- Email: [email protected]
Violent Juvenile Delinquents: Psychiatric, Neurological, Psychological, and Abuse Factors
Violent Juvenile Delinquents Ps ychiatric, Neurological, Psychological, and Abuse Factors
Dorothy Otnow Lewis, M.D., F.A.C.P., Shelley S. Shanok, M.P.H., Jonathan H. Pincus, M.D., and Gilbert H . Glaser, M.D. Abstract , Th is stud y compares the neu ropsychiatric. intellectual, a nd educational status of
ex tr emely violent and less violent inca rcerated boys. The more violent ch ild re n were more likel y to demonstrate psychotic symptomatology (paranoid ideation and rambling. illogical associa tio ns) and to have major and minor neurological abnorm aliti es . They were also more likely to have experienced and witn essed e xtre me physical abuse. The co nt r ibu tio n of these factors to violent delinquency is dis cu ssed a nd implications for therapeutic intervention are
suggested.
The purpose of this paper is to present the results of a comparison study of a sample of extremely violent delinquent boys and a sample of less aggressive, male juvenile offenders. This study sprang from earlier findings that man y delinquent children referred to a juvenile court clinic suffered from hitherto unrecognized, undiagnosed, and untreated psychiatric, neurological, and educational disorders which seemed to contribute to their antisocial behaviors (Lewis, 1976a; Lewis and Balla, 1976; Lewis and Shanok, 1976). This raised the question : are there any associations between the severity of a child's antisocial behaviors and his psychiatric, neurological, and psychoeducational status? There is disagreement in the literature whether psychiatric and neurological disorders are more prevalent in the delinquent and
Dr. Leuns is Associate Clinical Pr ofessor of Psychiatry at the Yair University Child Study Center; M s. Shanok is Associate in Re search, Yair University Child Study Center; Dr. Pincus is Professor of Nru rology at Yale University School of Medicine; and Dr. Glaspy is Professor and Chainnan of Neu rology at Yair University School of Medicine. R eprints may br requested from Dr. Le uns at the Yair University Child Study Center, 333 Cedar Stre et, N ew Haven, CT 06510. This research was supported by La w Enforcement A ssistancp Administration grant;; 76AE418003X and Ford Foundation grant;; 730A-41-57478. 0002-7 I 38179/1802--{)307 $01.16 e Ame rican Academ y of Child Psych iatry.
307
308
Dorothy Otnow Lewis et al.
criminal population than in the general population (Durbin et aI., 1976; Guze et aI., 1969; Healy and Bronner, 1926; Kloek, 1968; Nameche et aI., 1964; Petrich, 1976; Robins, 1966; Shanok and Lewis, 1976; Shoor and Speed, 1969; Stephenson et aI., 1973; Warren, 1960; Zitrin et aI., 1976). Similar disagreement exists in the literature concerning the medical histories of delinquents (Carper, 1974; Eilenberg, 1961; Gibbens, 1963; Glueck and Glueck, 1950; McCord and McCord, 1959; Scott, 1975; Stott, 1962). There is, however, a growing body of evidence that certain forms of violence are associated with disorders of the central nervous system (Bach-y-rita and Veno, 1974; Bach-y-rita et aI., 1971; Glaser and Dixon, 1956; Lewis, 1976a, 1976b; Mark and Ervin, 1970; Ounsted, 1969; Pincus and Tucker, 1978; Serafetinides, 1965; Treffert, 1964). Clearly, there was a need for a systematic study of the neuropsychiatric status of violent delinquent children. METHOD
Setting
Because the study was primarily concerned with serious offenders, the only correctional school serving the entire state of Connecticut was chosen as the setting. Previous work, emanating from a juvenile court setting, focused on the less serious juvenile delinquents from only one district of the state. The correctional school setting allowed the evaluation of some of the most seriously delinquent children throughout the entire state. Furthermore, within the correctional school, a secure unit had recently been constructed to house especially violent juveniles. Thus the possibility existed of comparing more and less violent children at the school itself. Samples
The study samples consisted of 97 boys, all of whom were incarcerated at the correctional school at some time during the 18-month duration of the project and were evaluated by the clinical team. For research purposes, after all data had been collected on the 97 boys, three of the investigators independently rated each child on a scale of violence from 1 (least violent) to 4 (most violent), using only data regarding behaviors and offenses, without taking into account psychiatric or neurological data. Raters attempted to use the follow-
Violent Juvenile Delinquents
309
ing broad criteria for group placement: A child was to be rated 1 if there were no evidence of his having committed any offense against person or having committed arson. Since almost all of the children at the school had, at one time or another, been involved in fist fights with peers, such behaviors were not considered violent unless a peer had been seriously injured and required hospital care, or if a weapon had been used. There were 8 in this group. Children were rated 2 if there was some indication in their charts of a potential for violence (e.g., isolated episodes of firesetting, isolated episodes of threatening with unloaded weapons, isolated episodes of threatening violence). There were 11 boys in this group. Children were rated 3 if they had actually committed serious offenses against person (e.g., murder, rape, multiple episodes of arson, armed robbery, assault). All of these children had committed more than one offense that endangered others. There were 55 in this group. Children who were considered to have demonstrated extraordinary brutality toward others were rated 4. For example, in this group was a boy who assaulted a woman, stomping on her face and leaving imprints of his shoes; a boy who raped and beat a young girl and subsequently raped and stabbed another; and a boy who shot a gun in the direction of a group of women, and who secreted knives on himself and around his room. There were 23 boys in this group. All ratings of violence were subjective. In almost all cases, there was agreement among raters. When disagreements occurred, a child was rated according to the assessments of two of the raters. DIAGNOSTIC EVALUATION
Psychiatric Assessment
Children in the study were evaluated by a child psychiatrist and a neurologist. The psychiatric evaluation has already been described (Lewis, 1976b; Lewis and Balla, 1976). However, the criteria for the presence or absence of particular psychiatric symptomatology will be reviewed briefly. Auditory or visual hallucinations or paranoid delusions were considered to be significant if children reported such experiences
310
Dorothy Otnow Lewis et al.
and acted in response to them or were clearly troubled by them. For instance, many children reported hearing their mothers calling their names. This was not considered abnormal unless it was associated with panic, an attempt to leave the room in which the voice was heard, or search for the source of the voice, such as looking under the pillow or bed, or out in the hallway or courtyard. Depression was considered to be significant if the children felt sad, said they had crying spells, and reported vegetative symptoms. Sadness alone was not considered to be an adequate basis for diagnosing depression. In the course of the psychiatric and neurological evaluations, the presence or absence of paranoid ideation was assessed. It took much experience to ascertain, as far as possible, the difference between appropriate, adaptive wariness and inordinate suspiciousness. Paranoid symptomatology was felt definitely to be present if children had mistakenly believed that someone was going to hurt them and could provide several examples of this, or if they admitted to constantly feeling the need to carry weapons such as guns and metal pipes for their own protection in the absence of identifiable dangers. Unusual suspiciousness manifested during the interview was not considered an indication of paranoid symptomatology unless it was associated with the kinds of pervasive delusional beliefs and behaviors mentioned above. Neurological Assessment
Standard neurological examinations were carried out. These included measurement of head circumference; evaluation of cranial nerves; motor, sensory, and reflex functions. Tests of coordination included quantification of the number of alternating palm strikes the child could perform during pronation-supination movements in 10 seconds and the number of finger taps the child could perform with his index finger in 10 seconds with each hand. The presence of choreiform movements was determined by asking the child to extend his arms and fingers in front of him and above his head for 5 seconds each. All were asked to skip after the examiner demonstrated the requested pattern of movement. Mental status examination in the course of the neurological evaluation included orientation for time and place. The ability to remember up to 6 numbers forward and 4 numbers backward was tested. Calculation skills were tested by 4 serial subtractions of 7, starting with 100.
