Neurologic abnormalities in murderers

Violent crime, particularly homicide, is one of the most pressing social problems in the United States today. The incidence of violent crime more than doubled in the last 20 years, and more than 23,000 acts of murder or non-negligent manslaughter were committed in 1992. ^[1] To stem the rising tide of violence, efforts must be made to understand its causes. Previous attempts have narrowly focused on environmental factors, physical abnormalities, or genetic abnormalities alone that could predispose an individual to violent behavior. ^[2,3]

There is a considerable body of literature indicating that malfunctioning of certain brain centers, particularly within the limbic system and the temporal lobes, ^[4-6] occurs with intense emotion and may result in violent behavior. These studies in animals and clinical reports have suggested that violent individuals suffer from intrinsic neurologic deficits and that their violent acts may be the direct result of epileptiform-type discharges of neurons in these sensitive brain regions. The term ``episodic dyscontrol'' <sup>[4]</sup> has been used to describe paroxysmal outbursts of violent behavior thought to occur in this manner. The Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV), uses the term ``impulse control disorder: intermittent explosive disorder'' to indicate a similar behavioral pattern.

Our own early hypotheses were much influenced by the concept that a limbic seizure or similar limbic or temporal lobe-mediated event was the cause of violent behavior, ^[7,8] but our evaluations of violent offenders have not fully supported it. ^[9,10] Temporolimbic abnormalities explain the acts of only a minority of violent offenders, and seizures are uncommon; however, neurologic abnormalities of some sort have been the rule among the many violent offenders we have evaluated. We have heretofore provided little definition of the neurologic abnormalities we have encountered in violent individuals--for example, the frequency, quality, and correlation of abnormalities on the clinical neurologic examination, EEG, neuropsychological testing, and neuroimaging studies.

This report concerns 31 of the 34 murderers evaluated by one of us (J.H.P.) since 1989. It describes in detail the evidence of neurologic dysfunction in these individuals and proposes a theory of how neurologic impairment may contribute to violence.

Methods.

Thirty-one murderers who were awaiting either trial (3) or sentencing (10) or were undergoing clemency appeals (18) when seen were evaluated neurologically at the request of their defense counsel between 1989 and 1993. The three individuals awaiting trial were confessed perpetrators; the issue of responsibility had not yet been settled in court. Three other subjects were examined but are not reported here. In two of these three subjects, data were insufficient because the subjects would not allow full physical neurologic examination; in the third, examination was unreliable because he had sustained a penetrating brain injury at the time of his arrest.

Our subjects did not commit their violence in time of war. None was a paid political assassin, terrorist, or Mafia figure. Several had been involved in gangs, driveby shootings, and rapes. Six were serial murderers and two were mass murderers. One had killed his infant son, and another had killed his three siblings. One had killed his girlfriend. Five had killed during an armed robbery. Four had killed a woman either after sexual intercourse or after the woman had refused. Five individuals had killed in an impulsive, unexpected manner.

In all cases counsel hoped to uncover any evidence that might assist the defense, but only one of the 31 subjects, a young man with cerebral palsy, was impaired by a neurologic deficit that would have been obvious to a nonphysician. All the others appeared ``normal'' to their counsel and the authorities. Medical history and neurologic examination were performed on all by one board-certified neurologist (J.H.P.). Each history and physical examination required about 3 hours. Review of pertinent records required the same. Particular attention was paid to evidence of ``frontal'' and ``cortical'' dysfunction. Signs that were routinely sought were those generally considered to be reflective of general cortical dysfunction or, perhaps more specifically, frontal lobe dysfunction. These include the snout, ^[11] suck, ^[11] and grasp ^[11,12] reflexes, paratonia, ^[11] abnormal smooth pursuit eye movements, ^[13] abnormal word fluency, ^[14] and an abnormal response to the test for assessing reciprocal coordination of the movements of the hands (a simultaneous change of position of clenched fist with outstretched hand). ^[15,16]

Complete psychiatric evaluation was undertaken in 12 subjects. Each of these evaluations required 6 to 10 hours. In others, the psychiatric findings and diagnosis rested with the neurologist.

Neuropsychological testing, including the Wisconsin Card Sorting Test and the Halstead-Reitan Battery (including Trail Making Test Parts A and B and the Category Test), was recommended in all subjects but was not obtained in all. IQ testing was also recommended. Results of previous IQ testing were used if current testing was not performed. Twenty-five subjects had some or all of the above testing performed, but the results on any given test were available for review in only a minority of subjects. In the remainder, the report containing the impression of the examining psychologist was noted, with particular reference to the presence of a frontal or nonfrontal pattern of abnormality. Any definite abnormality on the Wisconsin Card Sorting Test, ^[17] Trail Making Test Part A or B, or the Category Test was considered frontal.

