Alcohol-induced sleepwalking or confusional arousal as a defense to criminal behavior: a review of scientific evidence, methods and forensic considerations



    1. Sleep Medicine Services, The Lankenau Hospital, Wynnewood, PA; Paoli Hospital, Paoli, PA; Department of Medicine, Jefferson Medical College Philadelphia, PA
    Search for more papers by this author

    1. Department of Neurology, Minnesota Regional Sleep Disorders Center, Hennepin County Medical, University of Minnesota Medical School, Minneapolis, MN
    Search for more papers by this author

    1. Department of Psychiatry, Minnesota Regional Sleep Disorders Center, Hennepin Count Medical Center and University of Minnesota Medical School, Minneapolis, MN, USA
    Search for more papers by this author

    1. Department of Neurology, Minnesota Regional Sleep Disorders Center, Hennepin County Medical, University of Minnesota Medical School, Minneapolis, MN
    Search for more papers by this author

Mark R. Pressman, PhD, Director, Sleep Medicine Services, The Lankenau Hospital, 100 Lancaster Ave., Wynnewood, PA 19096, USA. Tel.: +1 610 645 8273; fax: +1 610 645 2291; e-mail:


An increasing number of criminal cases have claimed the defendant to be in a state of sleepwalking or related disorders induced by high quantities of alcohol. Sleepwalkers who commit violent acts, sexual assaults and other criminal acts are thought to be in a state of automatism, lacking conscious awareness and criminal intent. They may be acquitted in criminal trials. On the other hand, criminal acts performed as the result of voluntary alcohol intoxication alone cannot be used as a complete defense. The alcohol-induced sleepwalking criminal defense is most often based on past clinical or legal reports that ingestion of alcohol directly ‘triggers’ sleepwalking or increased the risk of sleepwalking by increasing the quantity of slow wave sleep (SWS). A review of the sleep medicine literature found no sleep laboratory studies of the effects of alcohol on the sleep of clinically diagnosed sleepwalkers. However, 19 sleep laboratory studies of the effects of alcohol on the sleep of healthy non-drinkers or social drinkers were identified with none reporting a change in SWS as a percentage of total sleep time. However, in six of 19 studies, a modest but statistically significant increase in SWS was found in the first 2–4 h. Among studies of sleep in alcohol abusers and abstinent abusers, the quantity and percentage of SWS was most often reduced and sometimes absent. Claims that direct alcohol provocation tests can assist in the forensic assessment of these cases found no support of any kind in the medical literature with not a single report of testing in normative or patient groups and no reports of validation testing of any sort. There is no direct experimental evidence that alcohol predisposes or triggers sleepwalking or related disorders. A legal defense of sleepwalking resulting from voluntarily ingested alcohol should be consistent with the current state of art sleep science and meet generally accepted requirements for the diagnosis of sleepwalking and other parasomnias.


It is a generally accepted legal principle that criminal responsibility requires criminal intent and conscious awareness (Denno, 2003). The requisite intent and awareness are thought to be absent in states of automatism. States of automatism that fulfill these requirements occur during or a result of epilepsy, brain trauma, emotional trauma, hypoglycemia and parasomnias, such as sleepwalking among others (Eichelberger, 2000). A successful automatism defense based on the proven existence of one of these disorders at the time of the criminal act may result in acquittal of the criminal charge.

The medico-legal literature contains numerous reports of criminal cases in which the defendants allege to have been sleepwalking or in similar states (parasomnias) during murders, attempted murders, assaults and rapes (Arizona v. Falater, 1999; Bonkalo, 1974; R v. Parks, 1985; Broughton et al., 1994; Ohayon, 2000; Ohayon et al., 1997). More recently, defendants have claimed that alcohol intoxication, often at very high blood alcohol (BAC) levels, was responsible for the occurrence of sleepwalking violence, sexual molestation or other criminal acts (Ebrahim, 2006; Ebrahim et al., 2005; Finegan v Heywood, 2000; R v Sadler, 2005; R v. Catling, 2005; R v. Bilton, 2005; R v. Lowe, 2005; R. v. Balenko, 2000; R. v. Seymour, 2006; R. v. Wade, 1992; Schenck and Mahowald, 1998). Defenses of alcohol-induced sleepwalking have depended on classic scientific reports that alcohol causes an increase in slow wave sleep (SWS) and/or in the arousal threshold, both factors that could be associated with incomplete arousals during sleep and sleepwalking (Broughton, 2000; Roehrs and Roth, 2001a). However, almost all of the typically cited sleep and alcohol literature is now more than 25 years old and is rarely cited in detail (MacLean and Cairns, 1982; Yules et al., 1966; Yules et al., 1967).

A review of the sleepwalking research literature shows that it is settled science that sleepwalking and related disorders such as confusional arousals occur due to incomplete arousals from deep sleep (Broughton, 2000). For this reason, they are called disorders of arousal (Broughton, 1968). Instability of the brain systems subserving SWS has also been implicated in the pathophysiology of these disorders (Guilleminault, 2006; Guilleminault et al., 2005a, 2005b, 2006). Their occurrence is not related to dreaming or psychological trauma (Broughton, 1968; Kales et al., 1966). While in the sleepwalking state, the sleepwalker lacks conscious awareness and suffers from severely impaired cognitive functions such as attention, memory, social interaction and planning. Nevertheless, the sleepwalker is capable of complex motor behaviors.

Sleepwalking and confusional arousal episodes do not occur spontaneously but are thought to result from a complex set of predisposing, priming and precipitating factors (Broughton, 2000; Pressman, 2007). A predisposition or susceptibility to development of disorders of arousal has a genetic basis and familial pattern (Hublin and Kaprio, 2003; Hublin et al., 1997; Hublin et al., 2001; Lecendreux et al., 2003). Sleepwalking and confusional arousals are primed by factors that increase the quantity and depth of SWS or increase the arousal threshold. Priming factors have been reported to include prior sleep deprivation, situational stress, medications, fever and alcohol. However, individual episodes of sleepwalking and related disorders generally occur infrequently, suggesting that a proximal trigger should also be present. These triggers have been reported to include sleep-disordered breathing, periodic leg movements in sleep, noise and touch (Guilleminault et al., 2003, 2005a; Pressman, 2007; Pressman et al., 1995). These triggers produce arousals in sleep. However, in susceptible individuals these arousals do not result in full wakefulness, but leave the individual in a state not quite wakefulness and not quite sleep.

With higher cognitive functions absent or impaired, social inhibitions may also be absent or impaired and primitive behaviors may appear unexpectedly. For this reason, eating, sexual behavior and violence may all occur during sleepwalking episodes (Schenck and Mahowald, 1994).