Violent Juvenile Delinquents
311
Certain issues were covered both by the neurologist and the psychiatrist. Both attempted to obtain as detailed a medical history as possible. Both tried to ascertain whether or not the child had been a victim of abuse or had witnessed extreme violence. A child was considered to have been abused by his parents or guardians if he had been punched; beaten with a stick, board, pipe, or belt buckle; or beaten with a belt or switch other than on the buttocks. He was also considered to have been abused if he had been deliberately cut, burned, or thrown downstairs or across a room . A child was considered not to have been abused if he was only struck with an open hand, beaten with the leather part of a belt, or beaten with a switch on the buttocks only . Whenever possible, a sleep electroencephalogram was performed. Results of previous electroencephalograms were also obtained whenever possible. Evidence of major neurological abnormality was defined as a documented history of grand mal epilepsy, an abnormal electroencephalogram, positive Babinski sign, or head circumference that was plus or minus two standard deviation from the mean for the child's age . Evidence of minor neurological abnormality was defined as inability to skip, choreiform movements, abnormal reflexes, abnormalities in coordination, or mixed dominance of hands and feet. Subjects were considered to have ps ychomotor epileptic symptomatology if at least two of the following were present: observed staring episodes with loss of fully conscious contact with reality; loss of memory for violent o r nonviolent acts, followed by fatigue; episodes of inability to understand conversations despite an attempt to do so ; dizzy spells followed by fatigue or headache ; episodes of unprovoked extreme anxiety; olfactory or gustatory hallucinations; many recurrent episodes of deja vu; macropsia or micropsia; forced thinking; and automatic repetitive behavior. Psychoeducational Testing
Psychological testing consisting of the Wechsler Intelligence Scale for Children (revised), the Bender-Gestalt Test, and the Rorschach Test was performed by a research developmental child psychologist. When time factors precluded testing by our own psychologist, or when recent test results were available, testing was not repeated, and previously charted test results were used for purposes of obtaining scores on tests and subtests. Educational evaluations were
Dorothy Otnow Lewis et al.
312
performed by a learning disabilities specialist. Reading grade discrepancies were calculated by subtracting the child's current reading grade score from his actual grade placement in school. Possible bias was introduced in the study because in most cases the investigators knew whether or not a subject was considered violent by virtue of the fact that most violent subjects were seen on the secure unit, while most of the less violent subjects lived in open settings and were seen at the school infirmary. Ethically it would have been impossible to have nonviolent children randomly placed on a securely locked unit. Thus it is possible that the preconceived notions of the examiners influenced the quality of the data collected. Our previously described rating process, based on behaviors only, without access to other data, was an attempt to correct for possible bias. FINDINGS
Comparison of Nonviolent Subjects with Violent Subjects
Of the 97 children evaluated. only 8 were considered by all raters to be nonviolent. When the psychiatric, neurological, and psychoeducational status of these children was compared with the violent subjects, a number of significant differences was found. Psychiatrically, more of the violent children suffered from paranoid symptoms (75.5% vs. 0%, X~ = 15.386, P < .001); were loose, rambling and illogical (57.1 % vs. 0%, X; = 7.000, P = .009); and were unable to remember 4 digits backward (55.2% vs. 12.5%, X; = 3.556, P = .060). Although differences did not reach conventional levels of significance in the following factors, none of the nonviolent children had experienced auditory hallucinations, compared with 40.8% of the violent boys; none had experienced visual hallucinations, compared with 28.2% of the violent boys; and none had experienced olfactory or gustatory hallucinations, compared with 12.9% of the violent group. Similarly, the nonviolent group was more neurologically intact than the violent group. For example, none of the nonviolent sample had major neurological signs, compared with 41.6% of the violent group. Of the violent group, 58.7% had choreiform movements compared with 14.3% of the nonviolent group, and 38% of the violent group were unable to skip, compared with 14.3% of the nonviolent group. Because of the very small number of subjects in the nonviolent group, however, these neurological differences did
313
Violent Juvenile Delinquents
not reach conventional levels of significance. Statistically significant differences between the nonviolent group and the rest of the sample were found in the proportion of each group with psychomotor epileptic symptomatology-25.0% (2) nonviolent vs. 69.4% (50) violent, X~ = 4.451, P = .035-and in the proportion of children with a greater than 10% discrepancy between right and left palm strikes-14.3% (1) nonviolent vs. 60.8% (45) violent, X~ = 3.904, P = .049. Intellectually there were no significant differences between the groups in terms of overall IQ as measured by the WISe or on verbal or performance scores, although there was a tendency for the violent group to score lower in the verbal area (91.500 nonviolent vs. 83.158 violent, t = 1.726, P = .088). We wondered whether neuropsychiatric differences between the more and the less violent children existed. In order to investigate this question, the 97 subjects were divided into two groups, those children rated 1 or 2 constituting the less violent group, and those rated 3 or 4 constituting the more violent gr,Jup.
Comparison of More and Less Violent Delinquents Psychiatric Findings. These are summarized in table 1. The most striking difference psychiatrically between the two groups was the finding that a significantly greater proportion of very violent children demonstrated or gave clear histories of paranoid symptomatolTable I A Comparison of the Xumbcr and Percentage of :\fore Violent and Less Violent Delinquent Boys with Specific Psych ia t ric Symptomatology More \'iolent Depressive symptoms Visual hall urinations Auditory hallucinations Olfactory or gustatory hallucinations Tactile hallucinations Paranoid symptomatologv Loose. rambling, illogical Inability to subtract serial 7s Inability to remember 6 digits forward Inability to remember 4 digits backward
X" • r
p value
84.fi 11.1 17.6
0.8.-,2 1.695 2.704
.sss
2.126 0.65!J 24.618 5.12fi 6.l~8
.145 .417 < .001 .024 .014 >.5
Less Violent
#
(f'lI.
,e,
#
2~
fi:;.7
1\
18 28
~O.O
2
43.3
3
II
15.1
0 0
,,,n
(I'
(j
8
5·\ 41
81.8 59.fi 69.5
4 6
0 0 16.7 23.5 33.:{
l-1
26.9
13
18.8
0.109
31
60.8
2
13.3
8.627
2H
OJ
s
.193
.I 01
.004
• Percentages arc based upon the actual number of subjects on whom information was available for each category.
