It didn’t take long to find the man who did it: within a couple of days, her cousin Damon Thibodeaux confessed to police on tape that he had raped and murdered her.

In fact, Thibodeaux had nothing to do with the crime, as DNA evidence would later confirm. But he paid a heavy price for his false statement, spending 15 years in solitary confinement on death row before being released.

Although hard to fathom, false confessions happen surprisingly often; they are thought to play a role in up to a quarter of wrongful convictions in the US, according to the campaign group the Innocence Project. In many cases, as in Thibodeaux’s, the suspect was profoundly sleep deprived during their police interviews.

Now a study has shed more light on how easily severe exhaustion can lead to this type of false confession. Legal experts are predicting it will be cited in future court cases. “It’s a milestone,” says Lawrence Sherman, head of the Institute of Criminology at the University of Cambridge.

Questionable tactics

It might seem obvious that people who are dead tired can make stupid decisions – but there is a long history of police and army interrogators using sleep deprivation as an interview technique.

In the UK some notorious 1970s miscarriages of justice involving suspected IRA bombers hinged on false confessions made after profound sleep deprivation. It is now illegal for police in the UK to interview people who haven’t had 8 hours’ sleep in the past 24, unless in an emergency. The entire interview process must also be filmed.

But many other countries including the US have no such rules. Thibodeaux’s confession came after an all-night interrogation, and he had been up the previous night helping the family search for Champagne.

It’s not the only high-profile miscarriage of justice involving sleep deprivation. Amanda Knox falsely confessed to murdering Meredith Kercher in Italy, after being interrogated from 10pm to 6am.

Misguided gamble

“To the average person it’s inconceivable how a false confession can happen,” says Saul Kassin of the John Jay College of Criminal Justice in New York, who has been an expert witness in dozens of wrongful conviction cases. He says the suspect usually sees it as a short-term measure, thinking that when all the evidence is in, their innocence will become obvious. “They believe that in the end they won’t have to pay for the confession.”

Such a gamble is hard for juries to understand, he says, but the latest study might help. In this, 88 people did various computer tasks as part of a fake experiment, then either slept for 8 hours or had to stay awake all night. The next morning they were accused of losing all the study data by pressing the “Escape” key, something they had been repeatedly warned against.

“It’s not as awful as confessing to murder but some of these people feel really bad – they think the experiment is ruined,” says Elizabeth Loftus of the University of California, Irvine, who took part in the work.

When asked to sign a statement admitting their guilt, half of those who were sleep deprived complied, compared with only 18 per cent of those who got a night’s rest.

Shaky defence

Peter Neyroud of the University of Cambridge says the study would be more relevant if it had used prisoners instead of university students, because students could in theory be more suggestible. “They may be bright but they aren’t necessarily savvy.”

But it still fills a gap in the research, says Kassin; while other studies have shown sleep deprivation can impair people in various ways, this is the first to show it can lead to false admissions of guilt. “In court there’s nothing more persuasive than a study that goes right to the point.”

Kassin says that people typically retract false confessions as soon as they have been allowed to sleep – but it still derails the police investigation and, if the case goes to court, it undermines the defence.

For instance, witnesses to the suspect’s alibi, if they have any, may become less confident or even back out when they learn of the false confession. This was shown by another study published this month involving a fake theft. Subjects spent some time doing tasks with a partner and were then told some money had gone missing from a room next door.

When asked if their partner had ever left the room, at first 95 per cent of them said no; this dropped to 45 per cent when told their partner had confessed to the theft.

Kassin wants the US to adopt similar rules to the UK, with time limits and mandatory taping of interviews. Breaking the rules is always possible, he says, but once you impose limits, this would require more than just a single police officer, it would have to involve the entire department.