The triggering of sleepwalking or confusional arousal episodes in the wrong place or the wrong time has been reported to be associated with violence for hundreds of years and has occasionally resulted in criminal charges (Bonkalo, 1974; Ohayon, 2000). More recently, sexual behavior in sleep has been described as a variant of sleepwalking or confusional arousals and has also resulted in criminal charges in a number of cases (Guilleminault et al., 2002; Rosenfeld and Elhajjar, 1998; Shapiro et al., 2003).

The majority of sleepwalking defenses in the past have relied on evidence of severe prior sleep deprivation, situational stress, family and personal histories of sleepwalking, amnesia for the episode, inappropriate and out of character behavior along with other factors. More recently, several cases have put forth defenses of alcohol-induced sleepwalking in defendants who were severely intoxicated at the time of their criminal behavior. As noted above, alcohol is traditionally considered a factor that can influence the occurrence of sleepwalking or confusional arousals. Early studies and current studies citing those early studies report that alcohol is a ‘trigger’ or precipitating factor for confusional arousals and sleepwalking. Some defense experts have taken this classic finding a step further and introduced alcohol provocation or induction procedures in the sleep laboratory in an attempt to ‘recreate’ the circumstances of the night of the criminal act (Ebrahim, 2006; Ebrahim et al., 2005).

This review will focus on: (1) the evidence for and against alcohol as a factor that increases SWS; (2) the evidence for and against alcohol as a priming or predisposing factor for sleepwalking; (3) the ‘objective methods’ that have been used in an attempt to prove that defendants were in an alcohol-induced sleepwalking state when they committed their criminal acts; (4) differences in the behavior and cognitive function of those intoxicated with alcohol compared with those suffering from sleepwalking and related disorders; (5) epidemiological and other research regarding the relationship of alcohol intoxication and sleepwalking to physical and sexual violence; and (6) implications of legal findings of voluntary alcohol intoxication.


An Ovid and PubMed search was conducted for all English language articles in which alcohol, alcohol abuse, sleep, sleepwalking and/or parasomnias appeared as key words. Legal databases in North American and the United Kingdom were also searched for criminal cases involving the same key words.

Alcohol and sleep

Alcohol and sleepwalking or related disorders

A search of the medical literature failed to find any experimental studies of the direct effects of alcohol on the sleep of clinically diagnosed sleepwalkers. Instead, there is a scattered group of case reports most often involving severe intoxication and criminal behaviors (see Pressman, 2007, Table 2).

Table 2.   Sleep in alcoholics and abstinent alcoholics
Citation and details of protocolNo. and type of subjectDuration of abstinence and details of protocolSWS (min or %) of total sleep time during maintenance use of ETOHSWS (min or %) during or following abstinenceSWS (min or %) when ETOH readministeredAdditional data
Irwin et al. (2002)14 male alcoholics aged 39.1 ± 9.8 yearsAbstinent for at least 2 weeks before testing 2.1 ± 5.1%  
Drummond et al. (1998)29 male abstinent alcoholic patients aged 25–67 years and an average of 12 years of heavy drinking.All 29 patients were studied after 16 days and 19 weeks of abstinence. 9 subjects continued on and were studied after 14 months and 4 at 27 months.N/a16 days = 0.2 ± 1.4%
19 weeks = 0.6 ± 6.9%
14 mnths = 1.0 ± 7.9%
27 mnths = 4.5 ± 14.3%
Snyder and Karacan (1985)26 male alcoholics aged 25–55 years with a history of 5–28 years of heavy drinking25 daysN/aS3% = 4.77 ± 4
S4% = 6.51 ± 7.96
Williams and Rundell (1981)48 male abstinent alcoholics aged 23–63 years. With an average of 16 years. Heavy drinking.Longitudinal study for 5 of 48 subject abstinent for 21 monthsN/a3 weeks = 2.6 ± 2.5%
13 weeks = 5.2 ± 4.7%
9 months = 8.3 ± 8.4%
15 months = 11.4 ± 10.6%
21 months = 7.7 ± 8.2
N/a24 subjects participated in 9 month follow-up. 14 subjects had returned to drinking. Compared to abstinent alcoholics SWS% was 7.4 ±5.6 versus 6.5 ± 4.9%, ns. No. of brief arousals for Ss abstinent for 35 days 124.2 + 62 versus 36.1 + 22 (controls) P < 0.001. No. of brief arousals for abstinent alcoholics at 92 days = 124.1 + 55.
Zarcone et al. (1980)9 alcoholic men aged 28–58 years. With history of alcoholism of 3–32 yearsAll subjects abstained for alcohol for 30 days. 5 days of sleep recording while abstinent followed by one day with readministration of alcoholN/a32.1 ± 31.7 min69.9 ± 46.2 min
BAC 20 min before lights out was 166.1 ± 26.7 mg/100 mL
Authors noted ETOH only appears to increase SWS% in abstinent alcohols in whom baseline SWS% remains relatively normal . 1 subject had no SWS in baseline or alcohol studies
Lester et al. (1973)20 male alcoholics aged 24–55 years with histories of heavy drinking of 3–30 yearsAbstinent for at least 3 weeks. 1 week later returned to sleep laboratory for 2 days during which alcohol was administered. Mean BAC before bedtime on night 1 was 145 mg/100 mL. On night 2 it was 155/100 mL S3% = 4.2 ± 4.4
S4% = 1.1 ± 2.6
During first of recording:
S3% = 7.8 ± 7.6
S4% = 2.2 ± 4.9
145 mg/100 mL
S3% 13.0 ± 10.4
S4% 10.2 ± 13.5%
S3% 11.8 ± 10.5
S4% 8.0 ± 10.9
Increase in SWS was statistically significant for night 1 only.
Increases in SWS with ETOH occurred exclusively in pts. aged 24–39 years and no increase in older pt.
Arousals from sleep baseline 97.9 + 27.5 (abstinent alcoholics) versus 32.2 + 24.9 (controls) P < 0.001. During alcohol administration 77.0 + 23.6.
Wagman and Allen (1975), study 16 male alcoholics with mean age of 35.5 years. with an average history of 8 years. Heavy drinking7 days withdrawal followed by 5 days of 18 oz 95% ETOH followed by 26 oz 95% ETOH, 1 day 32 oz 95% ETOHN/aAfter initial 7 days abstinence = 3%
After 7 days of abstinence at end of protocol = 5%
18 oz. = 10%
26 oz. = 12%
32 0z. = 15%
Highest increase in SWS with ETOH occurred with subjects with highest SWS during abstinence.
Adamson and Burdick, 197310 male alcoholics abstinent for 1–2 years3 consecutive nights in sleep laboratory. Results compared to JohnsonN/aS3% = 9.6 ± 5.8
S4% = 1.6 ± 2.7
N/aStage changes 112 + 28.6
Allen et al., 19715 alcoholic men aged 23–39 years with a history of heavy drinking for 2–24 yearsOn admission all subjects were maintained on 30 mL of 95% ethanol every 2 h from 08:00 to 22:00 hours for 3–7 days. 2 weeks withdrawal with PSG0–2.6%
3 of 6 subjects had no SWS
3 of 6 subjects had no SWS
N/aSleep stage values required 6–8 days to return to normal except for SWS
Fewer years of drinking associated with best SWS%
Johnson et al., 197014 males alcoholics aged 33–45 years. With an average history of 17 years. Heavy drinking2 days of ETOH use. BAC at 20:00 hours = 0.072 mg/100 mL followed by 10 days of withdrawalS3% = 3.0 ± 3.9
S4% = 0.2 ± 0.4
8 subjects had no S3 on night 2. 13 subjects had no S4 on night 2.
S3% = 5.0 ± 4.8
S4% = 0.8 ± 1.45
On day 10 of withdrawal. 3 subjects had no S3 and 9 subjects no S4

