Close
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Biopsychosocial and forensic clinical correlates of schizophrenia and homicide
(USC Thesis Other)
Biopsychosocial and forensic clinical correlates of schizophrenia and homicide
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
BIOPSYCHOSOCIAL AND FORENSIC CLINICAL CORRELATES OF
SCHIZOPHRENIA AND HOMICIDE
by
Robert August Schug
_________________________________________________________________
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
August 2009
Copyright 2009 Robert August Schug
ii
Table of Contents
List of Tables iii
List of Figures iv
Abstract v
Chapter 1: Background and Significance 1
Chapter 2: Method 25
Participants 25
Screening for CCMD-3 and DSM-IV Disorders 27
Neuropsychological Measures 27
EEG Recording Procedures 29
Psychosocial Functioning 30
Childhood Psychosocial Deprivation 31
Forensic Assessment Data 32
Statistical Analyses: Traditional and Modern Methods 33
Chapter 3: Results 35
Substance Use Disorders, Antipsychotics, and Head Injury 35
Neuropsychology 38
EEG 45
Psychosocial Functioning 58
Childhood Psychosocial Deprivation 66
Forensic Clinical Correlates 67
Chapter 4: Discussion 76
Hypothesis 1: Intellectual Functioning 76
Hypothesis 2: Executive Functioning 78
Hypothesis 3: EEG 83
Hypothesis 4: Psychosocial Functioning and Childhood Psychosocial 94
Deprivation
Hypothesis 5: Forensic Clinical Correlates 109
A Biopsychosocial Interactional Pathway to Schizophrenic Violence 115
Schizophrenic Murderers: Distinct Subgroup or a “Perfect Storm”? 116
Homicide Psychiatric Controls 118
Limitations and Strengths of the Present Study 120
Implications 123
Conclusions 125
Bibliography 127
iii
List of Tables
Table 1: Diagnostic Group Demographic Characteristics 26
Table 2: Antipsychotic Medications 37
Table 3: IQ 39
Table 4: Color Trails Test Variable T Scores 41
Table 5: Gambling Task Results 43
Table 6: Resting EEG Amplitudes at Individual Electrode Sites: 46
Delta Band (0.01- 4.00 Hz)
Table 7: Resting EEG Amplitudes at Individual Electrode Sites: 47
Theta Band (4.01- 8.00 Hz)
Table 8: Resting EEG Amplitudes at Individual Electrode Sites: 48
Alpha1 Band (8.01- 10.50 Hz)
Table 9: Resting EEG Amplitudes at Individual Electrode Sites: 49
Alpha2 Band (10.51- 13.00 Hz)
Table 10: Resting EEG Amplitudes at Individual Electrode Sites: 50
Beta1 Band (13.01- 20.00 Hz)
Table 11: Resting EEG Amplitudes at Individual Electrode Sites: 51
Beta2 Band (20.01- 30.00 Hz)
Table 12: Resting EEG Amplitudes at Individual Electrode Sites: 52
Beta3 Band (30.01- 4.00 Hz)
Table 13: Sample-Wide Resting EEG Amplitudes: Antipsychotic-Free 59
Versus Antipsychotic-Medicated Participants
Table 14: Sample-Wide Resting EEG Amplitudes: Non-Head-Injured 60
Versus Head-Injured Participants
Table 15: Psychosocial Correlates 63
Table 16: Offense Characteristics and Concomitants 68
Table 17: Victimology 70
iv
List of Figures
Figure 1: Gambling Task Results 44
Figure 2: Offense-Related Emotional Experience 72
v
Abstract
The present study sought to examine a subtype of homicidal schizophrenic
individuals, and to elucidate biological and psychosocial differences between these
schizophrenic murderers and their non-schizophrenic and non-murdering counterparts.
Participants were recruited from a medical university brain hospital in Nanjing, China,
and grouped according to psychiatric diagnosis and homicide conviction: normal
controls (n = 47), schizophrenic non-murderers (n = 33), non-schizophrenic murderers
(n = 31), schizophrenic murderers (n = 32), and homicide psychiatric controls (n = 14).