314
Dorothy Otnow Lewis et al.
ogy. Furthermore, they were significantly more likely than their less aggressive peers to be loose, rambling, and illogical in their thought processes during interviewing. In the formal mental status evaluations, they had extreme difficulty remembering even 4 digits backward, and the majority was unable to subtract serial 7s, even when given the opportunity to practice the task. Depressive symptoms were common in both samples. We found that although there was a definite tendency for more of the more violent children to have experienced hallucinations, particularly auditory hallucinations, differences between the two groups in this regard did not reach conventional levels of significance. Neurological Findings. Neurological differences between the very violent and less violent samples of delinquents are summarized in table 2. Almost the entire sample of more violent children (98.6%) had one or more signs of minor neurological impairment, which was significantly different from the less violent sample, 66.7% of whom evidenced minor neurological abnormalities. Table 2 Specific Neurological Signs: Comparison of More Violent and Less Violent Boys More Violent One or more major neurological signs Abnormal EH; Positive Babinski One or more minor neurological si~ns
Inahilirv to skip Chord form movcmcut-, Psychomotor symptoma tolog v Greater than 10";, di"'H'pancy he· tween right and Idt palm strike (~rl'atlT than lUI;, d iscrcpancv hetween right and left finger taps
#
(r'_
"
Less Violent
X~
#
(I'a
I
li.7 0 'i.fi
1I.'i6!l
;-.
(ifi.7 11.1 :11.1
J(i.:!i:'> ·1.!l:!li 3.37'i
p value
,I'
:\1 I!I II
·lfi.:! :!!1.7 l'i."
71
I:!
·11i
!':-Ui 1:1.:1 (iO.ti 71.!1
Ii
:17.:;
;).22:\
.001 .0:!7 .0(i7 .023
10
lil.-,
Ii
:17.:;
2.1~:;
.11;
2ti
I I.!!
.;0.0
O.OO:!
~(i
·10
(I
I
'J
1i.·I!'!l
z.sso
.011 .I0!! >.'i
> ..;
• Percentages arc based upon the actual numher of children for whom data were available for each category.
The presence of major signs of neurological abnormality significantly distinguished the two groups. Of note, almost 30% of the very violent children had grossly abnormal electroencephalograms, usually of a paroxysmal or focal nature, and/or a history of grand mal epilepsy, compared with none of the less violent sample. Psychological Test Results. Differences in performance on psycho-
Violent Juvenile Delinquents
315
logical testing between the two groups are summarized in table 3. Intelligence, as measured on the WISe, indicated a tendency for the less violent children to function somewhat better intellectually, but overall differences between the two groups were not striking. The more violent group did less well on the verbal section of the WISe, the greatest difference between groups occurring on the arithmetic subtest. Both groups scored relatively well on the picture completion and picture arrangement subtests, and both groups received their poorest scores on the vocabulary subtest. They both were several years behind their expected reading grade. Differences between math grade discrepancies did not reach conventional levels of significance. Table 3 A Comparison of Psvchoeduc.u ional Test Results of :\Iore Violent and Less Yiokllt Dcl iuqucnt Children
Average full-scale IQ WISC WISC \erbal score wise performance score ( .omprclu-nsion subrcst Arithmetic suhtcst Similarities suhtcst Vocabulary subu-st Picture completion subtest Pict me a ITaugcmcn t Block design Object assembly Coding Reading grade discrepancy Math grade discrepancy
More Violent
Less Violent
RfiA",!! R:!.5!!7 9:!.H7:; 7.fiF, (j.!J81 8.1 !;; 6.3:!1 10.1i:!3
93.000 RR.RI:! 98.87:; 8.800 !!.!!O!! 8.000 7.400 10.:;1:; lOy,() 9.818
9.922
8.2:;0 9.694 7.60·1
4..1 3..1
ro.ssr
8.58;1 2.300 2.200
t-test 1.94;1 1.978 1.507 1.41;1 3.!JO:! 0.11", 1.395 0.034 0.91:; 1.478 0.580 0.934 2.148 1.602
p value .O:;fi .O:;:! .130 .163 < .001 > .500 .169 > .500 .364 .145 > .500 .354 .037 .115
History of Abuse. Another set of factors that strongly distinguished the more violent from the less violent children related to a history of abuse hy parents or parent suhstitutes. Differences are summarized in table 4. The more violent children had been physically abused by mothers, fathers, stepparents, other relatives, and "friends" of the family. The degree of abuse to which they were subjected was often extraordinary. One parent broke her son's legs with a broom; another broke his fingers and his sister's arm; another chained and burned his son; and yet another threw his son downstairs, injuring his head, following which the boy developed epilepsy. The two samples also differed significantly in their exposure to
Dorothy Otnow Lewis et al.
316 Table 4
Comparison of the Number and Percentage of More \ 'iolent and Less Violent Delinquents with a History of Abuse More Violent
# Abused by mother Abused by father Abused by others Ever abused \\'itness to extreme violence
21 29 23
52 44
pt .
10
43.8
54.7 45.1 75.4 78.6
.
Less Violent
#
"/
2 5 2 6 3
14.3 29.4 14.3 33.3 20.0
, 0
X} 2.869 2.364 3.200 !I.535 15.615
p value .091 .125 .074 .003 <.001
• Percentages are based upon the actual numbers of subjects on whom information was a"a ila ble for each ca tegory.
violence. The fact that 78.6% of the more violent children were known to have witnessed extreme violence directed at others, mostly in their homes, compared with 20.0% of the less violent children, tells only part of the story. The degree of violence witnessed by these children went beyond mere fist fights. Several children witnessed their fathers, stepfathers, or mothers' boyfriends slash their mothers with knives, They saw their siblings tortured wiLh cigaJ'cLlc butts , chained to beds, and thrown into walls. They saw their relatives-male and female-arm themselves with guns, knives, and other sharp instruments and, at times, use these weapons against each other. Some children ran away from home at the approach of certain relatives, while many children reported defending their mothers with pipes and sticks while their mothers were being attacked. We wondered whether degree of violence as rated 1-4 correlated in any way with having been abused. A correlation coefficient of r = 0 .373, P < .00 I was found, indicating that a child's degree of violence was strongly associated with his having been abused. Of note, the family constellation of the more and less violent delinquents (e.g., broken homes, mother in home, father in home) was similar in both groups. It was, rather, the quality of family interactions that distinguished the two groups from each other. Multiple Regression Analysis. In an effort to determine which combination of factors best distinguished the more violent from the less violent delinquents, we conducted a multiple regression analysis, using group membership (very violent vs. less violent) as the dependent variable. Predictor variables were: major neurological signs, minor neurological signs, paranoid symptomatology, visual hallucinations, auditory hallucinations, olfactory/gustatory halluci-
317
Violent Juvenile Delinquents
nations, depressive symptoms, memory for 4 numbers backward, verbal IQ, performance IQ, having been abused, reading grade discrepancy, and math grade discrepancy. It was found that two variables-paranoid symptomatology and minor neurological signs -produced an R of 0.633 and accounted for 40% of the total variance. Paranoid symptoms accounted for 29.3% of the total variance, and minor neurological signs accounted for the other 10.7% of the variance. DISCUSSION
Our data strongly suggest that extremely violent juvenile offenders differed significantly in a number of neurological and psychiatric ways from their less violent delinquent peers. They were significantly more likely to manifest both major and minor signs of neurological impairment and to suffer from pervasive paranoid ideation. They were also more likely to have suffered abuse. Their paranoid orientation, combined with their neurological dysfunction. seemed to contribute to their tendency to retaliate quickly and sometimes brutally for either real or imagined threats. They rarel y if ever perceived themselves as provocateurs, but rather interpreted almost every interpersonal encounter as a time of threat in which they had to protect themselves. Further clinical evidence of the possible contribution of neurological vulnerability to their violent behaviors was the observation that the violent child, once started, often seemed totally unable to curtail his violence. There were man y children who could remember well certain of their violent behaviors, but at times were either totally unable to recall an act of which they were accused , or recalled it in a semicoherent, distorted manner. In man y of these episodes, it seemed unlikely that the child's memory lapse was a reflection of deliberate lying or of the operation of a variety of defense mechanisms since the child often was extremely c o n fu se d and described a ct s which were far
worse than those he had performed. These children frequently insisted that they could recall everything they had done perfectly, when other witnesses to the same violent event described quite different behaviors. The severe verbal deficiencies of the more violent group may have co n tr ibuted further to their aggressive behaviors. Perhaps these inordinately suspicious, neurologically impaired, and impulsive ch ild ren had particular difficulty putting aggressive thoughts
318
Dorothy Otnow Lewis et al.