Alcohol and sleep in normal controls and social drinkers

Alcohol has been theorized to increase the possibility of sleepwalking episodes by increasing the quantity of SWS and/or by increasing the threshold for arousals (Broughton, 2000; Roehrs and Roth, 2001a, 2001b).

A review of the sleep medicine literature produced 19 sleep laboratory-based research reports in which alcohol was ingested before bedtime in subjects who were either nondrinkers or social drinkers (see Table 1). Compared with sleep without alcohol, no significant changes in SWS as a percentage of total sleep time or total recording time were noted. However, in six of the 19 studies, a statistically significant increase in SWS or stage 4 sleep only was noted in first 2–4 h of the sleep period. In the seventh study, a statistically significant decrease in SWS was noted with alcohol. In the remaining 12 studies, alcohol had no effect on SWS in the first half of the night.

Table 1.   Effects of alcohol on sleep of non-drinkers, social drinkers and patients with sleep apnea
CitationNo. and type of subject and details of protocolAlcohol level in blood or breath or quantity of ETOH administeredInterval between ETOH administration and sleep onsetSWS (min or %) 1st half of recording baselineSWS (min or %) of total sleep time baselineSWS (min or %) 1st half of recording ETOHSWS (min or %) of total sleep time ETOHLatency and no. of arousals to SWS: baseline and ETOH
  1. N, night of treatment with alcohol; AHI, apnea–hypopnea index; CPAP, continuous positive airway pressure.

(A) Alcohol and sleep
Feige et al. (2006)5 men mean age 29.8 ± 6.1 years and 5 females mean age 29.6 ±  9.3 years0.03%‘Before bedtime’22.58 ± 11.39%13.41 ± 7.01%N1 = 21.01 + 10.4%
N2 = 20.78 + 8.95%
N3 = 19.25 + 10.16%
N1 = 12.72 + 7.17%
N2 = 12.73 + 5.57%
N3 = 11.38 + 6.48%
0.10% 19.95 ± 11.98%12.03 + 7.89%N1 = 30.98 + 12.1%*
N2 = 26.92 + 8.78%**
N3 = 25.72 + 12.16%***
(*P < 0.0001, **P < 0.02, ***P < 0.006)
N1 = 16.49 ± 6.7%*
N2 = 14.47 + 5.0%
N3 = 14.20 + 7.47%
Van Reen et al. (2006)7 female aged 22–25 years0.043 g per % BrAC90 min1st 2 h = 56  ± 8 min (46 ± 6 min Stage 4 sleep)101 ± 301st 2 h = 65 ± 17 min (58 ± 16 min Stage 4 sleep)
(P < 0.03 for stage 4 only compared to placebo
106 ± 1714 ± 3 versus 11 ± 3 (ns)
Miyata et al. (2004): group 16 females aged 21.2 ± 0.8 years sensitive to ETOHB1 0.0 kg−130 min25.1 ± 13.8%16.8 ± 8.8%  N/a
N1 0.28 g kg−1   23.6 ± 20.0%15.7 ± 13.4% 
N2 0.69 g kg−1   25.7 ± 13.9 %14.2 ± 9.8% 
Miyata et al. (2004): group 27 females aged 21.1 ± 0.7 not sensitive to ETOHB1 0.0 kg−130 min9.3 ± 8.8 %12.1 ± 5.3%  N/a
N1 0.28 g kg−1   10.8 ± 5.1%9.3 ± 8.8% 
N2 0.69 g kg−1   12.8 ± 7.2%6.6 ± 3.5% 
Smith and Smith (2003)13 female, 2 male college students aged 19–24 years. As part of a ETOH Memory study0.7 g kg−130 minNot reported66.4 ± 13.6 minNot reported53.2 ± 13.6 min