All were assessed in areas of intellectual and executive functioning, resting EEG, current
psychosocial functioning and childhood psychosocial deprivation, and forensic clinical
homicide factors. Results indicated that schizophrenic murderers—along with all other
schizophrenia and homicide groups—demonstrated reduced Verbal, Performance, and
IQ performance relative to normal controls, while some evidence for increased executive
functioning in schizophrenic murderers compared to schizophrenic non-murderers and
non-schizophrenic murderers was suggested. Also, while schizophrenic non-murderers
were characterized by increased diffuse slow-wave EEG activity compared to all other
groups, schizophrenic murderers were not and instead demonstrated increased left-
hemispheric fast-wave activity relative to schizophrenic non-murderers. Additionally,
group differences in current psychosocial functioning were also detected, and
schizophrenic murderers reported abnormally elevated levels of childhood psychosocial
deprivation; and were characterized by higher birth order than non-schizophrenic and
non-murderer counterparts. Furthermore, the homicides of schizophrenic murderers
vi
appeared distinct from other types of murderers in various motivational, behavioral,
and emotional aspects. Results are discussed within the framework of a proposed
biopsychosocial interactional trajectory toward violence in schizophrenia—perhaps
stemming from paternal germ line mutations, and involving left hemispheric
hyperarousal deficits in interactions with early psychosocial deprivation and specific
psychosocial stressors—which may lead to a homicidally violent schizophrenic
outcome. Ultimately, results suggest further evidence for a biopsychosocially distinct
subgroup of schizophrenic murderers, which could have implications in research,
treatment, forensic applications, and social perception.
1
Chapter 1: Background and Significance
Current understanding of the link between schizophrenia and violence across
empirical and public spheres is characterized by three “ironies.” First, while significant
empirical evidence indicates a strong relationship between schizophrenia and crime and
violence, scientists appear transfixed upon the reification of this relationship rather than
moving forward with a new generation of research into common risk factors and
etiological mechanisms (Raine, 2006). Second, while researchers are hesitant—possibly
due to fears of amplifying negative stigmatization (Arseneault, Moffitt, Caspi, Taylor, &
Silva, 2000; Raine, 2006)—public opinion appears guided by the misperception that all
individuals with schizophrenia are violent and dangerous (Schwartz, Petersen, &
Skaggs, 2001), due largely to sensationalized accounts of schizophrenia-associated
violence which have proliferated through the popular media (Junginger, 2006; Raine,
2006). Third, while the public maintains this grossly distorted view of violence in
schizophrenia, most schizophrenic individuals in fact pose no risk for violence (Joyal,
Putkonen, Paavola, & Tiihonen, 2004). Together, these “ironies” indicate an urgent need
for clarification, and the failure to address one important question—what causes only
some individuals with schizophrenia to become violent. Ignoring this question, it seems,
precludes both reducing the stigma unjustly attached to schizophrenic persons who are
not violent, and advancing the understanding of those who are.
Schizophrenia, Violence, and Homicide: Scope of the Problem
The prevalence rates of violence among schizophrenic persons in general are
notably disproportional to their level of stigmatization—though rates do indicate a need
2
for concern. As mentioned above, most individuals with schizophrenia are not violent.
For example, self-report data from a national clinical antipsychotic trials sample of
1,410 schizophrenia patients indicated that the 6-month prevalence of any violence
among these individuals was 19.1% (15.5% reported minor violence such as simple
assault without injury or weapon use; and 3.6% reported serious violence such as
weapon- or injury-associated assault, threat with a lethal weapon, or sexual assault;
Swanson et al., 2006). However, significantly increased rates of violence have been
demonstrated in schizophrenic individuals when compared to non-schizophrenic control
groups. Birth cohort studies from Denmark (Brennan, Mednick, & Hodgins, 2000),
Finland (Tiihonen, Isohanni, Räsänen, Koiranen, & Moring, 1997), and New Zealand
(Arseneault et al., 2000) have reported rates of official records and self-reported violent
crime as high as 11.3 – 33.3 percent in schizophrenic and schizophrenia-spectrum-
disordered individuals, compared to lower rates (i.e., 2.2 – 3.8%) in control groups with
no mental illness (rates reported in women were lower overall—0.1% and 2.8% for
controls and schizophrenic women, respectively; Brennan et al., 2000). Together, these
rates may speak to a smaller but significantly more violent subgroup of individuals with
schizophrenia.
