Not Guilty by Reason of Insanity of Murder: Clinical and Characteristics

Not Guilty by Reason of Insanity of Murder: Clinical and

Characteristics

Paul G. Nestor, PhD, and Joel Haycock, PhD

We examined archivally clinical status, neuropsychological functioning, and perpetrator-victim relationships of 28 adult patients who had committed homicide and had been subsequently involuntarily committed to a forensic hospital. We divided patients into two groups: (1) not guilty by reason of insanity (NGRI) acquittees (n = 13) and (2) convicted murderers (n = 15). In comparison with convicted murderers, NGRI acquittees were more likely to be seen as psychotic at the time of the index offense and also were more likely to have killed blood relatives, especially a parent. By contrast, convicted murderers were more likely to have killed a significant other, mainly a spouse or lover. At the time of the index offense, substance abuse was more likely to have occurred in the convicted murderers than in the NGRI acquittees. NGRI acquittees and convicted murderers did not differ on neuropsychological tests, with both groups generally scoring within normal limits on all tests. Taken together, these results suggested that NGRI murderers may be driven by acute psychosis directed toward blood relatives and occurring against a backdrop of relatively preserved neuropsychological functioning.

Violent criminal acts precipitated by ac- with socially. In the relatively rare in- tive mental illness occur rarely,'.2 but stance of a judicial finding of not guilty nonetheless generate considerable, often by reason of insanity (NGRI), a mentally heated philosophical, moral, and legal de- ill person is absolved of criminal wrong- bate about how such acts should be dealt doing for acts thought to be a direct con- sequence of a disordered mind. A person Dr. Nestor is affiliated with the Department of Psychol- ogy, University of Massachusetts, Boston; Department adjudicated as NGRI is typically commit- of Psychiatry, Harvard Medical School, Brockton Vet- ted to a state psychiatric facility until erans Affairs Medical Center, Brockton; and Bridgewa- ter State Hospital, Bridgewater, MA. Dr. Haycock is deemed no longer dangerous by reason of affiliated with the Department of Psychiatly, University mental illness. The distribution of crimes of Massachusetts Medical Center, Boston, and Bridge- water State Hospital, Bridgewater, MA. Preparation of varies among NGRI acquittees requiring this article was supported by an individual research grant continued hospitalization. In an eight- from the American Suicide Foundation to J.H. and awards from the Healey Endowment to P.G.N. and J.H. state study, Callahan et aL3 reported that Address correspondence to: Paul G. Nestor, PhD, Uni- 15 percent of hospitalized insanity acquit- versity of Massachusetts-Boston, Psychology Depart- ment, Boston, MA 02125-3393. tees committed homicide and another 38

J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 161 Nestor and Haycock percent had criminal charges of physical tween what may be "bad" and what may assault. More recently, Shah et aL4 re- be "mad" are lost.', * ported that 40 percent of NGRI acquittees The more serious the possible criminal in a regional New Jersey state hospital penalty, the greater the legal incentive for had been charged with murder or at- pursuing insanity acquittal. Since the tempted murder. Similarly, a study of mid-19th century until the last several New York State NGRI acquittees found decades, insanity acquittals often brought that 33 percent committed homicide or lifetime commitments to institutions for attempted homicide, 19 percent aggra- the criminally insane. Therefore, few vated assault, and 16 percent arson.5 crimes other than capital defenses gave a Perhaps no NGRI defense will inflame defendant an indisputable legal interest in popular and scholarly debate more than an insanity plea.9 Murder was preeminent that which occurs when a mentally ill among those offenses. Despite their po- person commits homicide. The debate is tential legal interest in them, few homi- as old as antiquity, capturing the interests cidal perpetrators receive insanity acquit- of moral philosophers no less than Aris- tals. The reasons may be multiple. totle who emphasized the importance of Mentally ill persons account for a very small number of total homicide^'^-'^; the distinguishing between "people of weak correlation between mental illness and vi- will who do wrong against their real olence appears modest2; and homicide ac- wishes and intentions, and vicious people count for a minute percentage of all of- who do wrong contentedly and with con- fenses committed by mentally disordered victionW6(p. 59). No act is arguably more persons. Further, since in recent decades vicious, tragic, and evil than that of ho- release to the community may or will micide. A NGRI finding in a case of eventually follow an insanity acquittal, murder ultimately distinguishes between judges and juries may view a judgment of those whose actions social policy deems a insanity with greater reluctance for the product of a "weak will," and hence not most serious offenses. Thus, Janofsky et morally blameworthy, from those who a1.13 found that the seriousness of the acted primarily with malice or evil.' Be- charge, together with diagnosis and ex- cause of the controversy that surrounds pert opinion, differentiated those eventu- such cases and because of the legal stat- ally found NGRI. Their sample of 143 utes involved, a NGRI defense for the defendants who pled insanity found none charge of murder may therefore represent charged with rape or murder who even- both the best and worst examples of clin- tually received such a verdict (NGRI). ical science providing information for the There are no national data on insanity courts in matters of culpability and acquittals. The best existing research has blameworthiness. In the best scenario, disclosed considerable variation in the boundaries between moral and clinical rate of insanity acquittals. 14* l5 Perhaps judgments are rigorously upheld, and in because of the small numbers involved, the worst scenarios, the distinctions be- the reported variations in homicide insan-