into words rather than actions. The contribution to violence of having experienced and witnessed physical abuse is complex. First, physical abuse often causes central nervous system damage, thus contributing to impulsivity, attention disorders, and learning disabilities. Second, it provides a model with which to identify. Finally, it engenders rage toward the abusing parent, rage that can then be displaced onto authority figures and other individuals, against whom the child may vent his anger. IMPLICATIONS
Especially violent juvenile delinquents were found to have paranoid symptoms and a multiplicity of major and minor neurological abnormalities. The most striking social factors distinguishing the more violent from the less violent group were not related to family constellation but rather were factors related to abuse. Violent delinquents, in contrast to less violent delinquents, had been severely abused and had witnessed extreme violence in their homes. The identification of the neuropsychiatric and emotional prob-
lems of these children was a subtle and time-consuming process. Furthermore, the kinds of problems uncovered were often chronic in nature, requiring ongoing support and treatment. Because of the need for lengthy diagnosis and prolonged treatment, violent juveniles are likely to be dismissed merely as incorrigible sociopaths and simply incarcerated. Our findings suggest that enlightened psychological, educational, and medical programs can and should be devised to meet the needs of these multiply damaged children. Programs designed to diminish violence which focus primarily on socioeconomic and psychological factors are likely to be unsuccessful if they ignore the medical problems (e.g., psychotic symptoms, neurological impairment) that contribute so strongly to the expression of violence. The role of specific medications as part of the treatment of violent adolescents (e.g., anticonvulsants, stimulants, anti psychotics, antidepressants) is an important area that has yet to be explored. REFERENCES BACH-Y-RITA, G. & VEND, A. (1974). Habitual violence. Am". j. Prychiat.• 131: 1015-1017. - - LION, J. R., CLIMENT, C. E., & ERVIN, F. R. (1971), Episodic dyscontrol. Am". j. Psychiat., 127: 1473-1478.
Violent Juvenile Delinquents
319
CARPER, .J. (1974), Medical care of delinquent adolescent boys. Pediat. Clin. N. Amer., 21 :423-433. DURBIN, J. R., PASEWARK, M. S., & ALBERS, D. (1976), Criminality and mental illness. Amer. I Psyrhiat., 134:80-83. EIU:NBERG, M. D. (1961), Remand home boys 1930-1955. Brit. I Crimin., 2: 111-131. GIBBENS, T. N. C. (1963), The effects of physical ill-health in adolescent delinquents. Proc. Roy. Soc. Med., 56: 1086-1088. GLASER, G. H. & DIXON, M. S. (1956), Psychomotor seizures in childhood. Neurology, 6:646-.Q55. GLUECK, S. & GLUECK, E. (1950) Unraveling Juvenile Delinquency. New York: Commonwealth Fund. Guzr., S. B., GOODWIN, D. W., & CRANt:, J. B. (1969), Criminality and psychiatric disorders.
Arch. Gen. Psychiat., 20:583-591. HEALY, W. & BRONNER, A. F. (1926), Delinquents and Criminals. New York: MacMillan. KLOEK, J. (1968), Schizophrenia and delinquency. Int. Psychiat. Clin., 5: 19-34. LEWIS, D. O. (l976a), Delinquency, psychomotor epileptic symptoms, and paranoid ideation. Amer. I Psychtat., 133: 1395-1398. - - - (1976b), Diagnostic evaluation of the juvenile offender. Child Psychiat. Hum. Deuelpm., 6: 198-213. - - - & BALLA, D. A. (1976), Delinquency and Psychopathology. New York: Grune & Stratton. - - - & SHANOK, S. S. (1976), Medical histories of delinquent and nondelinquent children.
Amer.I Psychial., 134:1020-1025. MCCORD, W. & MCCORD, J. (1959), Origins of Crime. New York: Columbia University Press. MARK, V. H. & ERVIN, F. R. (1970), Violence and the Brain. New York: Harper & Row. NAMECIIE, G., WARING, M., & RICKS, D. (1964), Early indicators of outcome in schizophrenia.
I Neru. Ment. Dis., 139:232-240. OUNSTt:I), C. (1969), Aggression and epilepsy. I Psychosom. Res., 13:237-242. PETRICH, J. (1976), Rate of psychiatric morbidity in a metropolitan county jail population.
Amer.I Psychiat., 133:1439-1444. PINCUS,.J. H. & TUCKt:R, G. (1978), Behavioral Neurology. New York: Oxford University Press. ROBINS, 1.. (1966), Deviant Children Grown Up. Baltimore: Williams & Wilkins. SCOTT, P. D. (1975), Medical aspects of delinquency. In: Contemporary Psychiatry, ed. T. Silverstone & B. Barraclough. Ashford, Kent: Headley Brothers, pp. 287-295. SERAFETINJDES, E. A. (1965), Aggressiveness in temporal lobe epileptics and its relation to cerebral dysfunction and environmental factors. Epilepsia, 6:33-42. SHANOK, S. S. & LEWIS, D. O. (1976), Juvenile court versus child guidance referral. Amer. I Psychiat., 134:1130-1133. SHOOR, M, & SPEED, M. H. (1969), Seven years of psychiatric consultation in a juvenile probation department. Psychiat. Quart., 43: 147-163. STEPHENSON, P. S., BLAKt:LY, B., & NICHOL, H. (1973), The psychiatric status and treatment needs of a random sample of juveniles charged with delinquency. Psychiat. Clin., 6:257270. STOTT, D. II. (1962), Evidence for a congenital factor in maladjustment and delinquency.
Amer. I Psychiat., 118:781-794, TRHFERT, D. A. (1964), The psychiatric patient with an EEG temporal lobe focus. Amer. I Psychiat., 120:765-771. WARREN, W. (1960), Some relationship between the psychiatry of children and adults. I
Ment. Sci., 106:815-826. ZITRIN, A., HARDESTY, A. S., BURDOCK, E. I., & DROSSMAN, A. K. (1976), Crime and violence among mental patients. Amer. I Psychiat., 133: 142-149.