For homicide—the most extreme form of violence—differences are even more
striking. For example, general population base rates for homicide in the United States,
calculated using official homicide offending rates (Fox & Zawitz, 2007) and census data
(United States Census Bureau, n.d), are quite low. In the year 2005 (the most recent year
of availability of both sources of data at the time of this writing), overall rates were
3
approximately 0.01 percent for men and 0.001 percent for women—though these
figures are limited in that they may include multiple homicides committed by the same
individual, and include all individuals—even those with mental illness. However, in
contrast, a recent meta-analysis of nine studies (Brennan & Alden, 2006) indicated a
nearly-threefold risk for homicide in male schizophrenic individuals relative to their
non-schizophrenic counterparts, and a nearly-twenty two-fold risk for homicide in
female schizophrenic individuals relative to their non-schizophrenic counterparts.
Furthermore, in samples of homicide offenders, schizophrenic individuals appear over-
represented even in comparison to individuals with other forms of mental illness such as
mood disorders (i.e., depression and depressive disorders). A recent review of eight
studies (Schanda et al., 2004) indicates in the majority of these samples a two- to five-
fold increase in prevalence rates of offenders with schizophrenia relative to those with
depressive mood disorders; though the direction of this difference was reversed for
female offenders in two studies. In all, these figures illustrate the very serious nature of
violence which may characterize some individuals with schizophrenia, and further
underscore the need for more research into its causes.
One effective approach to further elucidating the nature of the schizophrenia-
violence relationship may be to examine this subgroup of homicidally violent persons
with schizophrenia. Murderers in general have traditionally been viewed as a population
worthy of scientific study. Empirical focus upon homicide offenders has its roots in the
writings of early positivist criminology (Lombroso, 1876) and the psychoanalytic
interpretations of case material of murderers from the early twentieth-century (e.g.,
4
Lehrman, 1939; Wittels, 1937; Zilboorg, 1935). Studies of murderers within more
recent decades have acquired neuroscience data from these individuals (see below), and
have also focused upon personality assessment (e.g., Holcomb, Adams, & Ponder, 1985;
Kahn, 1967; Meloy, Gacono, & Kenney, 1994) and the relationship between homicide
and various other forms of Axis I and Axis II psychopathology (e.g., Arrigo & Purcell,
2001; Eronen, Hakola, & Tiihonen, 1996; Hirose, 1979; Raine, 1993). Though a recent
waning in interest in neuroscience investigations of homicide offenders in general is
evident, a brief re-emergence of these approaches was observed when several PET
studies on murderers by Raine and colleagues (e.g., Raine, Buchsbaum, & LaCasse,
1997) were published.
The literature related to homicide in schizophrenia is somewhat more limited and
dates back several decades. Initially, the schizophrenia-homicide relationship was
reflected in rates of schizophrenia prevalence reported among studies of homicide
offenders, as these reports tended to focus upon prison or special hospital samples. Rates
of schizophrenia among murderers studied since 1900 have ranged from four to 83
percent (Schanda et al., 2004; Yarvis, 1990). Though interesting, these reports were
largely descriptive in nature (as opposed to comparative) and unfortunately contributed
little to a greater understanding of the differences between murderers with and without
schizophrenia, and of the possible unique biological and social correlates of extreme
violence in schizophrenic individuals. Several key approaches—neuropsychological and
psychophysiological assessment (indices of brain functioning), along with examination
of early adverse environmental factors and social impairments—may help to quantify
5
and qualify the differences between schizophrenic individuals who murder versus those
who do not.
Neuropsychology of Schizophrenia
Both Blueler (1911) and Kraepelin (1919) described cognitive disturbances as
being among the core features of schizophrenia, and impairments in cognitive function
are still thought to be fundamental characteristics of the disorder (Joyce & Roiser,
2007). A sizeable literature on the neuropsychology of schizophrenia has been
developed documenting associated neuropsychological deficits (Wilk et al., 2005),
including generalized intellectual impairments or IQ deficits (Aylward, Walker, &
Bettes, 1984; Barkataki et al., 2005), along with a broad range of cognitive deficits in
domains such as attention, learning, memory, executive function, and processing speed
(Barkataki et al., 2005; Bilder, 2006; Dickinson, Ramsey, & Gold, 2007; Heinrichs &
Zakzanis, 1998; Wilk et al., 2005). These deficits suggest a pattern of widespread
cortical dysfunction in schizophrenic individuals which is relatively stable over time
(Gold, 2004; Mockler, Riordan, & Sharma, 1997; Rund et al., 2007) and independent of
symptomatology such as delusions, hallucinations, and negative symptoms (Gold, 2004;
Hughes et al., 2002). However, evidence suggests the existence of significant
heterogeneity among the neurocognitive profiles of individuals with schizophrenia
(Joyce & Roiser, 2007). For example, differential patterns of cognitive deficits have
been observed in paranoid versus non-paranoid schizophrenics, with paranoid
schizophrenics being characterized by better verbal intelligence, executive function, and
memory (Bilder, 2006; Kumar & Sharma, 1991; Seltzer, Conrad, & Cassens, 1997).