162 J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 NGRI of Murder ity acquittals have been considerable. In a including substance abuse, history of study of all Massachusetts insanity ac- mental illness, and intelligence. Among quittals during the years 1978, 1980, and the many findings was that, in compari- 1982, Phillips and ~ornikl"found only son with the two contrast groups, NGRI eight subjects, representing just 4.8 per- murders occurred during the acute phases cent of all insanity acquittees. had been of schizophrenia and were less likely to acquitted of murder. As of this study, be related to substance abuse at the time Massachusetts ranked lowest among all of the crime. those states in which the incidence of We have previously examined homi- NGRI findings in murder cases had been cides in relation to age, psychotic symp- studied, and in the intervening years, the toms, victim-perpetrator relationships, number of insanity acquittees has contin- and neuropsychological characteristics, ued to be very small. but not in relation to legal defense and Because of the relative infrequency of out~ome.'~'20 These findings indicated successful insanity defenses to murder, that, in comparison with less violent psy- there has been limited systematic empir- chotic patients, psychotic murderers ical research directed toward distinguish- scored higher on a variety of neuropsy- ing characteristics of NGRI murderers. In chological tests and were more likely to a comparison study of 50 homicide de- have imposter delusions and delusions in- fendants acquitted by reason of insanity corporating specific personal targets. Psy- and 50 defendants evaluated for criminal chotic homicides also involved a remark- responsibility and deemed responsible, ably high incidence of blood relatives as packer17 found that female sex, white victims. Yet, to our knowledge, no study race, fewer felony convictions, at least has yet to examine these variables simul- one prior hospitalization, a diagnosis of taneously, namely clinical characteristics, psychosis, and killing of a parent or child neuropsychological functioning, and per- were all significantly over-represented in petrator-victim relationships in NGRI the NGRI group. while spousal or murderers and in psychiatrically involved stranger victims were under-represented (but not necessarily psychotic), convicted in that group. The latter finding appears murderers. We therefore examine these to contrast with more recent reports of variables in two groups of patients com- Steadman and colleagues l5 who found mitted to Bridgewater State Hospital in that in relation to all types of assault, type Bridgewater, Massachusetts, the state's of victim had little relation to the success only secure psychiatric facility. For a of an insanity plea. In a recent study NGRI finding in Massachusetts, one of a performed in Quebec, Beaudoin, et al.I8 minority of states that did not revise the compared 14 schizophrenic NGRI mur- conditions of its insanity defense in derers with 12 schizophrenic patients the aftermath of the Hinckley verdict.15 convicted of murder and 15 convicted the burden of proof remains on the Com- murderers with no diagnosed mental dis- monwealth, and the standard of proof is order on a variety of clinical variables, beyond a reasonable doubt. In Common-