Dorothy Otnow Lewis, M.D., F.A.C.P., Shelley S. Shanok, M.P.H., Jonathan H. Pincus, M.D., and Gilbert H . Glaser, M.D. Abstract , Th is stud y compares the neu ropsychiatric. intellectual, a nd educational status of
ex tr emely violent and less violent inca rcerated boys. The more violent ch ild re n were more likel y to demonstrate psychotic symptomatology (paranoid ideation and rambling. illogical associa tio ns) and to have major and minor neurological abnorm aliti es . They were also more likely to have experienced and witn essed e xtre me physical abuse. The co nt r ibu tio n of these factors to violent delinquency is dis cu ssed a nd implications for therapeutic intervention are
suggested.
The purpose of this paper is to present the results of a comparison study of a sample of extremely violent delinquent boys and a sample of less aggressive, male juvenile offenders. This study sprang from earlier findings that man y delinquent children referred to a juvenile court clinic suffered from hitherto unrecognized, undiagnosed, and untreated psychiatric, neurological, and educational disorders which seemed to contribute to their antisocial behaviors (Lewis, 1976a; Lewis and Balla, 1976; Lewis and Shanok, 1976). This raised the question : are there any associations between the severity of a child's antisocial behaviors and his psychiatric, neurological, and psychoeducational status? There is disagreement in the literature whether psychiatric and neurological disorders are more prevalent in the delinquent and
Dr. Leuns is Associate Clinical Pr ofessor of Psychiatry at the Yair University Child Study Center; M s. Shanok is Associate in Re search, Yair University Child Study Center; Dr. Pincus is Professor of Nru rology at Yale University School of Medicine; and Dr. Glaspy is Professor and Chainnan of Neu rology at Yair University School of Medicine. R eprints may br requested from Dr. Le uns at the Yair University Child Study Center, 333 Cedar Stre et, N ew Haven, CT 06510. This research was supported by La w Enforcement A ssistancp Administration grant;; 76AE418003X and Ford Foundation grant;; 730A-41-57478. 0002-7 I 38179/1802--{)307 $01.16 e Ame rican Academ y of Child Psych iatry.
307
308
Dorothy Otnow Lewis et al.
criminal population than in the general population (Durbin et aI., 1976; Guze et aI., 1969; Healy and Bronner, 1926; Kloek, 1968; Nameche et aI., 1964; Petrich, 1976; Robins, 1966; Shanok and Lewis, 1976; Shoor and Speed, 1969; Stephenson et aI., 1973; Warren, 1960; Zitrin et aI., 1976). Similar disagreement exists in the literature concerning the medical histories of delinquents (Carper, 1974; Eilenberg, 1961; Gibbens, 1963; Glueck and Glueck, 1950; McCord and McCord, 1959; Scott, 1975; Stott, 1962). There is, however, a growing body of evidence that certain forms of violence are associated with disorders of the central nervous system (Bach-y-rita and Veno, 1974; Bach-y-rita et aI., 1971; Glaser and Dixon, 1956; Lewis, 1976a, 1976b; Mark and Ervin, 1970; Ounsted, 1969; Pincus and Tucker, 1978; Serafetinides, 1965; Treffert, 1964). Clearly, there was a need for a systematic study of the neuropsychiatric status of violent delinquent children. METHOD
Setting
Because the study was primarily concerned with serious offenders, the only correctional school serving the entire state of Connecticut was chosen as the setting. Previous work, emanating from a juvenile court setting, focused on the less serious juvenile delinquents from only one district of the state. The correctional school setting allowed the evaluation of some of the most seriously delinquent children throughout the entire state. Furthermore, within the correctional school, a secure unit had recently been constructed to house especially violent juveniles. Thus the possibility existed of comparing more and less violent children at the school itself. Samples
The study samples consisted of 97 boys, all of whom were incarcerated at the correctional school at some time during the 18-month duration of the project and were evaluated by the clinical team. For research purposes, after all data had been collected on the 97 boys, three of the investigators independently rated each child on a scale of violence from 1 (least violent) to 4 (most violent), using only data regarding behaviors and offenses, without taking into account psychiatric or neurological data. Raters attempted to use the follow-
Violent Juvenile Delinquents
309
ing broad criteria for group placement: A child was to be rated 1 if there were no evidence of his having committed any offense against person or having committed arson. Since almost all of the children at the school had, at one time or another, been involved in fist fights with peers, such behaviors were not considered violent unless a peer had been seriously injured and required hospital care, or if a weapon had been used. There were 8 in this group. Children were rated 2 if there was some indication in their charts of a potential for violence (e.g., isolated episodes of firesetting, isolated episodes of threatening with unloaded weapons, isolated episodes of threatening violence). There were 11 boys in this group. Children were rated 3 if they had actually committed serious offenses against person (e.g., murder, rape, multiple episodes of arson, armed robbery, assault). All of these children had committed more than one offense that endangered others. There were 55 in this group. Children who were considered to have demonstrated extraordinary brutality toward others were rated 4. For example, in this group was a boy who assaulted a woman, stomping on her face and leaving imprints of his shoes; a boy who raped and beat a young girl and subsequently raped and stabbed another; and a boy who shot a gun in the direction of a group of women, and who secreted knives on himself and around his room. There were 23 boys in this group. All ratings of violence were subjective. In almost all cases, there was agreement among raters. When disagreements occurred, a child was rated according to the assessments of two of the raters. DIAGNOSTIC EVALUATION
Psychiatric Assessment
Children in the study were evaluated by a child psychiatrist and a neurologist. The psychiatric evaluation has already been described (Lewis, 1976b; Lewis and Balla, 1976). However, the criteria for the presence or absence of particular psychiatric symptomatology will be reviewed briefly. Auditory or visual hallucinations or paranoid delusions were considered to be significant if children reported such experiences
310
Dorothy Otnow Lewis et al.
and acted in response to them or were clearly troubled by them. For instance, many children reported hearing their mothers calling their names. This was not considered abnormal unless it was associated with panic, an attempt to leave the room in which the voice was heard, or search for the source of the voice, such as looking under the pillow or bed, or out in the hallway or courtyard. Depression was considered to be significant if the children felt sad, said they had crying spells, and reported vegetative symptoms. Sadness alone was not considered to be an adequate basis for diagnosing depression. In the course of the psychiatric and neurological evaluations, the presence or absence of paranoid ideation was assessed. It took much experience to ascertain, as far as possible, the difference between appropriate, adaptive wariness and inordinate suspiciousness. Paranoid symptomatology was felt definitely to be present if children had mistakenly believed that someone was going to hurt them and could provide several examples of this, or if they admitted to constantly feeling the need to carry weapons such as guns and metal pipes for their own protection in the absence of identifiable dangers. Unusual suspiciousness manifested during the interview was not considered an indication of paranoid symptomatology unless it was associated with the kinds of pervasive delusional beliefs and behaviors mentioned above. Neurological Assessment
Standard neurological examinations were carried out. These included measurement of head circumference; evaluation of cranial nerves; motor, sensory, and reflex functions. Tests of coordination included quantification of the number of alternating palm strikes the child could perform during pronation-supination movements in 10 seconds and the number of finger taps the child could perform with his index finger in 10 seconds with each hand. The presence of choreiform movements was determined by asking the child to extend his arms and fingers in front of him and above his head for 5 seconds each. All were asked to skip after the examiner demonstrated the requested pattern of movement. Mental status examination in the course of the neurological evaluation included orientation for time and place. The ability to remember up to 6 numbers forward and 4 numbers backward was tested. Calculation skills were tested by 4 serial subtractions of 7, starting with 100.