6
Differential neuropsychological performance has been observed among schizophrenia
patients with dynamic (i.e., progressively increasing) versus static ventricular volume
(Osuji et al., 2007). Several authors (Bilder, 2006; Osuji et al., 2007) have proposed
discrete neuropsychological subtypes of schizophrenia, and two subgroups of
schizophrenia (one characterized by generalized impairment and one by normal
performance, except executive function) have been reliably identified by two research
groups. Additionally, other authors have examined dichotomization into memory
impaired and unimpaired groups of schizophrenics (review in Joyce & Roiser, 2007).
Given this information, it is not unreasonable to speculate there may exist a distinct
neuropsychological profile of schizophrenia predisposed to antisocial behavior and even
extreme violence—which may explain why some individuals with schizophrenia
commit homicide whereas others do not.
Neuropsychology of Homicide
The neuropsychology of the broader spectrum of antisocial behavior has also
received substantial empirical attention, and has implicated deficits in general
intelligence (Eriksson, 2006; Wilson & Herrnstein, 1985) and executive function (Dolan
& Park, 2002; Morgan & Lilienfeld, 2000; Stanford, Conklin, Helfritz, & Kockler, 2007;
Teichner, Golden, Van Hasselt, & Peterson, 2001) in violent, criminal, and antisocial
persons. In stark contrast to this body of literature—and the extensive body pertaining to
the neuropsychology of schizophrenia—the literature related to neurocognitive
performance in homicide offenders is smaller, of relatively poor quality, and has
produced mixed results. Neuropsychological data of any type from murderers is
7
especially scarce and has been largely incidental (i.e., when reported, it has not been
the primary focus or incorporated into the main research hypothesis of a given study)
and almost exclusively related to intellectual functioning. This incidental data, which
began to appear toward the middle of the last century in the form of case histories of
murderers (e.g., Bender & Curran, 1940; Patterson, 1942; Rosanoff, 1943), appears
significantly varied—describing IQ scores ranging from 32 to 153. Limited IQ data have
also appeared among the demographic descriptors of samples of murderers (e.g., Eronen
et al., 1996; Harbort & Mokros, 2001; Myers, Scott, Burgess, & Burgess, 1995; Pagan &
Smith, 1979), and other studies have subsumed murderers within larger categories of
offenders (Cornell, Roberts, & Oram, 1997; Valliant, Asu, Cooper, & Mammola, 1984).
Descriptive studies of murderers have reported mean IQ scores around one SD
below normal (Cole, Fisher, & Cole, 1968; Fisher, 1999; Warnick, 2007), while others
report relatively normal distributions (e.g., Snook, Cullen, Mokros, & Harbort, 2005).
Earlier comparative studies of intellectual functioning in murderers involved
comparisons to other types of offenders (Hays, Solway, & Schreiner, 1978; Kahn, 1967;
Wagner & Klein, 1977), and the general population (Deiker, 1973), in which murderers
overall demonstrated intellectual deficits. Some earlier studies suggested that murderers
are not characterized by PIQ > VIQ discrepancies (Deiker, 1973; DeWolfe & Ryan,
1984; Kahn, 1968); and more recent comparative studies between murderers and other
criminal subgroups have not found significant group differences in IQ performance
(Jamison, 2006; Langevin, 2003). The intellectual abilities of specific types and
8
subtypes of murderers have also been compared (Blackburn, 1971; Lester, Purdue, &
Brookhart, 1974), suggesting cognitive heterogeneity in intellectual functioning.
Heilbrun, Jr. (1982) conducted the first study of murderers (albeit not a sample
exclusively comprised of murderers) to both report on a specific neuropsychological
domain, executive functioning, rather than the global construct of intelligence alone, and
to employ more than one instrument in the process (i.e., the Stroop Interference test).
Other tests of executive function (i.e., the category test, TMT, TOL, WCST) have been
employed in subsequent investigations of murderer samples (Langevin, Ben-Aron,
Wortzman, Dickey, & Handy, 1987; Langevin, Ben-Aron, Wright, Marchese, & Handy;
1988; Ticehurst, Ryan, & Hughes; 1992) and comparisons of murderers and controls
(Cornell et al., 1997).
