doing hand brake turns in the car park near where they were sleeping. "As I went to sleep, it must have been on my mind that the camper wasn't secure. Then – I don't know how much later it was – I recall seeing Chris in bed over the other side of the camper and someone on top of her. All I said was: 'You bastards, you got in here.' I grabbed this man round the neck and pulled him off."
The following morning Brian woke up next to Christine's strangled body. He rang 999 and said: "I think I've killed my wife. Oh my God. I thought someone had broken in, I must have been dreaming or something. What have I done?"
Last month Thomas, a retired steel worker, was acquitted at Swansea Crown court of strangling Christine, 57. Lawyers prosecuting Thomas originally argued for a verdict of "not guilty by way of insanity", but then dropped the case, and Thomas was found simply "not guilty".
Judge Nigel Davies concluded that Thomas was "a decent man and a devoted husband". And Thomas has since said, "I'll never forgive myself, ever. It's like a hatred – a hatred of myself. Why did I do it?" But the case raises uncomfortable questions. Can a sleep disorder qualify as insanity? Could other, guilty, defendants use a similar claim to get away with murder? And what turns a chronic sleepwalker into an unwitting killer?
Dr Chris Idzikowski has been a sleep specialist for 30 years. When I meet him at the Edinburgh Sleep Centre he is still trying to understand the complexities of the Thomas trial, at which he was an expert witness for the defence. "The court accepted that this was a case of automatism," he says. "Meaning you have a condition that can lead to automatic behaviour, be that epilepsy, hypoglycaemia or – as in Thomas's case – a sleep disorder. In such cases you can't be held responsible for your actions." However, the plea of automatism is a complicated and controversial one.
Idzikowski tells me he typically receives one inquiry a week from lawyers seeking to use automatism as a defence for clients charged with offences ranging from assault to rape and murder. He first met Thomas before the trial and did two sleep studies on him, first at the Birmingham sleep centre and then at Swansea Prison, to test his claim that it was a parasomnia, or sleep disorder, that had led him to kill his wife.
"The recordings show that he had a condition that was consistent with night terrors," says Idzikowski. "But there's more than just a history of sleepwalking. There is a problem of him not breathing properly in the night, which can be a trigger for sleep behaviours."
Thomas proved particularly difficult to diagnose. "Initially we thought it was a case of REM [rapid eye movement] disorder behaviour, brought on by withdrawing from his drugs." REM is the more shallow form of sleep, during which we dream, so that made sense, Idzikowksi says. "He seemed to be acting out his dream of fighting with an intruder, and acting out dreams is a classic REM sleep disorder." But Thomas's night terrors and sleepwalking are both parasomnias associated with deeper, dreamless non-REM sleep.
Non-REM sleep disorders happen when your cognitive functions are not engaged, but your behavioural patterns are. "They often start because of an abrupt arousal like a car backfiring, or being too hot or too cold," explains Professor Colin Espie of Glasgow University's Sleep Centre. "That trigger rockets them to wakefulness. But thinking, planning and memory formation activities are still asleep."
Idzikowski is still unsure which form of sleep disorder caused Thomas to kill his wife, but he supports the court's decision. "The prosecution withdrew their case, which was right." Is Thomas insane? "No: it's a legal definition, not a medical one. Legally, there's non-insane automatism and insane automatism. The former is used if you've had a blow to the head, or you withdraw from drugs, and that creates the condition. Insane auto–matism is when it's intrinsic to the person's behaviour. Thomas withdrew from his drugs, but he also had a history of sleep disorders. So you could argue either non-insane or insane automatism. But to call him medically insane is not right."
An estimated 10 million Britons have sleep problems. A small proportion of those suffer from parasomnias, which are a range of sleep disorders that involve abnormal movements and behaviours during sleep. Two per cent of us suffer from sleepwalking (most are children, but for some the condition persists into adulthood). In 2005, a teenage sleepwalker had to be rescued after being found asleep on the arm of a 130ft crane.
Andrew, 27, has suffered from sleep disorders for as long as he can remember. "A lot of people have parasomnias in childhood and grow out of it. I didn't. My parents say that when I was a child I would cry out in my sleep a lot, but they didn't think it was treatable.
"When I started living with other people it became a problem. I had a flatmate who kept a rosary above her bed because she thought the noise was being made in the night by some sort of howling ghost." One night, his flatmates found him in his attic bedroom screaming and trying to prop up the roof with his arms. "The next morning, I had no memory of what I'd done."
Night terrors must make long-term relationships difficult. "Absolutely. The worst thing was that I wouldn't remember what I'd done in the night, which made me feel terribly guilty. I would never be aggressive to people – it's more of an inward aggression and fear – but there was always a danger I would lash out in my sleep and hurt someone I cared about."
It was only last year that Andrew sought treatment. "My fiancée nagged me because she was worried she was going to get hurt." Andrew's GP referred him to St Thomas's Hospital in London where his sleep patterns were monitored overnight. "I remember my consultant showing me the results. He pointed to one part of the graph for what's called 'fourth stage non-REM sleep' and said to me, jokingly, 'This is the automaton stage... as it would be described in court.' That's where I'd get off with murder, because the things you do in your sleep at that point, you don't know you're doing."
The first case in which the defence of automatism was successfully argued is thought to be that of Albert Tirrell, who in 1846 was accused in Boston, Massachusetts, of murdering his lover, a prostitute called Maria Bickford. His lawyer, Rufus Choate, convinced the jury that Tirrell did not cut Bickford's throat – or, if he did so, did it while sleepwalking; under the "insanity of sleep".
Since then, the controversial defence has been used in murder trials more than 60 times. In 2005, for example, Jules Lowe, 32, was found not guilty of murdering his 83-year-old father Edward after claiming the attack at his home in Walkden, Greater Manchester, took place while he was sleepwalking. Edward suffered 90 separate injuries which Manchester Crown court heard were consistent with being punched, kicked and stamped on, but his son had no memory of the attack.