J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 163 Nestor and Haycock wealth v. McHoul, 226 NE 2d 556 (1967), From the same sample of 310 patients Massachusetts adopted the American referred for neuropsychological evalua- Law Institute (ALI) Model Penal Code2' tions between 1987 and 1995, we selected definition of insanity: "An individual is a contrast group of 15 patients, all of not responsible for criminal conduct if at whom were eventually convicted of mur- the time of such conduct as a result of der, had IQ scores above 70, and were mental disease or defect he lacks substan- similar in age to the NGRI murderers at tial capacity either to appreciate the crim- the time of the index offense. For contrast inality (wrongfulness) of his conduct or to subjects, we selected only those patients conform his conduct to the requirements who had been referred for neuropsycho- of law" (p. 557).21 logical testing, had been either charged with or convicted of murder, and had Method undergone complete neuropsychological Sample We used an archival de- testing. Thus, the contrast subjects did not ~ign'~.~'to examine hospital records of represent a random sample of persons approximately 3 10 cases, all males conl- referred for neuropsychological testing, mitted to Bridgewater State Hospital in but rather were selected on the basis of Massachusetts. the state's only secure fo- age, crime, and completion of a test bat- rensic psychiatric facility, and referred tery. for neuropsychological evaluations be- Rating Scales To minimize subjec- tween 1987 and 1995. Records included tivity in ratings, we used a simple dichot- forensic psychiatric evaluations, neuro- omy (yedno) to rate records on the basis psychological evaluations and test scores, of presence of a (1) psychotic disorder at and police reports where available. From the time of the murder, as defined by a the sample of 3 10 referrals, 17 patients DSM-111-R diagnosis of schizophrenia, had been adjudicated as NGRI on the delusional disorder, psychotic disorder charge of murder. Four of the 17 patients not otherwise specified, major depressive were not included; 3 subjects completed disorder with psychotic features, or a bi- only one or two of the neuropsychologi- polar disorder with psychotic features:22 cal tests, and 1 subject scored in the men- and (2) substance abuse at the time of the tally retarded range on formal intelligence murder, as defined by a documented testing. We therefore examined the DSM-111-R diagnosis of psychoactive records of 13 patients adjudicated as substance abuse or dependence.22 or NGRI on the charge of murder, ranging records describing the perpetrator as in- between the ages of 20 and 52 years old at toxicated and/or under the influence of the time of the offense, with IQs above illicit drugs. In addition, we rated perpe- 70, and who had been formally evaluated trator-victim relationships in terms of on neuropsychological tests. Eleven of stranger, significant other (spouse or lov- the 13 subjects had been acquitted NGRI er), and blood relative, including specific of murder between 1985 and 1990 and 2 blood relationship (child or parent) and subjects in the 1970s. unrelated child.

J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 NGRI of Murder

Neuropsychological Measures Test Table 1 data included standardized measures of Characteristics of Perpetrator-Victim Relationships (1) intelligence, as assessed by the Wech- sler Adult Intelligence Scale-Revised NGRI Convicted (n=13) (n=15) (WAIS-R), which provides intelligence No. % No. % quotients (i.e., Full Scale IQ; Verbal IQ; and Performance IQ) derived from 11 Perpetrator Psychotic 12 92 1 7** subscale scores;23 (2) memory, as as- Substance abuse 0 0 4 27* sessed by the Wechsler Memory Scale- Revised (WMS-R),~~which provides Victim memory quotients or indexes (i.e., Gen- Parents 8 62 0 0** Significant other 3 23 9 60* eral Memory Index, Verbal Memory In- Related children 215 1 7 dex. Visual Memory Index) derived from Unrelated children 0 0 2 13 7 subscales scores; (3) attention and con- Stranger 18213 centration, as assessed by the Attention *p < .05. **p < .01. and Concentration Index of the WMS-R derived from 5 subscales scores24 and by the Trail Making ~est;~"4) problem chusetts legal standard, in relation to con- solving and executive functions, as as- victed murderers, NGRI murderers were sessed by the Wisconsin Card Sorting significantly more likely to be described Test (WCST);*~and (5) academic abili- as psychotic around the time of the crime ties of reading. spelling, and arithmetic as (2= 20.53, df = 1, p < .0001). For assessed by the Wide Range Achieve- NGRI murderers, records described all ment Test-Revised (WRAT-R).*~ but one subject (92%) as psychotic at the time of the crime. By contrast, records Results described only 1 of the 15 convicted mur- Both groups did not differ significantly derers (7%) as psychotic at the time of the in age at the time of the murder, with a crime. The groups also differed signifi- mean age of 35.2 years for NGRI and cantly for substance abuse at the time of 34.1 years for convicted murderers. Both the crime. Records revealed that none of groups also had similar levels of educa- the NGRI murderers but 27 percent of the tion, with a mean level of education of convicted murderers abused substances at 13.3 years for NGRI and 1 3.1 years for the time of the crime (X2 = 4.04, df = 1, convicted murderers. Both groups had p < .05). identical total numbers of victims: 17 Table 1 presents characteristics of per- deaths for the 13 NGRI murderers and 17 petrator-victim relationships for both deaths for the 15 convicted murderers. groups. As in packer's17 data, chi-square For NGRI murderers, 23 percent involved analysis revealed a significantly higher multiple deaths compared with 13% for incidence of blood relatives as victims for the convicted murderers. NGRI murderers than for convicted mur- As would be predicted by the Massa- derers (2= 11.87, df = 1,p < .001). For