EEG was recommended in all and obtained in 20 subjects. Neuroimaging was recommended in all and obtained in 19 (MRI in 15, CT in 4).

Extensive history was obtained with regard to the presence of physical or sexual abuse and other aspects of intrafamily violence. A subject was considered to have been physically abused if he had been beaten with an object other than an open hand or the leather part of a belt on any part of the body other than the buttocks. A subject was considered to have been overtly sexually abused if any fondling of the genitalia of the child, or of the abuser by the child, had occurred, or if any oral, vaginal, or anal penetration had occurred. The presence of a history of abuse was suspected in all and pursued systematically during interviews with the subject and his family and from careful review of his medical, social, school, and criminal records. Family members were interviewed by a member of the defense team in all cases.

Any records of attention deficit disorder ^[18] (with or without hyperactivity), learning disability, or criminal or deviant behavior (particularly fire-setting and cruelty to animals) were reviewed, as was any available documentation of earlier medical, psychological, or psychiatric evaluations of the subject. The presence of learning disabilities, alcoholism, or neurologic or psychiatric disease in the subject's family was also noted.

None of the observers of the dependent variables was blinded. Neurologic examination, psychological testing, MRI, and EEG were performed on manacled prisoners attended by guards.

Results.

There were 19 white subjects, 11 black subjects, and one Hispanic subject. The average age at the time of homicide was 32.7 years (SD, plus minus 9.21; range, 14 to 63). Four subjects were younger than 18 at the time of homicide. All subjects were male.

Subjects were divided into six groups based on the findings on neurologic examination Table 1. One of the frontal findings was elicited in seven of the 31 subjects examined, or 22.6% (group 1); two of these findings were elicited in three subjects, or 9.7% (group 2); and three or more of these findings were present in 10, or 32.3% (group 3). Thus, a total of 20 of the 31 subjects, or 64.5%, had some physical evidence of frontal dysfunction.

Neuropsychological testing was performed in four of the seven subjects in group 1. All four of the studies were reported to be abnormal, and a frontal pattern of abnormality was demonstrated in one. Results of IQ testing performed previously were available in two other subjects and were normal in both. All three of the subjects in group 2 underwent neuropsychological testing, and frontal lobe abnormality was demonstrated in two. IQ testing was also abnormal in both of these. All 10 subjects in group 3 received some testing, and frontal lobe dysfunction was found in five. IQ testing results were reported in nine of the subjects in this group; of these, two were mentally retarded (IQ less than 70), three were borderline mentally retarded (IQ less than 80), and four had normal Full Scale IQs.

EEG was performed in three of the subjects in group 1 (1 abnormal), two of those in group 2 (both normal), and seven of those in group 3 (4 abnormal). None of the EEGs in groups 1 to 3 showed epileptiform or paroxysmal activity; all showed slowing, which was diffuse or symmetric.

MRI or CT was performed in three subjects in group 1 (all normal), two in group 2 (both abnormal), and six in group 3 (all abnormal).

Seven individuals (group 4) had unequivocal asymmetric findings, which included a hemiparesis in four, a unilateral Babinski sign in two, and a cape-like sensory loss in one who was found on MRI to have a large cervical syrinx. Five members of group 4 were also in group 1, 2, or 3--they had both frontal and asymmetric findings. Four of these underwent neuropsychological testing, which revealed abnormalities in all four. A frontal pattern of dysfunction was found in three. EEG was obtained in three and was abnormal in two. Neuroimaging was performed in five and was abnormal in three.

Six subjects had a normal physical neurologic examination but abnormal mental status testing ^[19] (group 5). These included incorrect serial seven subtractions (6) and the inability to register and recall three objects (1). Only one of the six subjects had more than one abnormal response on mental status testing. Neuropsychological testing was performed in five of these six subjects. Frontal tests were abnormal in both of the two subjects who were specifically tested. IQ testing was normal in the four who were tested. EEG was obtained in four subjects and was abnormal in two, one with bitemporal spikes and a second with sharp waves and slowing in the temporo-parietal area. MRI was obtained in five subjects and was normal in all.