Kobayashi et al. (2002)10 male subjects mean age 22.5  ±  0.6Administered 0.8 g alcohol/bodyweight‘before bedtime’N/a7.1 ± 5.5%N/a6.2 ± 7.0% (ns)Arousals = 6.4 ± 6.3 baseline and 9.3 ± 5.5 with alcohol
Roehrs et al. (1999)Normal controls, 3 females, 6 males aged 26.1 ± 3.7 years 5.8 drinks per week0.04 ± 0.02%60 min44.3 ± 28.8 min53.0 ± 36.545.3 ± 31.153.4 ± 39.6 
Landolt and Gillin (2001)10 male subjects aged 61.6 ± 0.9  years. Visual scoring0.55 g kg−16 h 37.3 ± 6.1 min 36.6 ± 6.4 min
Landolt and Gillin (2001)10 male subjects aged 61.6 ± 0.9 years. Frequency analysis of SWS activity0.55 g kg−16 h   0.75–45 hz activity increased in 1st and 4th SWS periods and decreased in second period compared to control conditionN/a
Lobo and Tufik (1997)12 males aged 25–30  years. ETOH ±  partial SD (4 h) ± REMD138.3 mg dL−130 min87.1 ± 6.9 min28.9 ± 7.4%82.8 ± 12.1 min (ns)28.6 ± 7.1% (ns)11.6 ± 4.8 versus 12.9 ±  ± 4.0 (ns
Lobo and Tufik (1997)12 males aged 25–30  years. ETOH ± TSD143.3 mg dL−130 min86.2 ± 8.0 min28.2 ± 9.3%96.2 ± 5.8 (ns)26.1 ± 6.89.7 ± 3.8 versus 6.7 ± 3.4 (ns)
Dijk et al. (1992)8 male mean age 24.3  years. Visually scored0.61 ± 0.10 (BEC)35 min1st 2 h = 29.0 m89.8 ± 7.41st 2 h = 40 min
P < 0.05 Stage 4 only
89.1 ± 11.9N/a
Dijk et al. (1992)8 male mean age 24.3  years. Frequency analysis0.61 ± 0.10 (BEC)35 min  1st 2 h = significant increase in 0.25–1 hz band (P < 0.05) N/a
Roehrs et al. (1991)5 males aged 25–6 years0.06 BEC30 min16.2 ± 10.8%18.1 ± 5.2%17.7 ± 11.0%
(P < 0.03)
16.2 ± 10.8N/a
Rouhani et al. (1989)7 men and 7 women aged 30–50 years. During afternoon naps starting at 14:30 hours and running for 90 min9 to 29 mg/100 mL50 minN/a20.14 ± 10.4%N/a8.71 ± 9.41%*
(*P < 0.05)
No. of awakenings 3.42 ± 1.95 baseline to 9.64 ± 3.57 ETOH*
(P < 0.001)
Williams et al. (1983)11 females aged 18–21 years0.060 min59.5 ± 15.8 min18.4–5.2%  10.6 ± 2.4 versus 10.3 ± 3.0 versus 9.2 ± 3.0 min
54 ± 19/100 mL   68.4 ± 20.2 min18.7 ± 6.1% 
81 ± 11/100 mL   73.9 ± 17.5 min*
*(P < 0.01)
17.4 ± 3.9% 
MacLean and Cairns (1982)10 males aged 23.6  years. ETOH administered at 5 doses0 ± 1/100 mL47 min70.3 ± 19.6 min22.5 ± 5.0%  32.5 ± 23.5
8 ± 5/100 mL   72.6 ± 17.6 min24.4 ± 5.7%31.3 ± 17.5
30 ± 8/100 mL   69.8 ± 15.6 min23.9 ± 4.4%20.4 ± 15.2
49 ± 5/100 mL   77.6 ± 12.8 min22.0 ± 3.425.8 ± 22.1
74 ± 8/100 mL   82.2 ± 18.6 min (P < 0.05)24.4 ± 4.222.8 ± 13.7 (ns)
Prinz et al. (1980)5 males aged 21–25 years. Acute administration of ETOH80.5 ± 5.02 mg/100 mL<30 min30.2 ± 2.33 minNo significant difference40.9 ± 2.22 min
(P < 0.05)
No significant differenceN/a
Prinz et al. (1980)5 males aged 21–25  years. Chronic (After 7 nights of ETOH before LO)80.4 ± 3.05 mg/100 mL<30 min30.2 ± 2.33 minNot reported31.2 ± 0.67 min (ns)Not reportedN/a
Stone (1979)6 males aged 20–31 years. Administered alcohol at 3 levelsPlacebo30–60 minN/a88.7 minN/a 16.6 min
6.25 (2–11) mg%    78.3 min16.8 min
26.5 (10–40) mg%    70.1 min13.4 min
51.4 (32–71) mg%    79.8 min11.8 min (ns)
Rundell et al. (1972)10 male students aged 21–30  years. Single dose study75 mg (50–90 mg)30 min30.1 ± 7.5% 32.5 ± 7.8% (ns) 23.7 versus 15.4 min (P < 0.01) (from stage 2 sleep)
Rundell et al. (1972)7 male students aged 21–27  years. Repeated dose (1 baseline, 3 ETOH77 mg (50–90)30 min29.9 ± 4.8% N1 = 33.3 ± 14.2%
N2 = 37.9 ± 10.3%
N3 = 32.6 ± 10.5%
Williams and Salamy (1971)10 subjects Single dose0.5 mL lb−1?27.7% 31.0%
(no statistical tests reported)
Williams and Salamy (1971)6 subjects Repeated dose0.5 mL lb−1 28.6% N1 = 32.0%
N2 = 36.3%
N3 = 31.4%
Yules et al. (1967)4 male graduate students50–60%4 h  Stage 4 increased in 2 Ss and decreased in 2 Ss. No change in Stage 3  
Yules et al. (1966)3 male graduate students85–105%15 min19% (estimated from graphs) 18% (ns) N/a
(B) Alcohol and obstructive sleep apnea
Scanlan et al. (2000)21 male habitual snorers0.7 g-dL BAC90 min 13 ± 1%
AHI ± 7.1 ± 1.9
 13 ± 1%
AHI = 9.7 ± 2.1
Teschler et al. (1996)14 obese males with untreated OSA0.45 ± 0.1 mg mL−1 (without CPAP)60 min22 ± 6%
AHI = 45.6.2
16. ± 4%
44.1 ± 7.8
21 ± 6%
AHI = 41.0 ± 7.2
13 ± 3%
AHI 36 ± 4
0.47 ± 0.2 mg mL−1 (with CPAP) 39 ± 6%
AHI = 3.3 ± 1.2
32 ± 3%
AHI = 3.3 ± 1.2
51 ± 4%
AHI = 4.5 ± 1.7
50.6 ± 8%
AHI = 3.1 ± 1.1
Collop (1994)8 untreated pts with Dx of OSA1 mL kg−1Immediate-17.4 ± 2.3%
AHI = 9.6 ± 5.3
-16.3 ± 2.0
AHI ± 20.2 ± 16.0
Berry et al. (1991)10 obese males with dx of severe OSA (AHI > 40 h−1) Pt. Used CPAP. ETOH administered at 2 different doses63.7 ± 17.3 mg dL−160 min 13.6 ± 10.1% 9.7 ± 8.5% 
108.6 ± 20.6 mg dL−1  22.6 ± 11.3%
AHI = 3.6 ± 3.7 h−1
 23.7 ± 14.7%
AHI = 1.9 ± 2.7 h−1

An increased quantity or percentage of SWS in the first 2–4 h of sleep could be consistent with an increased risk of sleepwalking episodes as sleepwalking and related disorders occur most frequently in the first 2–4 h of sleep. However, although the increases in SWS were statistically significant, the reported amount of the increase in these studies was not clinically impressive. The largest increases noted in SWS were from 70.3 ± 19.6 to 82.2 ± 18.6 min (BAC of 74 ± 8/100 mL) and in stage 4 only sleep from 46 ± 6 to 58 ± 16 min (BrAC of 0.043/%). As individual patient data were not presented, the presence of an individual with a more extreme increase in SWS in response to alcohol cannot be excluded.

It is generally accepted that alcohol will affect the arousal threshold of sleeping patients because of its well-known CNS inhibitory effects. However, a review of the sleep and alcohol literature failed to find any studies that directly tested this proposition.