Violent Juvenile Delinquents
311
Certain issues were covered both by the neurologist and the psychiatrist. Both attempted to obtain as detailed a medical history as possible. Both tried to ascertain whether or not the child had been a victim of abuse or had witnessed extreme violence. A child was considered to have been abused by his parents or guardians if he had been punched; beaten with a stick, board, pipe, or belt buckle; or beaten with a belt or switch other than on the buttocks. He was also considered to have been abused if he had been deliberately cut, burned, or thrown downstairs or across a room . A child was considered not to have been abused if he was only struck with an open hand, beaten with the leather part of a belt, or beaten with a switch on the buttocks only . Whenever possible, a sleep electroencephalogram was performed. Results of previous electroencephalograms were also obtained whenever possible. Evidence of major neurological abnormality was defined as a documented history of grand mal epilepsy, an abnormal electroencephalogram, positive Babinski sign, or head circumference that was plus or minus two standard deviation from the mean for the child's age . Evidence of minor neurological abnormality was defined as inability to skip, choreiform movements, abnormal reflexes, abnormalities in coordination, or mixed dominance of hands and feet. Subjects were considered to have ps ychomotor epileptic symptomatology if at least two of the following were present: observed staring episodes with loss of fully conscious contact with reality; loss of memory for violent o r nonviolent acts, followed by fatigue; episodes of inability to understand conversations despite an attempt to do so ; dizzy spells followed by fatigue or headache ; episodes of unprovoked extreme anxiety; olfactory or gustatory hallucinations; many recurrent episodes of deja vu; macropsia or micropsia; forced thinking; and automatic repetitive behavior. Psychoeducational Testing
Psychological testing consisting of the Wechsler Intelligence Scale for Children (revised), the Bender-Gestalt Test, and the Rorschach Test was performed by a research developmental child psychologist. When time factors precluded testing by our own psychologist, or when recent test results were available, testing was not repeated, and previously charted test results were used for purposes of obtaining scores on tests and subtests. Educational evaluations were
Dorothy Otnow Lewis et al.
312
performed by a learning disabilities specialist. Reading grade discrepancies were calculated by subtracting the child's current reading grade score from his actual grade placement in school. Possible bias was introduced in the study because in most cases the investigators knew whether or not a subject was considered violent by virtue of the fact that most violent subjects were seen on the secure unit, while most of the less violent subjects lived in open settings and were seen at the school infirmary. Ethically it would have been impossible to have nonviolent children randomly placed on a securely locked unit. Thus it is possible that the preconceived notions of the examiners influenced the quality of the data collected. Our previously described rating process, based on behaviors only, without access to other data, was an attempt to correct for possible bias. FINDINGS
Comparison of Nonviolent Subjects with Violent Subjects
Of the 97 children evaluated. only 8 were considered by all raters to be nonviolent. When the psychiatric, neurological, and psychoeducational status of these children was compared with the violent subjects, a number of significant differences was found. Psychiatrically, more of the violent children suffered from paranoid symptoms (75.5% vs. 0%, X~ = 15.386, P < .001); were loose, rambling and illogical (57.1 % vs. 0%, X; = 7.000, P = .009); and were unable to remember 4 digits backward (55.2% vs. 12.5%, X; = 3.556, P = .060). Although differences did not reach conventional levels of significance in the following factors, none of the nonviolent children had experienced auditory hallucinations, compared with 40.8% of the violent boys; none had experienced visual hallucinations, compared with 28.2% of the violent boys; and none had experienced olfactory or gustatory hallucinations, compared with 12.9% of the violent group. Similarly, the nonviolent group was more neurologically intact than the violent group. For example, none of the nonviolent sample had major neurological signs, compared with 41.6% of the violent group. Of the violent group, 58.7% had choreiform movements compared with 14.3% of the nonviolent group, and 38% of the violent group were unable to skip, compared with 14.3% of the nonviolent group. Because of the very small number of subjects in the nonviolent group, however, these neurological differences did
313
Violent Juvenile Delinquents
not reach conventional levels of significance. Statistically significant differences between the nonviolent group and the rest of the sample were found in the proportion of each group with psychomotor epileptic symptomatology-25.0% (2) nonviolent vs. 69.4% (50) violent, X~ = 4.451, P = .035-and in the proportion of children with a greater than 10% discrepancy between right and left palm strikes-14.3% (1) nonviolent vs. 60.8% (45) violent, X~ = 3.904, P = .049. Intellectually there were no significant differences between the groups in terms of overall IQ as measured by the WISe or on verbal or performance scores, although there was a tendency for the violent group to score lower in the verbal area (91.500 nonviolent vs. 83.158 violent, t = 1.726, P = .088). We wondered whether neuropsychiatric differences between the more and the less violent children existed. In order to investigate this question, the 97 subjects were divided into two groups, those children rated 1 or 2 constituting the less violent group, and those rated 3 or 4 constituting the more violent gr,Jup.
Comparison of More and Less Violent Delinquents Psychiatric Findings. These are summarized in table 1. The most striking difference psychiatrically between the two groups was the finding that a significantly greater proportion of very violent children demonstrated or gave clear histories of paranoid symptomatolTable I A Comparison of the Xumbcr and Percentage of :\fore Violent and Less Violent Delinquent Boys with Specific Psych ia t ric Symptomatology More \'iolent Depressive symptoms Visual hall urinations Auditory hallucinations Olfactory or gustatory hallucinations Tactile hallucinations Paranoid symptomatologv Loose. rambling, illogical Inability to subtract serial 7s Inability to remember 6 digits forward Inability to remember 4 digits backward
X" • r
p value
84.fi 11.1 17.6
0.8.-,2 1.695 2.704
.sss
2.126 0.65!J 24.618 5.12fi 6.l~8
.145 .417 < .001 .024 .014 >.5
Less Violent
#
(f'lI.
,e,
#
2~
fi:;.7
1\
18 28
~O.O
2
43.3
3
II
15.1
0 0
,,,n
(I'
(j
8
5·\ 41
81.8 59.fi 69.5
4 6
0 0 16.7 23.5 33.:{
l-1
26.9
13
18.8
0.109
31
60.8
2
13.3
8.627
2H
OJ
s
.193
.I 01
.004
• Percentages arc based upon the actual number of subjects on whom information was available for each category.