Three subjects had completely normal neurologic and mental status examinations (group 6). Neuropsychological testing, EEG, and MRI were performed in two of these three. Neuropsychological testing was abnormal in both, one of whom had a frontal and the other a nonfrontal pattern of dysfunction. EEG was abnormal in one, showing focal slowing, and MRI was normal in both.

Thirty of the 31 subjects underwent some type of testing other than the neurologic examination. Of these 30, no subject was normal in all spheres (ie, neurologic examination, EEG, neuropsychological testing, and neuroimaging). Of the 19 individuals who underwent neuroimaging, nine had abnormal findings. Abnormal brain MRI findings correlated significantly with abnormality on physical examination (p less than 0.05).

Specific neurologic diagnoses could be established in 20 of the 31 subjects (some had more than one neurologic diagnosis). The diagnoses included fetal alcohol syndrome (5), mental retardation with Full Scale IQ less than 70 (5), and borderline mental retardation with Full Scale IQ greater than 70 but less than 80 (4). There was one each of the following diagnoses: athetoid cerebral palsy, hypothyroidism with psychosis, mild cerebral palsy with mental retardation, pituitary microadenoma with acromegaly and borderline mental retardation, hydrocephalus with elevated pressure, and dementia. Two had a history of epilepsy and three had trauma-induced brain injury. Two subjects had dementia that was thought to be alcohol-induced.

Psychiatric symptoms included varying degrees of paranoia in all subjects, most typically manifested as excessive suspiciousness, an inability to sustain relationships, the carrying of weapons for protection in the absence of an identified threat, and an inability to make or maintain eye contact during interviewing. In addition, eight subjects manifested frank paranoid delusions and met the DSM-IV diagnostic criteria for schizophrenia, paranoid type. ^[20] Depression was present in nine, and a convincing history of bipolar affective disorder was present in three. Four subjects manifested a history strongly suggestive of dissociative identity disorder, ^[21] although logistical constraints in these did not permit demonstration of the two or more personalities as required by DSM-IV for the diagnosis of multiple personality disorder.

Reviews of school records and psychological assessments revealed evidence of attention deficit disorder in 11, two of whom had once been treated with methylphenidate. Nine had been either suspected of or diagnosed as having a learning disability by school officials. Several of the subjects also had a family history of learning disabilities. Fifteen subjects met the DSM-IV diagnostic criteria for conduct disorder ^[22] as juveniles.

A history of alcohol abuse in a persistent and prolonged fashion was obtained in 17 individuals (55%), all but one of whom had ingested alcohol shortly before committing homicide. Fifteen individuals had a family history of severe alcoholism in at least one parent.

There was a documented history of severe, prolonged physical abuse in 26 of the 31 subjects. In 13 of these subjects, the history was substantiated by the presence of scars and other signs of physical injury. A family member was also frequently able to corroborate the history. One subject denied abuse by his family but did sustain severe beatings (enough to result in severe bleeding) while incarcerated during his early teenage years. Abuse was strongly suspected but could not be substantiated in three subjects, one of whom had multiple unexplained scars on his back and another of whom had several emergency room visits for unexplained trauma. In only two of the 31 subjects was there neither evidence nor suspicion of abuse.

There was a documented history of overt sexual abuse in 10 subjects. A history of covert abuse was elicited from one subject, whose father instructed him in oral sex at the age of 5 years. Sexual abuse was suspected in four other subjects, one of whom had been cared for by a relative who was a known pedophile. No history of sexual abuse could be elicited in 16 subjects.

A history of symptoms suggestive of complex partial seizures, manifested primarily by memory lapses with possible staring spells, was elicited in five individuals. These were infrequently accompanied by symptoms of ``dizziness'' or macropsia. No subject had been treated for complex partial seizures. Of the 20 subjects who underwent EEG, eight had EEG abnormalities. Of these, four demonstrated abnormality arising from the temporal lobe. Two showed focal temporal slowing, one showed bitemporal spikes, and one was positive for sharp waves and slowing.

Of the 19 subjects who underwent neuroimaging, three demonstrated temporal lobe abnormality with reduced volume of one of the temporal lobes. In all, only nine of the 31 subjects either had symptoms suggestive of seizures arising from the temporal lobe or had temporal lobe abnormalities on EEG, MRI, or both. Of these nine, one had both symptoms and EEG abnormalities. Of these nine, six also demonstrated frontal signs.

Discussion.