J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 165 Nestor and Haycock

Table 2 Neuropsychological Test Scores of NGRI Subjects and Convicted Murderers NGRl Convicted (n = 13) (n = 15) Wechsler Adult Intelligence Scale-Revised Full Scale IQ 100.1 ? 12.8 97.6 2 13.2 Verbal IQ 103.8 ? 13.1, 97.2 ? 15.8 Performance IQ 95.1 ? 10.2 98.4 2 13.4

Wechsler Memory Scale-Revised Verbal Memory lndex Visual Memory lndex Attention/Concentration Index

Wisconsin Card Sorting Test Categories achieved

Trail Making Testa Trails A Trails B

Wide Range Achievement Test-Revised Reading Spelling Arithmetic "Seconds. *p < .05.

NGRI murderers, 69 percent involved 15 convicted murderers (7%) involved blood relatives as victims in comparison related children as victims. None of the to 7 percent for convicted murderers. For NGRI murderers involved unrelated chil- NGRI murderers, 62 percent involved dren, whereas 2 of the 15 convicted mur- parents, whereas no convicted murderer derer (13%) involved unrelated children. had killed a parent (2= 12.9, df = 1, Table 2 presents neuropsychological p < .001). Groups also differed signifi- test scores for both groups. On standard- cantly in percentages of victims who were ized tests of intelligence, both groups significant others, mainly a spouse or a scored in the average range of abilities. lover. For convicted murderers, 60 per- NGRI and convicted murderers did not cent of the crimes involved a significant differ significantly on Full Scale IQ, Ver- other in comparison to 23 percent for bal IQ, or Performance IQ. Mean Full NGRI murderer (X2 = 3.88, df = 1, y < Scale IQ was 100.1 for NGRI murderers .05). The groups did not differ signifi- and 97.5 for convicted murderers. Mean cantly in percentages of either related or Verbal IQ was 103.8 for NGRI and 97.2 unrelated children as victims. Two of the for convicted murderers, and mean Per- 13 NGRI murderers (15%) and 1 of the formance IQ was 95.1 for NGRI murder-

166 J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 NGRI of Murder ers and 98.3 for convicted murderers. derers and convicted murderers did not NGRI and convicted murderers also did differ on most neuropsychological tests. not differ significantly on standardized NGRI murderers thus showed a similar tests of memory, attention and concentra- profile to that of psychotic homicides, but tion, or executive abilities. By contrast, a very different profile from that of psy- on standardized tests of academic abili- chiatrically involved, convicted murder- ties, NGRI murderers scored significantly ers. Similar to our findings regarding psy- higher on standardized academic tests of chotic homicide^'^. 20 and the findings of spelling (t = 2.57, df = 23, p < .05) and Packer,17 NGRI murderers are generally written arithmetic (t = 2.36, df = 23, p < psychotic and may kill a blood relative .05). while in the throes of an acute psychotic episode. The nature of the psychotic epi- Discussion sode and the selection of the victim may The current results indicate that homi- very well be inextricably linked. For psy- cide perpetrators likely to receive an in- chotic acts of lethal or near lethal vio- sanity acquittal may be distinguished lence, personal victims are often specifi- from psychiatrically involved, convicted cally incorporated into the psychotic murderers on a number of variables in- symptoms experienced at the time of the volving both mental status at the time of crime, as in the case of the psychotic the offense and the objective characteris- perpetrator with impostor delusions at- tics of the homicide. As expected, all but tacking a victim believed to have been one of the NGRI murderers showed clear replaced by an impostor.20 In the current evidence of psychosis around the time of study, for example, two of the NGRI psy- the murder, in comparison to only 1 of the chotic murderers killed parents whom 15 convicted murderers. In this sense, they believed had been replaced by im- NGRI murderers in this study may very postors. In another case, a NGRI psy- well represent a legitimate and reasonable chotic murderer with fixed, religious de- representative sample of clinically valid lusions killed his father in response to a NGRI murderers. In addition, consistent divine message he reported experiencing with previous findings,'' substance abuse prior to committing the murder. In this at the time of the index offense was less case, the NGRI murderer killed his father, likely to be a factor in NGRI murders. whom he loved, as an example of his The groups also differed dramatically in devotion to a higher being. terms of the nature of victim-perpetrator For NGRI homicide perpetrators, relationships. NGRI murderers were records often describe exceptionally per- more likely to kill blood relatives, espe- sonal, intimate experiences of what may cially parents, whereas convicted murder- be viewed as some type of alien force ers were more likely to kill a significant directly controlling their actions or the other, such as a lover or a spouse. In actions of others. The NGRI murderer contrast to these rather robust group dif- may experience these alien forces as hav- ferences on these variables, NGRI mur- ing their own autonomous purposes.