Neuropsychological performance differences have not been found in
comparative studies of murderers and controls on tests of other domains of
neuropsychological function, such as attention (Raine, Buchsbaum, & LaCasse, 1997).
In this study, no differences in Continuous Performance Task (CPT) performance were
observed between murderers and controls; nor were CPT differences found in
subsequent studies between individual subgroups of murderers (i.e., murderers
with/without psychosocial deprivation, and predatory/affective murderers; Raine, Meloy
et al., 1998; Raine, Stoddard, Bihrle, & Buchsbaum, 1998). Together, these results speak
to the cognitive heterogeneity of murderers, and the enormous heterogeneity in general
intelligence observed in the literature may be due to failure to consider schizophrenia as
9
a moderating factor (in fact, the lowest of the reported IQ scores—32, by Rymer
[1942]—was from a schizophrenic murder).
Neuropsychology of Schizophrenia and Homicide. The neuropsychological
correlates of schizophrenic homicide have rarely been explored in the literature. While a
handful of studies (Barkataki et al., 2005; Eriksson, 2006; Kumari et al., 2006; Nestor,
Haycock, Doiron, Kelly, & Kelly, 1995; Silver, Goodman, Knoll, Isakov, & Modai,
2005; Wong et al., 1997) report the inclusion within their samples of homicidal
individuals with schizophrenia, these individuals are subsumed within larger groups of
schizophrenic violent offenders and thus cannot be characterized on their own. To date,
only one formal neuropsychological comparison of schizophrenic murderers and non-
murderers (Barber, 1994) has been undertaken. In this unpublished master’s thesis, the
performance of 14 male schizophrenic murderers from a maximum security psychiatric
hospital and 14 male schizophrenia outpatients from a psychiatric hospital was
compared across ten neuropsychological tests, including attention, executive function,
language, and memory tasks. Results indicated that schizophrenic murderers
demonstrated significantly better performance on two tests of executive function
(Classical Weigl and WISC Mazes errors), and overall outperformed non-murderers on
all but two of the tests. However, compared to murderers, non-murderers demonstrated
significantly higher psychotic symptomatology (i.e., delusions, flat affect, and
psychomotor retardation) and higher rates of hallucinations, incoherence, and poverty of
speech—which may have adversely affected their overall performance.
10
One other investigation has compared murderers with schizophrenia to those
without. Nestor, Kimble, Berman, and Haycock (2002), in a study of 26 murderers in a
maximum security forensic hospital, used the WAIS-R, the Wechsler Memory Scale—
Revised (WMS-R; Wechsler, 1987), WRAT-R, and WCST to distinguish between high
psychopathy and high psychosis individuals; each with distinct NP differences in
intellectual abilities. Results indicated that psychotic group had higher verbal than
performance IQs, and the reverse was true for the psychopathic group. To date, no
studies have made neuropsychological comparisons among schizophrenic murderers,
schizophrenic non-murderers, non-schizophrenic murderers, and controls.
A recent series of meta-analyses of 43 studies across neuropsychological
domains (Schug & Raine, 2009) indicates antisocial schizophrenic individuals are
characterized by widespread deficits across multiple domains (Full Scale IQ, Verbal and
Performance IQ, attention, broadly-defined executive function, and memory) in
comparison to their antisocial counterparts; and instead by reduced general intellectual
functioning and memory dysfunction in comparison to their schizophrenic counterparts.
This was the first quantitative synthesis of a somewhat limited body of research related
to broader constructs of schizophrenia and antisocial behavior, and to date this
neuropsychological deficit profile has not been examined in comparisons of
schizophrenic murderers to their non-murdering and non-schizophrenic counterparts.
Electroencephalography
Electroencephalography is another potentially promising avenue for elucidating
the biological correlates of extreme violence in schizophrenia. The
11
electroencephalogram (EEG) is a recording of the difference in electrical potential
between various points on the surface of the scalp (Hugdahl, 1995), likely reflecting the
depolarizations of the dendritic trees of pyramidal cells in the cortex (Raine, 1993). EEG
can be broken down into different frequency components, most commonly delta (0 – 4
Hz), theta (4 – 8 Hz), alpha (8 – 12 Hz), and beta (13 – 30 Hz). These components have
been aligned with a continuum of consciousness. Delta is associated with deep sleep, as
well as brain pathologies such as tumors when present in awake individuals. Theta is
associated predominantly with drowsiness and low levels of alertness, and has shown
relationships with the cessation of pleasurable activity, hypnagogic imagery, REM
(rapid eye movement) sleep, problem solving, attention, and hypnosis. Alpha is
associated with relaxed wakefulness and lack of active cognitive processes, and beta
with alertness and vigilance. Traditionally, lower-frequency variations (18-30 Hz) have
been referred to as beta and higher frequency variations (30-70 Hz or higher) as gamma.