314
Dorothy Otnow Lewis et al.
ogy. Furthermore, they were significantly more likely than their less aggressive peers to be loose, rambling, and illogical in their thought processes during interviewing. In the formal mental status evaluations, they had extreme difficulty remembering even 4 digits backward, and the majority was unable to subtract serial 7s, even when given the opportunity to practice the task. Depressive symptoms were common in both samples. We found that although there was a definite tendency for more of the more violent children to have experienced hallucinations, particularly auditory hallucinations, differences between the two groups in this regard did not reach conventional levels of significance. Neurological Findings. Neurological differences between the very violent and less violent samples of delinquents are summarized in table 2. Almost the entire sample of more violent children (98.6%) had one or more signs of minor neurological impairment, which was significantly different from the less violent sample, 66.7% of whom evidenced minor neurological abnormalities. Table 2 Specific Neurological Signs: Comparison of More Violent and Less Violent Boys More Violent One or more major neurological signs Abnormal EH; Positive Babinski One or more minor neurological si~ns
Inahilirv to skip Chord form movcmcut-, Psychomotor symptoma tolog v Greater than 10";, di"'H'pancy he· tween right and Idt palm strike (~rl'atlT than lUI;, d iscrcpancv hetween right and left finger taps
#
(r'_
"
Less Violent
X~
#
(I'a
I
li.7 0 'i.fi
1I.'i6!l
;-.
(ifi.7 11.1 :11.1
J(i.:!i:'> ·1.!l:!li 3.37'i
p value
,I'
:\1 I!I II
·lfi.:! :!!1.7 l'i."
71
I:!
·11i
!':-Ui 1:1.:1 (iO.ti 71.!1
Ii
:17.:;
;).22:\
.001 .0:!7 .0(i7 .023
10
lil.-,
Ii
:17.:;
2.1~:;
.11;
2ti
I I.!!
.;0.0
O.OO:!
~(i
·10
(I
I
'J
1i.·I!'!l
z.sso
.011 .I0!! >.'i
> ..;
• Percentages arc based upon the actual numher of children for whom data were available for each category.
The presence of major signs of neurological abnormality significantly distinguished the two groups. Of note, almost 30% of the very violent children had grossly abnormal electroencephalograms, usually of a paroxysmal or focal nature, and/or a history of grand mal epilepsy, compared with none of the less violent sample. Psychological Test Results. Differences in performance on psycho-
Violent Juvenile Delinquents
315
logical testing between the two groups are summarized in table 3. Intelligence, as measured on the WISe, indicated a tendency for the less violent children to function somewhat better intellectually, but overall differences between the two groups were not striking. The more violent group did less well on the verbal section of the WISe, the greatest difference between groups occurring on the arithmetic subtest. Both groups scored relatively well on the picture completion and picture arrangement subtests, and both groups received their poorest scores on the vocabulary subtest. They both were several years behind their expected reading grade. Differences between math grade discrepancies did not reach conventional levels of significance. Table 3 A Comparison of Psvchoeduc.u ional Test Results of :\Iore Violent and Less Yiokllt Dcl iuqucnt Children
Average full-scale IQ WISC WISC \erbal score wise performance score ( .omprclu-nsion subrcst Arithmetic suhtcst Similarities suhtcst Vocabulary subu-st Picture completion subtest Pict me a ITaugcmcn t Block design Object assembly Coding Reading grade discrepancy Math grade discrepancy
More Violent
Less Violent
RfiA",!! R:!.5!!7 9:!.H7:; 7.fiF, (j.!J81 8.1 !;; 6.3:!1 10.1i:!3
93.000 RR.RI:! 98.87:; 8.800 !!.!!O!! 8.000 7.400 10.:;1:; lOy,() 9.818
9.922
8.2:;0 9.694 7.60·1
4..1 3..1
ro.ssr
8.58;1 2.300 2.200
t-test 1.94;1 1.978 1.507 1.41;1 3.!JO:! 0.11", 1.395 0.034 0.91:; 1.478 0.580 0.934 2.148 1.602
p value .O:;fi .O:;:! .130 .163 < .001 > .500 .169 > .500 .364 .145 > .500 .354 .037 .115
History of Abuse. Another set of factors that strongly distinguished the more violent from the less violent children related to a history of abuse hy parents or parent suhstitutes. Differences are summarized in table 4. The more violent children had been physically abused by mothers, fathers, stepparents, other relatives, and "friends" of the family. The degree of abuse to which they were subjected was often extraordinary. One parent broke her son's legs with a broom; another broke his fingers and his sister's arm; another chained and burned his son; and yet another threw his son downstairs, injuring his head, following which the boy developed epilepsy. The two samples also differed significantly in their exposure to
Dorothy Otnow Lewis et al.
316 Table 4
Comparison of the Number and Percentage of More \ 'iolent and Less Violent Delinquents with a History of Abuse More Violent
# Abused by mother Abused by father Abused by others Ever abused \\'itness to extreme violence
21 29 23
52 44
pt .
10
43.8
54.7 45.1 75.4 78.6
.
Less Violent
#
"/
2 5 2 6 3
14.3 29.4 14.3 33.3 20.0
, 0
X} 2.869 2.364 3.200 !I.535 15.615
p value .091 .125 .074 .003 <.001
• Percentages are based upon the actual numbers of subjects on whom information was a"a ila ble for each ca tegory.
violence. The fact that 78.6% of the more violent children were known to have witnessed extreme violence directed at others, mostly in their homes, compared with 20.0% of the less violent children, tells only part of the story. The degree of violence witnessed by these children went beyond mere fist fights. Several children witnessed their fathers, stepfathers, or mothers' boyfriends slash their mothers with knives, They saw their siblings tortured wiLh cigaJ'cLlc butts , chained to beds, and thrown into walls. They saw their relatives-male and female-arm themselves with guns, knives, and other sharp instruments and, at times, use these weapons against each other. Some children ran away from home at the approach of certain relatives, while many children reported defending their mothers with pipes and sticks while their mothers were being attacked. We wondered whether degree of violence as rated 1-4 correlated in any way with having been abused. A correlation coefficient of r = 0 .373, P < .00 I was found, indicating that a child's degree of violence was strongly associated with his having been abused. Of note, the family constellation of the more and less violent delinquents (e.g., broken homes, mother in home, father in home) was similar in both groups. It was, rather, the quality of family interactions that distinguished the two groups from each other. Multiple Regression Analysis. In an effort to determine which combination of factors best distinguished the more violent from the less violent delinquents, we conducted a multiple regression analysis, using group membership (very violent vs. less violent) as the dependent variable. Predictor variables were: major neurological signs, minor neurological signs, paranoid symptomatology, visual hallucinations, auditory hallucinations, olfactory/gustatory halluci-
317
Violent Juvenile Delinquents
nations, depressive symptoms, memory for 4 numbers backward, verbal IQ, performance IQ, having been abused, reading grade discrepancy, and math grade discrepancy. It was found that two variables-paranoid symptomatology and minor neurological signs -produced an R of 0.633 and accounted for 40% of the total variance. Paranoid symptoms accounted for 29.3% of the total variance, and minor neurological signs accounted for the other 10.7% of the variance. DISCUSSION
Our data strongly suggest that extremely violent juvenile offenders differed significantly in a number of neurological and psychiatric ways from their less violent delinquent peers. They were significantly more likely to manifest both major and minor signs of neurological impairment and to suffer from pervasive paranoid ideation. They were also more likely to have suffered abuse. Their paranoid orientation, combined with their neurological dysfunction. seemed to contribute to their tendency to retaliate quickly and sometimes brutally for either real or imagined threats. They rarel y if ever perceived themselves as provocateurs, but rather interpreted almost every interpersonal encounter as a time of threat in which they had to protect themselves. Further clinical evidence of the possible contribution of neurological vulnerability to their violent behaviors was the observation that the violent child, once started, often seemed totally unable to curtail his violence. There were man y children who could remember well certain of their violent behaviors, but at times were either totally unable to recall an act of which they were accused , or recalled it in a semicoherent, distorted manner. In man y of these episodes, it seemed unlikely that the child's memory lapse was a reflection of deliberate lying or of the operation of a variety of defense mechanisms since the child often was extremely c o n fu se d and described a ct s which were far
worse than those he had performed. These children frequently insisted that they could recall everything they had done perfectly, when other witnesses to the same violent event described quite different behaviors. The severe verbal deficiencies of the more violent group may have co n tr ibuted further to their aggressive behaviors. Perhaps these inordinately suspicious, neurologically impaired, and impulsive ch ild ren had particular difficulty putting aggressive thoughts
318
Dorothy Otnow Lewis et al.