J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 167 Nestor and Haycock

These forces may command the NGRI The current study did not examine spe- murderer to attack and kill a personal cific symptomology but did suggest that target, perhaps because of religious delu- command hallucinations and paranoid de- sions, as in the case of the son killing his lusions are likely to be present in NGRI father in response to a perceived divine murders. While clearly not unique to any message. In other instances, these forces one diagnosis, these symptoms often do may be experienced as imminent threats characterize paranoid schizophrenia, to life and are believed to originate from which may be associated with increased a particular personal target whom the homicide rates in relation to both control NGRI murderer kills in an apparent act of samples and mentally ill offender^.^'-^^ psychotic self-defense. Similar types of As an example, in relation to a contrast paranoid psychotic experiences may be group of hospitalized psychiatric patients, predictive of violent acts in mentally ill Shah et al.4 reported a significantly persons living in the community.28 higher incidence of paranoid schizophre- For NGRI murderers, these experi- nia in NGRI acquittees. 40 percent of ences may be more intense, elaborate, and whom had been charged with either homicide or attempted homicide. The rela- systematized than those associated with tively intact neuropsychological function- nonlethal violence in mentally ill persons ing of the NGRI murderers of the current living in the community. In addition, for study would also be consistent with para- NGRI murderers, these experiences may noid schizophrenia, which, unlike other occur against a backdrop of relatively in- types of schizophrenia, may not substan- tact cognitive abilities. In the current tially compromise cognitive function- sample, neuropsychological test scores ing.34 However, NGRI murderers may did not distinguish NGRI murderers, all also show other symptoms, which may but one of whom had clear histories of not be entirely consistent with paranoid psychosis, from psychiatrically involved, schizophrenia. In this regard, some stud- convicted murderers, all but one of whom ies have suggested that violence in men- had no history of psychosis. Likewise, the tally ill persons may be linked to both psychotic NGRI murderers generally paranoid delusions and a mood distur- scored higher on neuropsychological tests ban~e.~~,36 Future studies will clearly than patients with chronic schizophre- need to examine both sets of symptoms, nia.29 Overall. these data indicate that affective and psychotic, in persons adju- NGRI murderers may represent a distinct dicated as NGRI of murder. group of psychotic patients, characterized The dynamics of NGRI murders may by malignant paranoid delusions yet rel- be best understood by examining the in- atively preserved cognitive abilities. teraction between the perpetrator and the These relatively intact cognitive abilities innocent victim. In the current study, the may provide the wherewithal and organi- characteristics of the perpetrator-victim zational skills to act upon psychotic per- relationships separated NGRI murderers ceptions and beliefs.20*30 from convicted murders to a remarkable