Gamma activity may reflect the brain’s integration of multiple stimuli into a coherent
whole (Raine, 1993; Stern, Ray, & Quigley, 2001). All frequencies are present in the
awake individual, but individuals differ in the relative amount of power existing in these
main frequency bands (Raine, 1993). Resting EEG is characteristic and is associated
with specific personality and cognitive features (Vogel & Schalt, 1979).
EEG studies of schizophrenia. The classic finding in the literature on resting
EEG in schizophrenia is increased slow-wave activity (i.e., delta and theta) in
schizophrenic persons (Clementz, Sponheim, Iacono, & Beiser, 1994). Increased delta
and increased theta and decreased alpha activity have been observed in the resting EEG
12
of schizophrenia patients compared to controls (Clementz, Sponheim, Iacono, &
Beiser, 1994; Sponheim, Clementz, Iacono, & Beiser, 1994), though no significant
differences in frequency composition were observed between first-episode and chronic
schizophrenia patients (Sponheim et al., 1994)—suggesting these abnormalities are
stable characteristics and not related to treatment in schizophrenia. Mientus et al. (2002)
found a significant increase in delta activity over the whole cortex, most strongly in the
anterior cingulated gyrus and temporal lobe, in the resting EEGs of schizophrenia
patients. Other authors (Miyauchi et al., 1990) have reported an increase of slow band
activity in the parieto-occipital regions of schizophrenia patients. More recent studies
using advanced 123-channel EEG montage have shown increased topographic
complexity and fragmentation in the slow potentials of schizophrenic individuals
(Basile, Yacubian, Ferreira, Valim, & Gattaz, 2004).
One interpretation is that frontally pronounced slowing of EEG activity in
schizophrenia represents hypofrontality—one of the most prominent and consistent
findings in neuroimaging studies of schizophrenia (see reviews in Basile, Yacubian,
Ferreira, Valim, & Gattaz, 2004, and Mientus et al., 2002). EEG slowing correlates with
reduced cerebral blood flow and glucose utilization in schizophrenia patients (Wuebben
& Winterer, 2001). Another interpretation is that increased slow-wave activity
represents greater brain activation (Clementz et al., 1994; Miller, 1989)—consistent with
findings of increased slow band activity during hallucinations in schizophrenia and
heightened creativity in normal controls, and increased cerebral blood flow observed in
some samples of schizophrenic patients (Miller, 1989).
13
Some authors have noted reduced alpha along with increased delta in the EEG
spectra of schizophrenic individuals (Iacono, 1982; Itil, Saletu, & Davis, 1972; Itil,
Saletu, Davis, & Allen, 1974). Decreased alpha activity may be associated with
prominent negative symptoms in schizophrenia, though findings are mixed (Merrin &
Floyd, 1996). Increased power has been reported in higher EEG frequency bands (i.e.,
alpha and beta) though details of results tend to be more variable than for slower
frequencies (Miller, 1989).
EEG studies of homicide. Extensive reviews indicate that hundreds of studies
have assessed EEG in antisocial populations such as criminals, delinquents,
psychopaths, and violent offenders (see Raine, 1993). While a large number of studies
have implicated EEG abnormalities in violent recidivistic offenders, results from EEG
studies of psychopathic individuals are much more inconsistent. EEG arousal deficits in
criminals are most frequently observed in the frontal and temporal regions of the brain
(Ishikawa & Raine, 2002).
Though an initial flourish of interest was displayed via descriptive studies toward
the middle of the last century, the literature on the EEG recordings of murderers remains
significantly limited in comparison to the larger body of work compiled on EEG in
antisocial populations in general. These earlier studies focused primarily upon reporting
prevalence of EEG abnormalities among samples of murderers, and describing the
particular types (i.e., mild nonspecific, severe nonspecific, and focal/epileptic) and
location of abnormalities observed—including abnormalities specific to individual EEG
components (i.e., delta, theta, alpha, beta) and other complexes or activities (e.g., spikes,
14
sharp waves, etc.; Hill & Pond, 1952; Mundy-Castle, 1953; Stafford-Clark & Taylor,
1949). While these abnormalities varied greatly in nature and scope, many tended to be
described in temporal regions.