into words rather than actions. The contribution to violence of having experienced and witnessed physical abuse is complex. First, physical abuse often causes central nervous system damage, thus contributing to impulsivity, attention disorders, and learning disabilities. Second, it provides a model with which to identify. Finally, it engenders rage toward the abusing parent, rage that can then be displaced onto authority figures and other individuals, against whom the child may vent his anger. IMPLICATIONS
Especially violent juvenile delinquents were found to have paranoid symptoms and a multiplicity of major and minor neurological abnormalities. The most striking social factors distinguishing the more violent from the less violent group were not related to family constellation but rather were factors related to abuse. Violent delinquents, in contrast to less violent delinquents, had been severely abused and had witnessed extreme violence in their homes. The identification of the neuropsychiatric and emotional prob-
lems of these children was a subtle and time-consuming process. Furthermore, the kinds of problems uncovered were often chronic in nature, requiring ongoing support and treatment. Because of the need for lengthy diagnosis and prolonged treatment, violent juveniles are likely to be dismissed merely as incorrigible sociopaths and simply incarcerated. Our findings suggest that enlightened psychological, educational, and medical programs can and should be devised to meet the needs of these multiply damaged children. Programs designed to diminish violence which focus primarily on socioeconomic and psychological factors are likely to be unsuccessful if they ignore the medical problems (e.g., psychotic symptoms, neurological impairment) that contribute so strongly to the expression of violence. The role of specific medications as part of the treatment of violent adolescents (e.g., anticonvulsants, stimulants, anti psychotics, antidepressants) is an important area that has yet to be explored. REFERENCES BACH-Y-RITA, G. & VEND, A. (1974). Habitual violence. Am". j. Prychiat.• 131: 1015-1017. - - LION, J. R., CLIMENT, C. E., & ERVIN, F. R. (1971), Episodic dyscontrol. Am". j. Psychiat., 127: 1473-1478.
Violent Juvenile Delinquents
319
CARPER, .J. (1974), Medical care of delinquent adolescent boys. Pediat. Clin. N. Amer., 21 :423-433. DURBIN, J. R., PASEWARK, M. S., & ALBERS, D. (1976), Criminality and mental illness. Amer. I Psyrhiat., 134:80-83. EIU:NBERG, M. D. (1961), Remand home boys 1930-1955. Brit. I Crimin., 2: 111-131. GIBBENS, T. N. C. (1963), The effects of physical ill-health in adolescent delinquents. Proc. Roy. Soc. Med., 56: 1086-1088. GLASER, G. H. & DIXON, M. S. (1956), Psychomotor seizures in childhood. Neurology, 6:646-.Q55. GLUECK, S. & GLUECK, E. (1950) Unraveling Juvenile Delinquency. New York: Commonwealth Fund. Guzr., S. B., GOODWIN, D. W., & CRANt:, J. B. (1969), Criminality and psychiatric disorders.
Arch. Gen. Psychiat., 20:583-591. HEALY, W. & BRONNER, A. F. (1926), Delinquents and Criminals. New York: MacMillan. KLOEK, J. (1968), Schizophrenia and delinquency. Int. Psychiat. Clin., 5: 19-34. LEWIS, D. O. (l976a), Delinquency, psychomotor epileptic symptoms, and paranoid ideation. Amer. I Psychtat., 133: 1395-1398. - - - (1976b), Diagnostic evaluation of the juvenile offender. Child Psychiat. Hum. Deuelpm., 6: 198-213. - - - & BALLA, D. A. (1976), Delinquency and Psychopathology. New York: Grune & Stratton. - - - & SHANOK, S. S. (1976), Medical histories of delinquent and nondelinquent children.
Amer.I Psychial., 134:1020-1025. MCCORD, W. & MCCORD, J. (1959), Origins of Crime. New York: Columbia University Press. MARK, V. H. & ERVIN, F. R. (1970), Violence and the Brain. New York: Harper & Row. NAMECIIE, G., WARING, M., & RICKS, D. (1964), Early indicators of outcome in schizophrenia.
I Neru. Ment. Dis., 139:232-240. OUNSTt:I), C. (1969), Aggression and epilepsy. I Psychosom. Res., 13:237-242. PETRICH, J. (1976), Rate of psychiatric morbidity in a metropolitan county jail population.
Amer.I Psychiat., 133:1439-1444. PINCUS,.J. H. & TUCKt:R, G. (1978), Behavioral Neurology. New York: Oxford University Press. ROBINS, 1.. (1966), Deviant Children Grown Up. Baltimore: Williams & Wilkins. SCOTT, P. D. (1975), Medical aspects of delinquency. In: Contemporary Psychiatry, ed. T. Silverstone & B. Barraclough. Ashford, Kent: Headley Brothers, pp. 287-295. SERAFETINJDES, E. A. (1965), Aggressiveness in temporal lobe epileptics and its relation to cerebral dysfunction and environmental factors. Epilepsia, 6:33-42. SHANOK, S. S. & LEWIS, D. O. (1976), Juvenile court versus child guidance referral. Amer. I Psychiat., 134:1130-1133. SHOOR, M, & SPEED, M. H. (1969), Seven years of psychiatric consultation in a juvenile probation department. Psychiat. Quart., 43: 147-163. STEPHENSON, P. S., BLAKt:LY, B., & NICHOL, H. (1973), The psychiatric status and treatment needs of a random sample of juveniles charged with delinquency. Psychiat. Clin., 6:257270. STOTT, D. II. (1962), Evidence for a congenital factor in maladjustment and delinquency.
Amer. I Psychiat., 118:781-794, TRHFERT, D. A. (1964), The psychiatric patient with an EEG temporal lobe focus. Amer. I Psychiat., 120:765-771. WARREN, W. (1960), Some relationship between the psychiatry of children and adults. I
Ment. Sci., 106:815-826. ZITRIN, A., HARDESTY, A. S., BURDOCK, E. I., & DROSSMAN, A. K. (1976), Crime and violence among mental patients. Amer. I Psychiat., 133: 142-149.