168 J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 NGRI of Murder degree. From these and other similar data, are arguably driven by rational or at least interactive models may eventually be de- nonpsychotic motives, such as money. veloped that will take into account how defiance, anger, or revenge. certain psychotic symptoms are experi- The validity of such an interactive enced and how these uniquely personal model and its accompanying implications experiences influence victim selection. awaits further empirical studies. Suffice it motives, and ultimate homicidal actions. is to say that the current empirical find- For NGRI murders, a victim may have ings indicate that NGRI murderers may symbolic and psychotic significance only represent a distinct if not unique group of to the perpetrator. The motives of the mentally ill patients. For example, inpa- psychotic murder cannot be explained tient violence among mentally ill patients simply in terms of anger, rage, or jeal- has been related to thought disorganiza- ousy, but rather only by careful examina- tion, agitation, and neuropsychological tion of symptoms, experiences. and con- impairment.'7'" By contrast, highly or- sequent actions. ganized, systematized, paranoid delu- An interactive model may also help sions, coupled with relatively preserved neuropsychological functioning, may be distinguish NGRI murder from other related to lethal or near lethal acts of types of unusual murders, which often violence in mentally ill persons. including raise insanity defenses. Consider the those adjudicated as NGRI for murder. cases of those who murder in the service Thus, the nature of the mental illness. the of extreme political ideology. In these environmental context of the violence. instances, the perpetrator is not psychotic, and the level of the criminality all need to but may represent an aggrieved ideologue be considered in any heuristic model that driven by fanatical beliefs that are shared seeks to explain why some mentally ill by other like-minded members of the po- persons become violent. engage in felo- litical group. In that regard, the perpetra- nious crimes, or commit homicide (see tor may be viewed not so much as delu- also Lapierre et sional but rather as hateful and In summary, perhaps no criminal defense prejudiced. Targets such as persons rep- generates more controversy than that of the resenting an opposing political viewpoint NGRI defense for the charge of murder. are selected not in terms of psychotic The aforementioned distinguishing charac- delusions but exclusively in terms of po- teristics of successful NGRI defenses may litical significance. These perpetrators help to provide some perspective to this will likely crave public attention. Like- heated controversy of how a mentally ill wise, NGRI murderers who kill blood person may be adjudicated not criminally relatives. especially parents, may also be responsible for murder. Future studies will distinguished from other murderers who likely establish additional distinguishing kill their parents. These latter persons characteristics of NGRI murders. Epidemi- may be younger, as in the case of the ological studies are also needed to help to antisocial teenager, are not psychotic, and understand these current findings, which

J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 169 Nestor and Haycock were derived 'from objective but retrospec- 14. Bloom JD, Williams M: Management and Treatment of Insanity Acquittees: A Model tive analyses of cases selected on the basis for the 1990's. Washington, DC: American of clinical criteria. Psychiatric Press, 1994 15. Steadman HJ, McGreevy MA, Morrissey JP, Callahan LA, Robbins PC, Cirincione C: Be- References fore and After Hinckley: Evaluating Insanity Defense Reform. New York: Guilford Press. 1. Beck JC: Epidemiology of mental disorder 1993 and violence: beliefs and research findings. 16. Phillips BF, Hornik JA: The Insanity Defense Ham Rev Psychiatry 2:l-6, 1994 in Massachusetts. Boston: Massachusetts De- 2. Monahan J: Mental disorder and violent be- partment of Mental Health, 1984 havior: perception and evidence. Am Psychol 17. Packer IK: Homicide and the insanity defense: 47511-21, 1992 a comparison of sane and insane murderers. 3. Callahan LA, Steadman HJ, McGreevy MA, Behav Sci Law 5:25-36, 1987 Robbins PC: The volume and characteristics 18. Beaudoin MN, Hodgins S, Lavoie F: Homi- of insanity defense pleas: an eight-state study. cide, schizophrenia, and substance abuse or Bull Am Acad Psychiatry Law 19:331-8, dependency: Can J Psychiatry 38541-6, 1991 1993 4. Shah PJ, Greenberg W, Convit A: Hospital- 19. Nestor PG: Neuropsychological and clinical ized insanity acquittees level of functioning. correlates of murder and other forms of Bull Am Acad Psychiatry Law 22535-93, extreme violence in a forensic psychiatric 1994 population. J Nerv Ment Dis 180:418-23, 5. McGreevy MA, Steadmann HJ, Dvoskin JA, 1992 Dollard N: New York State's system of man- 20. Nestor PG, Haycock J, Doiron S, Kelly J, aging insanity acquittees in the conlmunity. Kelly D: Lethal violence and psychosis: a Hosp Community Psychiatry 42:s 12-7, 199 1 clinical profile. Bull Am Acad Psychiatry 6. Midgley M: Wickedness: A Philosophical Es- Law 23:331-41, 1995 say. London: Routledge, 1985 21. Model Penal Code, 5 4.01 (Proposed Official 7. Prins H: Psychiatry and the concept of evil: Draft, 1962) sick in heart or sick in mind? Brit J Psychiatry 22. American Psychiatric Association: Diagnostic l65:297-302, 1994 and Statistical Manual of Mental Disorders 8. Gunn J: Trials of psychiatry: insanity in the (ed 3 rev). Washington, DC: APA, 1987 twentieth century, in The Mentally Disordered 23. Wechsler D: Wechsler Adult Intelligence Offender. Edited by Herbst KR, Gunn J. Lon- Scale-Revised. New York: Harcourt Brace don: Butterworth-Heinemann in association Janovich, 198 1 with the Mental Health Foundation, 1991, pp. 24. Wechsler D: Wechsler Memory Scale-Re- 17-36 vised. New York: Psychological corporation, 9. Haycock J: The ideology of "dangerousness" 1987 in the creation of separate institutions for the 25. Lezak M: Neuropsychological Assessment. "criminally insane." Presented to the Section New York: Oxford University Press, 1983 on Crime, Law, and Deviance, Annual Meet- 26. Heaton RK: A Manual for the Wisconsin Card ing of the American Sociological Association, Sorting Test. Odessa, FL: Psychological As- Los Angeles, CA, August 5-9, 1994 sessment Resources, 1981 10. Coid J: The epidemiology of abnormal homi- 27. Jastak S, Wilkinson GS: WRAT-R. Wilming- cide and murder followed by suicide. Psychol ton, DE: Jastak Associates, 1985 Med 13:855-60, 1983 28. Link BG, Steuve A: Psychotic symptoms and 11. Cot6 G, Hodgins S: The prevalence of major the violent/illegal behavior of mental patients mental disorders among homicide offenders. compared to community controls, in Violence Int J Law Psychiatry 15:89-99, 1992 and Mental Disorder. Edited by Monahan J, 12. Wolfgang ME: Patterns in Criminal Homi- Steadnian H. Chicago: University of Chicago cide. Oxford: Oxford University Press, 1958 Press, 1994 13. Janofsky JS, Vandewalle MB, Rappeport JR: 29. Nestor PG, Shenton ME, McCarley RW, et al: Defendants pleading insanity: an analysis of Neuropsychological correlates of MRI tempo- outcome. Bull Am Acad Psychiatry Law 17: ral lobe abnormalities in schizophrenia. Am J 203-11, 1989 Psychiatry 150: 1849-55, 1993