Despite the incidental reporting of EEG in both larger descriptive and individual
case studies of murderers (Green, Leon-Barth, Venus, & Lucey, 2001; Lewis et al.,
1988; Lewis et al., 1985; Mouridsen & Tolstrup, 1988), the EEG-homicide literature is
often a bit richer in clinical detail than that from neuropsychological studies of the same
topic. While some recent studies (e.g., Blake, Pincus, & Buckner, 1995) report similar
prevalence rates of EEG abnormalities among murderers to those found in earlier
studies, rates of 9-60% have been reported recently in the literature (Driver, West, &
Faulk, 1974; Sakuta & Fukushama, 1998; Winkler & Kove, 1961—rates that remain are
more or less higher than the 5-20% reported in the general population (Mednick,
Volavka, Gabrielli, & Itil, 1982). Similar types of abnormalities (i.e., diffuse or
symmetric slowing, bi-temporal spikes, tempero-parietal sharp waves and slowing, and
focal slowing) have been recently reported (Blake et al., 1995). Other results have been
somewhat mixed. Evans and Park (1997), in a sample of 20 murderers, found the
greatest concentrations of multiple abnormalities in bilateral frontal, right temporal, and
parietal sites. Gatzke-Kopp, Raine, Buchsbaum, and LaCasse (2001) found significant
increases in slow-wave (delta and theta) and beta1 activity in the temporal, but not
frontal, lobe in murderers compared to controls. Green et al. (2001), however, report
frontal EEG abnormalities in a case of a juvenile murderer. Lindberg et al. (2005) found
15
overall reduced alpha power and bilaterally increased occipital delta and theta power
in murderers compared to controls.
EEG studies of schizophrenia and homicide. To date no formal EEG
comparisons of schizophrenic to non-schizophrenic murderers have been made, though
individuals with schizophrenia have been included within homicide samples. For
example, Gatzke-Kopp et al. (2001) included schizophrenic individuals within their
sample, but did not report the exact number of these individuals or any
schizophrenia/non-schizophrenia EEG comparisons. Mundy-Castle’s (1955) sample of
21 murderers was comprised predominantly (81%) of individuals with schizophrenia
and other psychotic disorders, though no comparisons were made between those with
schizophrenia and those without. Stafford-Clark and Taylor (1949) divided one of the
five murder subgroups (guilty but insane) into groups based upon psychiatric
diagnoses—epilepsy, schizophrenia, depression, psychopathic personality, and
paranoia—and reported (without statistical comparisons) the number of individuals
within each diagnostic category possessing EEG abnormalities.
The systematic exploration of EEG recordings in schizophrenic and non-
schizophrenic murderers may lead to evidence implicating schizophrenia as a
moderating influence on extreme forms of violence. For example, Mednick et al. (1982)
suggest that the prevalence of EEG abnormalities in violent individuals ranges from 25
to 50%, with the range of abnormalities in normals ranging from 5 to 20%. Though the
approximated prevalence of abnormalities in murderers (9-60%) is higher than that of
normals, no explanation has been offered as to why nearly half of murderers are
16
characterized by EEG abnormalities, whereas half are not. This may speak to the
biological heterogeneity of murderers—indexed by cortical activity and potentially
moderated by the presence of schizophrenia—which may be crucial in the overall
understanding of homicidal behavior. Additionally, it might offer important insight into
the biological reasons why some individuals with schizophrenia engage in these
behaviors, whereas others do not.
Psychosocial Correlates
Psychosocial impairment. The importance of the consideration of moderating
psychosocial influences in biological-violence relationships has been discussed in the
literature, though these influences are often not considered (see Raine, Stoddard et al.,
1998). Psychosocial functioning limitations remain a major source of disability in
schizophrenia, as unemployment, and lack of reliable friends and leisure activities are
associated with low subjective quality of life in individuals with this disorder (Prouteau
et al., 2005). To date, these factors remain largely unexamined in antisocial
schizophrenic populations in general and schizophrenic murderers in particular. In one
study, however, Moran and Hodgins (2004) used four indices of psychosocial
functioning in a sample of 232 men with schizophrenic disorders (including 38
murderers); and found more deficits in GAF, intimate relationships, employment
history, and compulsory military service in those with ASPD compared to those without
ASPD—though these differences were not significant. These results, though in need of
replication, offer some insight into the question of if schizophrenic murderers are
17
characterized by more pronounced patterns of psychosocial deficits than their non-
murderous or non-schizophrenic counterparts.