Effects of alcohol on social drinkers with sleep apnea and snoring

Obstructive sleep apnea (OSA) and snoring are among the most common sleep disorders affecting an estimated 24% of men and 9% of women (Pang and Terris, 2006). Alcohol is well known to exacerbate the severity of sleep apnea and snoring by increasing the arousal threshold and decreasing the muscle tone of the upper airway (Collop, 1994). When alcohol is administered to patients with OSA the frequency of sleep-disordered breathing increases along with severity of sleep fragmentation. As a result, SWS decreases or shows no change (see Table 2). When alcohol is administered to patients who are asymptomatic snorers, the snoring is converted into apneas and hypopneas. Thus, alcohol typically results in either no change or a reduction in SWS when ingested by patients with sleep-disordered breathing.

Sleep in active alcoholics, binge drinkers and heavy drinkers

The effect of alcohol on sleep differs significantly for patients who are heavy drinkers, alcoholics or binge drinkers. In the majority of studies reported (see Table 2), the percentage of SWS is very low and sometimes absent during active drinking. Sleep is also severely fragmented with numerous brief arousals.

Sleep in patients with acute alcoholic psychoses

A series of studies by Gross et al. is not included in Table 2 due to differences in the standards for the analysis of sleep as well as the severity of the alcohol abuse syndrome in their patients (Gross and Goodenough, 1968; Gross and Hastey, 1975; Gross et al., 1973). Their studies typically involved patients with the most severe alcoholic syndromes characterized by delirium and hallucinations. However, their research subjects were also withdrawn from alcohol for the longest periods before testing. Of special interest are reports of sleep studies in patients with very severe alcoholism associated with delirium and hallucinations as well as extremely high BACs. The total sleep time of one patient in their series consisted of 100% SWS. However, high-amplitude delta wave EEG identical to the delta waves of SWS was also noted while the patient was behaviorally awake, suggesting that the delta wave activity noted in severe alcoholism may be due to brain dysfunction (Gross and Goodenough, 1968).

Sleep in abstinent alcoholics

The pattern of low or absent SWS with frequent arousals is noted to persist for months or even years even when the alcoholic becomes abstinent (see Table 2). However, several studies have suggested that return of SWS during abstinence may occur and a return to alcohol may result in an increase in SWS. In some cases, the administration of alcohol in the abstinent alcoholic may result in a return to levels of SWS that do not differ from matched normal controls or may go even higher than normal for several days. However, SWS values thereafter slowly decline. The difference in response to alcohol appears to depend on how long the patients had been abstinent before receiving alcohol and what their baseline level of SWS was during drinking and abstinence from alcohol. Patients with extremely low SWS% during drinking and abstinence often did not show any increase in SWS%, while those who had maintained a higher percentage SWS did show an increase.

Effects of sleep deprivation on abstinent alcoholics

A single study has examined the effect of sleep deprivation on the deep sleep of abstinent alcoholics. Irwin et al. (2002) suggest that many of these alcohol abstinent individuals may have damage to the brain systems that regulates SWS. This theory was tested by sleep-depriving abstinent alcoholics. Partial and total sleep deprivation in normals will result in a compensatory rebound of SWS when normal sleep resumes. Irwin et al. partially sleep deprived 46 alcoholics following 2 weeks of abstinence. On the sleep deprivation night they were permitted to sleep from 03:00–06:30 hours only. The effect of this partial sleep deprivation was compared with groups of European-American and African-American controls. Results showed that especially in the European-American normal control group total SWS and stage 4% essentially doubled, while in the alcoholic groups no change was present in SWS or stage 4%. Abstinent alcoholics did not show this rebound SWS following sleep deprivation suggesting that the compensatory brain mechanism was no longer functioning. It is suggested that the low or absent level of SWS and the inability to compensate for sleep deprivation could be secondary to alcohol-induced cerebral atrophy, changes in the circadian pattern of melatonin or levels of cytokine activity.

Provocative testing

In the UK in recent years, a battery of three sleep tests has been introduced as a method for ‘recreating’ the circumstances thought to have been present when the hypothesized incident of alcohol-induced sleepwalking criminal behavior occurred (Ebrahim, 2006; Ebrahim et al., 2005; R v Matthew Sadler, 2005; R v. Catling, 2005; R v. James Bilton, 2005; R v. Lowe, 2003). The intention of these tests is to provoke an episode of sleep-related behavior in the sleep laboratory. This test battery consists of a night of baseline polysomnography followed by 24 h of sleep deprivation. The next night the patient returns for the alcohol provocation portion of the test battery. Based on an estimate of the patients BAC at the time of the alcohol-induced sleepwalking incident, the defendant is asked to ingest sufficient quantities of alcohol to raise his BAC to that level by bedtime along with any other drugs or medications taken. The patient is asked to essentially go on an alcohol binge in the sleep laboratory. Recording then commences and continues for the duration of the defendants usual sleep period.

The preparation for these studies is not clearly stated. In defendants who have been jailed pending their trials there is an assumption of sobriety sometimes for months prior to testing. These individuals who previously were heavy or binge drinkers would have gone through withdrawal from alcohol and currently have the status of an abstinent alcoholic. Defendant's who were not jailed prior to testing could have persisted in their usual drinking habits or not. This would leave their status prior to testing as undetermined.

Study 1

The first sleep study in this test battery is described as a standard diagnostic sleep test or polysomnogram (PSG). When clinically confirmed young sleepwalkers are studied in this manner and compared with normal controls they are reported to have increased number of arousals in SWS. However, although the number of arousals was found to be statistically higher than that found in normal controls, the mean number and range of arousals varies considerably between studies. For this reason, there is no clinically meaningful cut-off value that can be used. Additionally, this finding lacks specificity in that frequent arousals from SWS are reported in patients with common sleep disorders such as OSA or periodic leg movements in sleep (Pressman, 2004). For this reason, the number of arousals from SWS has been found to have no diagnostic or predictive value for the diagnosis of sleepwalking or related disorders (Brozen et al., 2003).

The frequency of arousals becomes even less useful when alcohol is involved. When alcohol is administered to non-drinkers or social drinkers no statistically or clinically significant changes in arousal frequency are reported (Rouhani et al., 1989). However, in chronic or binge drinkers as well as in abstinent alcohols after almost 2 years of sobriety, frequent arousals are noted. The number of arousals during chronic alcohol use or following abstinence for very long periods is reported to be 3–10 times higher than those noted in patients with sleepwalking.

Classic studies of sleepwalking have also reported the presence of so-called hypersynchronous delta waves (HSDW) as a marker of sleepwalking. However, recent studies have noted that this marker may occur infrequently in patients with clinically confirmed sleepwalking (Schenck et al., 1998) and may not even appear during sleep studies in which actual sleepwalking occurs (Kavey et al., 1990). Further, HSD is reported to occur commonly in patients with no history of sleepwalking or related disorders (Pilon et al., 2006; Pressman, 2004). Thus, HSD lacks both sensitivity and specificity as a diagnostic marker of sleepwalking.