There have been very few thorough studies of the neurologic status of violent perpetrators. EEG studies ^[23] of 64 prisoners charged with murder in the 1940s revealed that of those whose crime was without an apparent motive and of those found to be ``insane,'' more than 70% had severe and focal brain abnormalities. Episodes of directed violence have been attributed to complex partial seizures <sup>[4,24-26]</sup> or a similar entity once called ``episodic dyscontrol'' ^[4,27,28] and now labeled ``impulse control disorder.''

The relatively small population available for the present study yielded a high rate of neurologic abnormality. All but one (97%) had abnormalities that reflected brain dysfunction, and a specific neurologic diagnosis could be made in 20 (64.5%).

Previous studies of violent individuals implicated the limbic system, ^[4,29] but only nine of these murderers had any evidence of temporal lobe abnormality. Much more impressive was the high frequency of frontal signs or, more generally, cortical or subcortical signs in this population. Clearly, those chosen for examination by their lawyers were not randomly picked, and a control group was not available for comparison. Even with these limitations, the frequency of neurologic deficits was surprisingly high.

In an earlier study of 97 boys and 22 girls residing at a state correctional school, ^[30] we showed that more violent children were more likely than less violent incarcerated children to demonstrate neurologic abnormalities and paranoid ideas and to have been abused. We compared matched samples of 31 incarcerated delinquents and 31 nondelinquents. The constellation of abuse, family violence, psychiatric symptomatology, cognitive impairment, and neurologic signs correctly predicted group membership 84% of the time. The same constellation also distinguished the more aggressive from the less aggressive subjects in each group.

In this study, we defined frontal dysfunction as the presence of one or more ``frontal release signs.'' This definition may not precisely describe the anatomic site of dysfunction, as such signs can reflect general cortical or even subcortical dysfunction, and a single sign is encountered in about 5% of the general population over 50 years of age. ^[31] Such frontal signs were present in 20 of the 31 individuals described here, only one of whom was over age fifty.

Jenkyn et al ^[32] reported the predictive value of certain frontal signs for evidence of more diffuse cortical dysfunction on neuropsychological testing. They described 13 neurologic signs that correlated significantly with abnormal neuropsychological testing. These included an abnormal response to the following tests: nuchocephalic reflex, glabellar blink, conjugate gaze (upward and downward), smooth pursuit, suck reflex, lateral gaze impersistence, paratonia of arms and legs, limb placement, accurate spelling of ``world'' in reverse, accurate recall of three objects after 5 minutes, and accurate recall of the past three US presidents in reverse. The presence of three or more signs correctly predicted abnormal cortical function in 64% of mildly impaired subjects, in 96% of moderately impaired subjects, and in 100% of markedly impaired subjects. Among our subjects, eight (25.8%) exhibited three or more of these 13 signs originally validated by Jenkyn et al. Neuropsychological testing was abnormal in all eight of these subjects: four were mentally retarded, two had borderline mental retardation, one demonstrated low memory and learning capabilities, and one demonstrated severe frontal lobe dysfunction.

Mesulam ^[33] demonstrated the importance of frontal lobe functioning in behavior. The frontal lobes generally have an inhibitory effect upon other cortical areas ^[15] and may serve to ``filter out'' information that is not relevant, thus allowing one independence from environmental stimuli. The behavior of patients with frontal lobe lesions is inordinately dependent on environmental stimuli. ^[34,35]

A recent computer analysis of the type of damage sustained by the frontal lobes of Phineas Gage, a paradigm of frontal lobe dysfunction, revealed that his lesion was located in the ventromedial region of both frontal lobes ^[36]--the ``prefrontal cortex''--damage to which gave rise to behavioral changes commonly labeled the ``frontal lobe syndrome.'' This type of injury may lead to dramatic changes in personality and conduct while preserving cognitive, motor, and language functioning. The ability to make rational decisions that respect social convention and regulate emotional response presumably would be impaired by a lesion in this anatomic area. ^[15]

This type of emotional alteration, coupled with a form of ``environmental dependency syndrome,'' in which one's behavior is inordinately controlled by external stimuli, may explain some violent behaviors, lack of remorse, and moral and ethical insensitivity. An exaggerated response to an environmental stimulus in an individual with frontal lobe dysfunction may result in an action far exceeding the normal reaction. Many of the acts of homicide committed by the present series of individuals were evoked by stimuli that in most people would arouse simply a feeling of anger or frustration: the murder of women who rejected sexual advances, one young boy's murder of his three younger siblings while waiting to attack his abusive father, the fatal beating of a crying infant.