Dr Irshaad Ebrahim, director of the London Sleep Centre, was called in to carry out a series of overnight sleep studies on Lowe to test his claim, measuring brain waves, muscle activity and breathing activity. He also looked at factors that trigger sleepwalking episodes, such as alcohol and stress. "We think this was the first sleepwalking murder in the UK," he says. "Mr Lowe had a history of sleepwalking, and this was generally worse when he drank alcohol, but he had never been violent before the night of this offence. However, his stepmother had just died and there were several other stressful factors occurring in his life."
This was, however, an unusual case. "Extreme forms of violence, of sleepwalking or automatism, are extremely rare, so we usually view them with suspicion."
Rare, perhaps, but such cases often receive a great deal of publicity. One especially famous criminal case of automatism didn't involve a sleep disorder, but still made headlines. When REM guitarist Peter Buck attacked two cabin staff during a 2002 British Airways flight from Seattle to Heathrow, it was seen initially as a case of a boozed up rock star behaving badly.
Buck covered the cabin crew in yogurt, knocked over a trolley, tried to steal a knife and swore at the captain. But he was acquitted of common assault and damaging property. The court accepted he had no recollection of the incident because he was suffering from non-insane automatism at the time, brought on by combining alcohol and a sleeping pill at the start of the flight. After the verdict, pilots' union Balpa said the jury decision sent out "all the wrong signals".
Similar scepticism has inevitably greeted rape cases in which automatism has been used as a defence. In 2007, 26-year-old RAF serviceman Kenneth Ecott was found not guilty of raping a 15-year-old girl in Poole, Dorset. He told Bournemouth Crown court he had a condition called "sexsomnia". During a friend's birthday party, Ecott climbed into a bed with the girl and had sex with her. She awoke to find him naked on top of her. She screamed and then watched him get up slowly and go into the garden. Ecott told the court he had no memory of the rape. The girl's mother told reporters after Ecott's acquittal, "I worry that there are a lot more people who will get off lightly using the same defence. They could harm someone else and use sexomnia as an excuse."
How can someone unknowingly commit a violent act such as rape? "If sexual activity is a habitual behaviour, then it is possible," Professor Espie says. He points out that we carry out many tasks unconsciously every day. "It's only when you learn something new that you do it fully consciously. For example, when you're learning how to drive, you can't imagine that you'll ever be able to do it automatically. But eventually you do, with no memory of having got from A to B. It's similar with sleepwalkers. Often you'll find they carry out tasks – boiling a kettle, making toast – on automatic." Espie believes this automatic behaviour can stretch to making sexual advances. "We've had couples come in for sleep studies, with one partner complaining of unwanted sexual attention in the night from their sleeping partner."
Cases of sexsomnia, though, are unusual. Dr Adrian Williams of the London Sleep Centre says that most court cases involving defendants with sleep disorders are accidents involving sleepy drivers. In one case, in 2005, Colin Kane, 36, from Bishopton, Renfrewshire, was driving a truck that crashed into a tailback near Alexandria, Dunbartonshire. Three people died in the accident. "This driver managed to convince the court that he was not responsible for the crash because his sleep apnoea [disturbed breathing during sleep] meant he was not in control of his vehicle at the time," Dr Williams says.
Kane had reportedly experienced a blackout just before the crash, which may have been caused by his sleeping disorder. The jury found the case not proven. "So far, sleep apnoea hasn't been successfully used as a defence in such cases," Dr Williams says.
Most parasomniacs don't end up in court. Usually they seek help and, in most cases, their behaviour is brought under control. For the past year, Andrew has been taking varying doses of antidepressants, trimipramine and the muscle relaxant clonazepam. The side effects of the parasomnia have almost gone and he sleeps much better, "but I worry that I'm developing a tolerance to the drugs". For many parasomniacs, these drugs and their side-effects are a reality of life. In the Thomas case, his defence rested on the fact that he had come off his anti-depressants.
But should parasomniacs take more responsibility for their condition? "If, during sleepwalking for example, you regularly drive, shouldn't you or someone in your family take steps to hide the keys so you can't?" asks Professor Espie. "Say you killed a pedestrian by driving while asleep and you had a history of sleepwalking, shouldn't you deserve to be found culpable?"
So far, though, there have been no such convictions in Britain. Courts have, however, attempted to stop the circumstances that led to a sleep-disorder-related death recurring. "There was a case in Edinburgh in the 19th century in which a sleeping man thought his son was being attacked by wild animals," Dr Idzikowski says. "He thought he'd thrown him to safety, but in fact he threw him against the wall and killed him. The judge and jury accepted the defence of automatism, with the proviso that he could no longer sleep with anyone again."
Sleep-related automatism is controversial partly because there is a worry that even sleep experts can be hoodwinked by cunning defendants. "I have certainly seen people who I think are trying it on," Espie says. "It's very difficult to tell the difference between a genuine and a bogus case, and I've been in sleep research for 30 years. At one extreme you might have a lying psychopath who wants to get off, and at the other extreme an innocent who invariably admits to what they have done." How then do you tell them apart? "One sign is that the innocents are generally not defensive at all, because they can't remember it happened. They readily admit to what happened, and say it wasn't their fault."
"One judge said that automatism is the last refuge of the scoundrel," Dr Idzikowski says. "I'm sure there are people who have the disorder, commit a crime, and try to lean back on it to get away with it."
Could that have happened with Brian Thomas? "I'm convinced he was not guilty. That said, you never know. Maybe he's a genius who's tricked me and everybody else and is now going to claim lots of insurance money for his wife's death. Perhaps, but I don't think so."