Of course, the occurrence of an actual sleepwalking episode or confusional behavior during the baseline sleep study would be considered a positive finding at that time. However, sleepwalking episodes rarely occur in the sleep laboratory. If one did occur, a close examination of the polysomnographic record would be required to rule out the possibility of malingering by the defendant. The presence of a positive sleep study finding would not necessarily be of use to the defense. Most studies are conducted months and even years after the date of the crime. A positive finding would confirm that the defendant has certain characteristics of a sleepwalker at the time of testing and not that he was necessarily sleepwalking on the night of the crime.

Other studies have reported statistical differences in the quantity and distribution of SWS when analyzed by frequency or power analysis (Espa et al., 2000; Guilleminault et al., 2001). However, the authors of these studies have warned that this finding cannot be used to either prove or disprove that a certain individual is a sleepwalker (Gaudreau et al., 2000). Sleep studies may, however, have a value for ruling out other sleep disorders or identifying potential triggers (Mahowald et al., 2007; Pressman, 2007).

Study 2

The second part of the three-part test battery is a sleep study following 24 h of sleep deprivation. Sleep deprivation is a well-known priming or possibly precipitating factor for sleepwalking that usually occurs in conjunction with situational stress. The relationship of sleep deprivation to sleepwalking has recently been experimentally tested in the sleep laboratory. Four studies have sleep deprived clinically diagnosed sleepwalkers from 26 to 38 h and then studied them in the sleep laboratory (Guilleminault et al., 1998; Joncas et al., 2002; Mayer et al., 1998; Pilon et al., 2006). Three of the four studies reported an increase in complex behavior out of SWS. The fourth study, however, reported the opposite effect, a decrease in complex behaviors. The differences between the four studies may be due to the dose of sleep deprivation, subject selection or other factors. The authors of the fourth study suggest that sleep deprivation followed by subsequent increase in SWS caused a consolidation of SWS and a decrease in arousals. Other studies of sleep deprivation in sleepwalkers hypothesize that sleep deprivation leads to an increase in arousal threshold making sleepwalking and confusional arousal episodes more likely.

However, as of this date, problems that reflect on the tests reliability and validity remain to be sorted out. As a result, the use of sleep deprivation cannot be recommended as part of either a clinical or forensic test battery (Pressman, 2007).

The amount of increase in SWS following sleep deprivation sufficient to prime or trigger a sleepwalking episode is unknown. As noted above several recent experimental studies have sleep deprived known sleepwalkers for 24 to 38 h. However, only one of these studies reports the SWS quantity and percentage prior to and following sleep deprivation. After 38 h of sleep deprivation the quantity of SWS a statistically significant increase from 40.6 ± 14.8 to 71.7 ± 15.3 min (P < 0.0001) was reported. This was accompanied by an increase in complex behaviors (Joncas et al., 2002).

However, all of these sleep deprivation studies report that sleep deprivation either had no effect on the number of arousals or caused a decrease in their number. Thus, the number of SWS arousals following sleep deprivation has no clinical or forensic significance. Further, as the purpose of the sleep deprivation is to increase the amount of SWS it is worth repeating that abstinent alcoholics with a profound loss of SWS have been shown not to respond to sleep deprivation with an increase in SWS.

The final concern about use of sleep deprivation is that many of the defendants who are being tested only report ‘sleepwalking’ episodes following alcohol use and not following sleep deprivation. If a defendant has never had a sleep deprivation-related episode, then the use of sleep deprivation is inappropriate and unlikely to produce the desired results.

The final sleep study of test battery is the ‘alcohol provocation test’. The stated goal of this test is to recreate the circumstances that the defense hypothesizes resulted in the sleepwalking or confusional arousal episode and the subsequent criminal behavior. To this end, experts for the defense estimate what the defendant's BAC was on the night of the criminal act. They then calculate how many beers, lagers or other types of alcoholic beverages would need to be consumed prior to bedtime to raise the defendant's BAC to that level. In practice the defendant is brought to the sleep laboratory several hours before testing is to begin and is asked to consume the required amounts of alcohol. When the patient is done he gets into bed and a standard sleep study is then begun.

This ‘alcohol provocation’ protocol suffers from a large number of serious methodological problems:

  • 1This test was specifically created for legal purposes and has never been used in clinical settings.
  • 2This test lacks any data regarding its sensitivity, as there are no reports of the effects of alcohol at any dose on clinically diagnosed sleepwalkers.
  • 3This test lacks any data on its specificity, as it has never been tested on normal controls or patients with other types of sleep disorders.
  • 4There are no normative data on which to base clinical or forensic decisions.
  • 5The basic premise of alcohol testing that alcohol increases SWS and arousal threshold is supported by only six of 19 published research studies.
  • 6No episodes of sleepwalking have been reported during alcohol studies.
  • 7At the time of the criminal act most defendants were chronic alcohol users, but at the time of the sleep study they have been abstinent for months or years. The alcohol provocations thus cannot be a ‘recreation’.
  • 8Administering a large quantity of alcohol to someone who is no longer tolerant to large quantities of alcohol is potentially dangerous.
  • 9Other factors present at the time of the crime cannot be duplicated including:
    • a. sleep quantity and patterns for at least 1 week;
    • b.stress levels;
    • c.use of other legal or illegal drugs.
  • 10Trigger for episode unknown
  • 11Other people, including victims, not present in sleep laboratory.
  • 12Sleep laboratory is sound shielded.
  • 13Sleep laboratory is absolutely dark.
  • 14Numerous electrodes and sensors placed on patient may disrupt sleep.
  • 15Sleep laboratory practices can disrupt usual sleep patterns.
  • 16Abstinent alcoholics have very poor quantity and quality of sleep.
  • 17Abstinent alcoholics have frequent arousals in sleep.
  • 18Administration of alcohol may result in an increase in SWS in abstinent defendants whose SWS has not dropped to very low levels.
  • 19Administration of alcohol to individuals whose baseline level of SWS during abstinence has remained low may not show any increase in SWS.
  • 20Administration of alcohol is known to increase the severity of OSA and cause a change from simple snoring to sleep apnea. Sleep apnea is associated with frequent arousals from sleep. Frequent arousals from sleep may reduce SWS.
  • 21Presence of arousals or HSDWs is not diagnostic of sleepwalking.

Intoxicated behavior versus sleepwalking behavior

Alcohol induces changes in physiology and cognition at certain doses and BACs each time it is ingested (Greely and McDonald, 1989). However, sleepwalking and related disorders occur intermittently and are dependent on a complex interaction of factors. The effects of alcohol are well known to be dose dependent and are influenced by a number of well known factors, such as tolerance, food, alcohol quantity and concentration, water level, age and body weight. Sleepwalking is known to be influenced by factors that may predispose, prime and precipitate episodes. However, the number of sleepwalking episodes that occur each year varies widely.