In some subjects, the manner in which damage to the brain had been sustained was obvious. One had hydrocephalus, two had cerebral palsy, 17 were alcoholic (several of these profoundly so--for example, a subject with alcohol-related pancreatitis at age 16), and 10 had sustained moderate or significant head injury. Microcephaly and other stigmata of the fetal alcohol syndrome indicated congenital factors in five.

The history of physical and sexual abuse (frequent, prolonged, ongoing for years, and committed in most instances by a parent or parent substitute) was pervasive and extreme. There is a rapidly growing body of literature on the devastating effects of such abuse on the life of the victim and on society. ^[37] Aggressive tendencies and poor impulse control, ^[38] depression frequently with psychotic features, ^[39,40] and suicidal behavior ^[41] are some of the long-term effects of sexual abuse. Many reports of abuse do not distinguish single incidents from ongoing, continuous abuse lasting a decade or more; the latter was the likely experience of all but two of the subjects reported here (94%).

The very word ``abuse'' does not convey the full meaning of the experience of our subjects. ``Torture'' might be a more accurate description of what was described and observed. We witnessed burns on the buttocks of one subject who, as a child, was forced to sit on a hot stove for punishment. Another individual was subjected to anal penetration with a cattle prod as a child. Chronic torture such as this may be responsible for the formation of dissociative states, ^[21] a syndrome that we encountered in four subjects. Complete lack of conscious recall for periods and other features (overidentification with a make-believe friend, a history of episodes for which the subject has no memory) characterize this condition, one that may be easily overlooked on history and examination unless the subject is carefully approached; such features are routinely missed on psychological testing and ordinary neurologic and psychiatric interview. Dissociative episodes can easily be mistaken for complex partial seizures, especially in a brain-damaged population.

The veracity of the history of abuse is supported by the fact that none of the subjects had offered physical or sexual abuse as a mitigating factor in his defense prior to the examination by the psychiatrist or neurologist. In some subjects, several interviews and examinations were necessary to establish, and to confirm independently, the presence of an abusive background that the subject often denied. Horizontal scars on the back from whipping, healed fractures, cigarette burns, and other stigmata of corporal punishment supplemented the independent recollection of family and friends and medical records to confirm and extend the reports of severe abuse.

The very high frequency of severe abuse, brain dysfunction, and paranoia in this population suggests that these three factors may have an etiologic role in violence. The experience of abuse alone, the presence of brain damage alone, or the symptom of paranoia alone cannot explain the formation of a violent individual, as the vast majority of individuals who have suffered or manifested each of these conditions alone are not violent.

However, abnormal psychological development caused by long-standing exposure to severe abuse, together with paranoia and an impaired ability to deal with frustrating environmental factors due to frontal dysfunction, may provide an explanation for the commission of a homicidal act. We have, therefore, evolved the theory that the interaction of abuse, paranoia, and neurologic dysfunction provides the matrix of violence. ^[42] Also significant is the very high rate of alcohol use (52%) at the time of homicide. Alcohol could serve to further disinhibit an already disinhibited personality, prompting an act of impulsive violence.

The two subjects in this series who did not come from an abusive background (as best could be determined) were psychotic. One was an unmedicated schizophrenic with myxedema, and the other was a psychotic 14-year-old adopted child born to a heroin-addicted mother from a family rife with manic-depressive psychosis. Thus, these individuals had very obvious sources of brain dysfunction and paranoia with poor impulse control. In these extreme cases, these features alone may have resulted in violence.

None of the murderers fully examined was neurologically or psychologically normal. The ramifications of this finding are important not only for the defense of these individuals but also possibly for the early recognition of potentially violent persons. Fully 25 (81%) of the subjects demonstrated behaviors at an early age that should be recognized as possible harbingers of more violent acts at a later age. Some examples of these included alcoholism, setting fires (to churches and even their own homes), torturing and killing animals, carrying knives or guns to school, and actual assault. Other worrisome factors included attention deficit disorder with hyperactivity, conduct disorder, and fetal alcohol syndrome.

Early recognition of children or youths who are at risk for being violent should prompt an evaluation for the presence of violence in the home and of a neurologic or psychiatric abnormality. Any intervention, if it is to be effective, probably depends on the early recognition of risk factors.

Acknowledgments

We thank Dorothy Otnow Lewis, MD, for assistance with the psychiatric evaluation of several of the subjects; Susan Matthews for assistance with preparation of the manuscript; and Pamela Reese, Sarah Morris, MS, and Gary Chase, PhD, for assistance with collection and interpretation of data.