Last month, psychiatrists at Stanford University announced that sleepwalking is on the rise. More than 8.4 million adult Americans—3.6 percent of the population over 18—are prone to sleepwalking. That's up from a 2 percent prevalence the same authors found a decade ago.

And as the latest issue of Scientific American Mind notes, a subset of these nighttime wanderers may be at risk for a disturbing and dangerous phenomenon: sleep violence. Aggressive somnambulance in the general population hovers at or below 2 percent in surveys conducted in North America and Europe. But not all sleepwalkers exhibit violent behavior and what causes the violence remains a puzzle to researchers.

In fact, three separate disorders are associated with sleep violence. In arousal disorders—discussed in-depth in this month's feature—an individual operates in a mental state between wakefulness and sleep, carrying out complex behaviors with no evident conscious awareness. In comparison, people with nocturnal frontal lobe epilepsy experience brief, repetitive and inadvertently violent actions, such as running or kicking, that precede a seizure. A third condition, rapid eye movement (REM) sleep behavior disorder, occurs when movement centers in the brainstem—which create paralysis during deep sleep—deteriorate, often due to a disease of the nervous system such as Parkinson's. Without this paralysis, the body is free to move around and act out dreams during REM sleep, often causing accidental harm to the sleeper and bedmate. In 2000, the Mayo Sleep Disorder Center's Eric Olson reviewed the records of 93 patients with REM sleep behavior disorder and found that 64 percent had assaulted their spouses and 32 percent had injured themselves during sleep.

Because various disorders can underlie sleep violence, investigating incidents is understandably challenging. Michel Cramer Bornemann, a sleep specialist at the Minnesota Regional Sleep Disorders Center, and his colleagues at the center's Sleep Forensics Associates have handled more than 200 forensic cases related to sleep disorders, often at the request of law enforcement. Of these, only arousal disorders have been associated with criminal behavior during sleepwalking. He estimates that about a third of cases the forensics associates encounter involve sleep drugs, such as Ambien, which may increase the risk of experiencing an arousal disorder. In a state somewhere between wakefulness and sleep, these individuals may walk around, eat, or drive a car while asleep. Yet even though it may be possible to assess an individual's likelihood of having a sleep disorder, deciding whether that individual was awake or asleep during a given incident is another matter.

In 1997, Scott Falater of Arizona stabbed his wife repeatedly and pushed her into their swimming pool. When police—tipped off by a neighbor—arrived, Falater appeared unaware of what had happened to his wife. He claimed to have been asleep throughout the time of the incident.

In 2004, psychologist Rosalind Cartwright—who had been consulted by Falater's defense—wrote a report of the case, paralleling it to an earlier sleepwalking murder in Canada. In each case, the murderer had no apparent motive and was reputed to have a positive relationship with his victim. Both men claimed to recall no portion of the attack. Cartwright adds that the perpetrators were undergoing intense personal stress and sleep deprivation at the time of the attack, heightening their risk for a sleep disorder. Falater, a Mormon, had been taking caffeine pills for the first time in years. Cartwright observed that the addition of this stimulant to his daily routine may have further raised his risk of disrupted sleep. The trials had very different outcomes, however. While the case in Canada ended in an acquittal, the jurors were skeptical of Falater's sleepwalking story. Falater was convicted of murder and sentenced to life in prison.

As Cartwright noted in her report, there is no single test for diagnosing a sleep disorder with certainty. She had conducted a battery of psychological tests and four nights of sleep studies before testifying that a sleep disorder might be involved in the Falater killing. Nonetheless, it is virtually impossible—and ethically problematic—to reconstruct the circumstances of a given night or provoke a patient into sleepwalking or talking.

Cramer-Borneman adds that sleep violence presents important challenges to the legal system. Although the current system recognizes only 'mens rhea,' a guilty mind as a requisite of guilt to go along with a guilty act--perhaps the all-or-nothing understanding of the mind is inappropriate. Instead, sleep violence may be best explained in terms of degrees of conscious awareness, degrees of wakefulness, self-control, and sleep.

For now, the possibility of wandering like Lady Macbeth, with eyes open "but their sense is shut," remains a haunting reality, a hint of the many mysteries the sleeping brain still holds.