Epidemiology of alcohol and violence versus sleepwalking and violence

More than 100 cases of sleepwalking or confusional arousal-related violence or sexual behavior have been reported since 1901 (Bonkalo, 1974; Guilleminault et al., 2002; Ohayon, 2000; Rosenfeld and Elhajjar, 1998). In comparison, alcohol has been reported to be highly correlated with violence crime in a large number of studies for an average of 50% of all reported cases of homicide and assault (Murdoch et al., 1990). A recent study reported that alcohol ingestion increases the risk of a violent crime by 13.9 times (Haggard-Grann et al., 2006). A frequency of 50% would amount to millions of cases of alcohol-related violence worldwide each year.

Alcohol and sexual assaults versus sleepwalking and sexual assaults

Reports to the police of sexual assaults have noted alcohol use in an average of 50% of perpetrators with individual reports ranging from 32% to 72% (Abbey et al., 2001; Collins and Messerschmidt, 1993). Community-based studies of sexual assault have reported that 60–65% of perpetrators used alcohol (Abbey et al., 1994). Additionally, a large number of victims also report alcohol use before their attacks.

Sexual behavior in sleep is most often reported to occur during sleepwalking or confusional arousals (Guilleminault et al., 2002; Shapiro et al., 2003). Of 16 published reports of sexual attacks during apparent sleep, only two cases involve a sexual assault on someone who was not the usual sexual partner of the attacker or who was not sharing a bed with the assaulter. Of 16 well-documented cases of sexual assault during apparent sleepwalking episodes five were reported to involve alcohol use by the attacker.

Alcohol, NREM Parasomnias and the International Classification of Sleep Disorders – 2nd Edition (ICSD-2)

The ICSD-2 briefly refers to alcohol as a ‘precipitating factor’ for confusional arousals and as ‘an additional risk factor’ for sleepwalking (American Academy of Sleep Medicine, 2005). This is inconsistent as both disorders have identical pathophysiologies. These statements are also not based on direct evidence of the effects of alcohol on sleepwalkers, as no experimental studies of this have been done. As noted above, a review of the alcohol and sleep literature finds no support for alcohol as a proximal trigger for NREM parasomnias and little evidence that alcohol results in a significant increase in the quantity of SWS. The ICSD-2 does not distinguish alcohol's effects based on quantity ingested or BAC. The studies reviewed strongly suggest that heavy use of alcohol results in a significant decrease in SWS as well as increased sleep fragmentation. These conditions are not consistent with the occurrence of sleepwalking episodes. The findings reviewed here strongly suggest that in the next edition of ICSD, the role of alcohol should be limited to that of a possible predisposing factor when ingested in relatively low levels in individuals who are not alcohol abusers.

Voluntary alcohol intoxication and sleepwalking

Voluntary alcohol intoxication, especially with high BAC levels, may result in severely impaired cognitive functioning including amnesia for criminal behaviors (Marlowe et al., 1999). In most jurisdictions, voluntary alcohol intoxication is considered a reckless behavior and cannot be presented in court as a complete defense. It cannot be considered an automatism for legal purposes.

Thus, a defense of alcohol-induced sleepwalking or confusional arousal becomes a very attractive legal strategy with the potential of acquittal for the defendant. In legal cases utilizing an alcohol-induced sleepwalking defense, the role of voluntary ingestion of alcohol is greatly minimized in favor of the ‘real’ problem, sleepwalking or confusional arousal. Occurrence of the parasomnia and its accompanying behavior are presented as the major cause of an otherwise random and unpredictable act. The criminal act becomes the fault of the parasomnia and not the fault of the person who ingested the alcohol. Of course, if the defendant had not ingested alcohol on that night, no criminal behavior would have occurred whether it is directly due to intoxication or to an alcohol-induced parasomnia.

A Scottish case found that even if the court accepted that the defendant was in an alcohol-induced parasomnic state, the fact that the defendant had voluntarily ingested the alcohol made this fact moot (Finegan v Heywood, 2000). The defendant in this case had consumed at least six pints of beer in the late evening with friends, fell asleep watching TV in his living room and next remembers being awakened by the police in his friend's car after having traveled 1.5 miles and stopping at a 45° angle on the curb. He was found to have a BAC of 0.079% when tested by the police, exceeding the legal limit in Scotland. In his initial trial, the court accepted on the basis of expert testimony that the defendant had been in an alcohol-induced parasomnic state. Among the evidence presented by the expert witness was the fact that alcohol was associated with parasomnia because it was known to increase SWS sleep and SWS sleep was the time during which parasomnias were most likely to occur. However, the court rejected the defense because the parasomnia had been induced by voluntary ingestion of alcohol. Additionally, the defendant was aware that alcohol had been associated with three parasomnic episodes in the past. The Lord Justice-General writing for the High Court of Justiciary rejected the defendant's appeal. He noted that the proper approach would be to follow a previous case, Brennan v H.M Advocate (1977):

‘…a person who voluntarily and deliberately consumes known toxicants, including drink or drugs, of whatever quantity, for their intoxicating effects, whether these effects are fully foreseen or not, cannot rely on the resulting intoxication as the foundation of a special defense of insanity at the time nor indeed can he plead diminished responsibility’ (p. 46).

The court held specifically in this case:

‘that the defense of automatism cannot be established upon proof that the appellant was in a transitory state of parasomnia which was the result of, and indeed induced by , deliberate and self-induced intoxication.’ (p. 460).

The court noted that in the absence of alcohol the defendant might very well have been acquitted and that it was only because his drinking induced the condition that he was convicted. At the court's instruction this case was treated as a case of voluntary alcohol intoxication only.

An American case (Lewis v State, 1943) came to a similar conclusion that a defense of sleepwalking was not a legal defense to murder if it were ‘artificially’ induced by high quantities of alcohol. The court refused to overturn the initial guilty verdict.

Other cases involving claims of alcohol-induced sleepwalking have resulted in acquittals with different legal reasoning. Some of the acquittals have resulted from strong expert testimony for the defense and others have depended on whether judges thought the defendant's ingestion of alcohol was reckless or if the defendant's criminal behavior could have been predicted by past behavior.