170 J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997 NGRl of Murder

30. Taylor PJ: Schizophrenia and crime: distinc- paranoid and nonparanoid schizophrenia. tive patterns in association, in Mental Disor- Schizophr Res 12:223-36, 1994 der and Crime. Edited by Hodgins S. New- 35. Buchanan A: Acting on delusion: a review. bury Park, CA: Sage Publications, 1993 Psycho1 Med 23: 123-34, 1993 3 1. Addad M, Benezech M, Bourgeois M, Yesav- 36. Kennedy HG, Kemp LI, Dyer DE: Fear and age J: Criminal acts anlong schizophrenics in anger in delusional (paranoid) disorder: the French mental hospitals. J Nerv Ment Dis association with violence. Br J Psychiatry 169:289-93, 1981 16O:488 -92, 1992 32. Benezech M, Bourgeois M, Yesavage J: Vio- 37. Adams JJ, Meloy JR, Moritz MS: Neuropsy- chological deficits and violent behavior in in- lence in the mentally ill: a study of 547 pa- carcerated schizophrenics. J Nerv Ment Dis tients at a French hospital for the criminally 178:253-6, 1990 insane. J Nerv Ment Dis 168:698-700, 1980 38. Krakowski ML, Convit A, Jaeger J, Lin S, 33. Taylor PJ, Gunn J: Violence and psychosis 11: Volavka J: Neurological impairment in vio- effect of psychiatric diagnosis on convictions lent schizophrenic inpatients. Am J Psychiatry and sentencing of offenders. Br Med J 288: 146:849-53, 1989 1945-49, 1984 39. Lapierre D, Braun CMJ, Hodgins S, Levelille 34. Kremen WS, Seidman LJ, Goldstein JM, S, Constantineau C: Neuropsychological cor- Faraone SV, Tsaung MT: Systematized delu- relates of violence in schizophrenia. Schizo- sions and neuropsychological function in phr Bull 21:253-62, 1995

J Am Acad Psychiatry Law, Vol. 25, No. 2, 1997