In a Canadian case (R v Granger, 1996; Schenck and Mahowald, 1998), the defendant ingested at least 36 beers over a 10-h period. He was observed to be severely intoxicated during that time. At approximately 02:00 hours he got into bed at the home of a friend. Sometime later the 4-year-old daughter of his friend got into bed with him. The child later claimed he had sexually molested her. An expert witness presented evidence for an alcohol-induced parasomnia and the defendant was acquitted. The final verdict made no mention of the defendant's culpability if he should become severely intoxicated at a future date and commit a similar criminal act. However, Schenck and Mahowald have written that the defendant should not be able to make use of the alcohol-induced sleepwalking defense a second time. A second episode would constitute reckless behavior on his part.

The appeals court ruling in another Canadian case is also consistent with Schenck and Mahowald's point of view (R. v. Balenko, 2000). This case involved a man with childhood history of parasomnias. Parasomnias were also present in adulthood, but were most often associated with alcohol. In 1999, the defendant was found apparently asleep in his car with the engine running, gear in drive and foot on the brake. His BAC was 0.2%. He was unresponsive to the police and could not walk without assistance. He defended his actions by saying he was in a state of alcohol-induced sleepwalking.

During the appeal it was argued by the prosecution that disposition of this case should fall under a previous Canadian Supreme Court decision (Penno v. R, 1990) and treated as a case of simple intoxication. It was noted in the Penno decision that:

‘By voluntarily taking the first drink, an individual can reasonably be held to have assumed the risk that intoxication would make him or her do what he or she otherwise would not normally do with a clear mind’ (p. 366).

On appeal, the defendant's conviction was overturned with appeals court judge noting the defendant could not have known that his voluntary ingestion of alcohol would have resulted in behavior that was this complex even though he had a history of less complex behaviors with alcohol in the past. The clear implication was that this defense could not be used the second time, as the defendant was now aware that voluntary ingestion of alcohol could result criminal behavior.

The Court concluded that Mr Balenko:

‘neither did foresee nor could have reasonably foreseen that any state of somnambulism that might result from his consumption of alcohol could have possibly lead to complex behaviour and much less going out of his house and doing these acts of care and control of his vehicle. As a general rule, mens rea in true criminal offences at the very least requires recklessness.’

However, having said this, the court further states:

‘Were Mr Balenko not to take in the future the appropriate means to prevent the reoccurrence of this unfortunate event, he would most probably be hard pressed to plead that he did not act recklessly if the conduct in question did reoccur.’

Knowledge of prior similar sleep-related behavior under the influence of alcohol would thus appear to meet the standards in all jurisdictions as reckless and negate a claim that the defendant was suffering from an automatism. However, it may be very difficult to decide what prior alcohol-related behavior might be similar enough to invoke a finding of recklessness. In an attempt to prove their client is indeed a sleepwalker, defense attorneys often bring to the attention of the court as many prior episodes of putative ‘sleepwalking’ as possible. It is common for the court to hear testimony from numerous family members, friends and acquaintances that have witnessed prior apparent ‘sleepwalking’ episodes. This evidence may buttress the defense to some degree, but can also be used by the prosecution to demonstrate that the defendant had a prior knowledge of the effects of alcohol on his sleep. Whether this prior knowledge can be used to demonstrate recklessness in those courts that require it remains a question for the courts to answer.


Use of a defense of alcohol-induced sleepwalking or confusional arousal has been based on a number of poorly supported, but frequently repeated suppositions. Alcohol may be a priming factor for sleepwalking under very limited circumstances and at lower BAC levels only. At the high levels of alcohol intoxication often put forth as evidence for induction of sleepwalking/confusional arousals the hypothesized increase in SWS is extremely unlikely to be present and sleep architecture very likely to be disrupted.

The relationship between alcohol, SWS and sleepwalking that is often proposed as objective evidence is based on clinical and forensic case reports only. A review of the sleep and alcohol medical literature did not reveal a single report in which the effect of alcohol on the sleep of a previously clinically diagnosed sleepwalker was reported. There are no experimental data on dose effects. A variety of case reports – both medical and legal – describe what the authors believe is the induction of sleepwalking or confusional arousals due to large quantities of alcohol. In some cases, the individuals appear to have a limited past history of sleepwalking or confusional arousal in the absence of alcohol ingestion especially in childhood, but in the majority of cases episodes of ‘sleepwalking’ in adulthood occurs only with alcohol. Most often individuals in these case reports do not fulfill the usual criteria for the diagnosis of sleepwalking or confusional arousals, but do fulfill diagnostic criteria for alcohol intoxication-related disorders. Defendants in these case reports most often did not report sleep deprivation or situational stress. Defendants in these cases most often did not describe recent prior unambiguous episodes of sleepwalking that occurred without alcohol.

Consideration should also be given to the relative risks of violent crimes and sexual crimes following alcohol intoxication alone and during sleepwalking or confusional arousals. Sleepwalking and confusional arousals have been reported to be implicated in one or two cases of violent behavior a year worldwide over the last 100 years (Ohayon et al., 1997). On the other hand, alcohol has been reported to be involved in an average of 50% of all violent crimes and sexual assaults each year in a large number of studies conducted worldwide. In the USA, the Department of Justice (Greenfeld, 1998) estimates that three million alcohol-related violent crimes are committed each year. The 20005/2006 British Crime Survey (Walker et al., 2006) reports 1 029 000 alcohol-related violent crimes in England and Wales alone. A conservative estimate would be that alcohol alone is five million times more likely than sleepwalking or confusional arousals to be the cause of violent behavior.

Even if it is accepted that a true state of alcohol-induced parasomnia existed before commission of the criminal act, it would not have occurred without the voluntary ingestion of alcohol. In almost all legal jurisdictions behaviors that result from voluntary ingestion of high quantities of alcohol cannot be used as a complete defense. Criminal behavior that results from voluntary ingestion of alcohol cannot be defended in the same way as a spontaneously occurring abnormal state of the brain during sleep. Indeed, appeals court decisions in the USA and the UK have refused to accept defenses of alcohol-induced parasomnias because the alcohol was voluntarily ingested.

The clear conclusion is that alcohol is frequently and independently associated with violent criminal behavior while sleepwalking and confusional arousals are rare causes. Further, research studies of the effects of alcohol on sleep show a weak – at best – effect of low-level alcohol on SWS while no scientific evidence can be found to support a role for severe alcohol intoxication. Sleep laboratory tests cannot prove a defendant was sleepwalking at the time of a prior criminal act and alcohol provocation tests have no basis in current science or practice. This review demonstrates that views of the relationship between alcohol and sleep as well as alcohol and sleepwalking need to be revised to reflect the current medical research literature.

Finally, a legal defense of an automatism resulting from a voluntarily ingested alcohol-induced parasomnia should be consistent with the current state of sleep science and meet generally accepted requirements for the diagnosis of parasomnias. Claims of alcohol-induced parasomnias presented solely to circumvent the laws of voluntary intoxication should be understood for what they are and rejected.