Early psychosocial deprivation. Psychosocial factors may also play a key role in
the etiological processes leading to the development of extreme violence in
schizophrenia; thus, it is important to examine dimensions of psychosocial deprivation
occurring during early formative years. Individuals who later become schizophrenic are
disproportionately likely to have suffered the disadvantages of social deprivation (i.e.,
low SES) both in utero and in early life (Venables, Raine, Dalais, Liu, & Mednick,
2006). The limited work on early psychosocial deprivation in antisocials stems from the
early pioneering psychopathy research of Bowlby (1946), who suggested that early
prolonged maternal separation resulted in an inability to form later emotional
attachments with others (i.e., “affectionless psychopathy”). McCord and McCord (1964)
argued that the development of psychopathy is influenced by parental rejection, an
adolescent parent, erratic discipline, and poor parental supervision. Other family factors
possibly associated with early psychosocial deprivation have received empirical support
in their relationship to psychopathic-like traits. These include low parental involvement,
abuse or neglect, parental conflict and disrupted families, large family size, criminal or
antisocial parents or siblings, other parental characteristics (substance abuse, stress or
depression, working mothers), and socioeconomic factors such as low income and poor
housing (see Farrington, 2006). Recent studies have shown that successful psychopaths
have demonstrated increased parental absence in comparison to unsuccessful
psychopaths and controls (Ishikawa, Raine, Lencz, Bihrle, & LaCasse, 2001).
18
Psychopaths reported significantly more childhood social stressors (low parental SES,
parental verbal and physical discord, childhood sexual abuse, and family alcohol abuse)
than both nonpsychopathic criminals and controls (Schug & Fung, 2007).
Studies of homicidal individuals have produced similar findings. Various early
psychosocial deprivation factors have been reported in samples of sexual murderers
(Ressler, Burgess, & Douglas, 1988) and juvenile murderers (Lewis et al., 1988).
However, murderers without any clear history psychosocial deprivation demonstrated
reduced prefrontal glucose metabolism than murderers with a history of psychosocial
deprivation and controls, suggesting a stronger predisposition to violence via brain
abnormalities in this group (Raine, Stoddard et al., 1998). Though schizophrenic
murderers and non-murderers have not to date been formally compared along
dimensions of childhood deprivation, Moran and Hodgins (2004), in the aforementioned
sample, found significantly lower parental SES, and higher rates of physical abuse,
childhood institutionalization, and parental criminality and substance abuse in those with
ASPD compared to those without ASPD. Higher rates of childhood sexual abuse and
parental violence were observed in schizophrenic ASPDs, though these differences were
not significant. To date, no studies have investigated the possibility of distinct patterns
of childhood psychosocial deprivation among schizophrenic murderers and their non-
murderous and non-schizophrenic counterparts.
Forensic Clinical Aspects
The motivational, behavioral, and affective correlates of homicide—though
reported in individual case studies of murderers in both scientific and nonscientific
19
literatures—are rarely examined in empirical investigations of homicide offender
samples. Rarer still are empirical comparisons of these forensic clinical aspects among
murderers with and without schizophrenia. Motivational factors such as antecedent
arguments or drug/alcohol use have been examined; for example, though psychotic
murderers are reportedly less often addicted to drugs than non-psychotic murderers
(Nijman, Cima, & Merckelbach, 2003), schizophrenic murderers with ASPD are
reportedly more likely to have used alcohol and to be involved in an altercation with the
victim prior to the homicide than those without ASPD (Joyal et al., 2004).
Differential behavioral aspects of schizophrenic homicides have also been
reported. For example, Varma and Jha (1966) found murders committed by mentally ill
persons to be characterized by an absence of apparent motive, no attempts at crime
concealment, and no apparent pre-planning—and it has been suggested that the crime
scenes and personal characteristics of psychotic murderers reflect more disorganization
than those of psychopathic murderers (Ressler et al., 1988). Studies addressing
victimology have noted that schizophrenic homicides usually involve only one victim,
who is typically known—often intimately—by the offender (i.e., a family member,
partner, or acquaintance; Häkkänen & Laajasalo, 2006; Joyal et al., 2004; Nijman et al.,
2003; Varma & Jha, 1966). Schizophrenic murderers with ASPD, however, are more
likely to target a non-relative than those without ASPD (Joyal et al., 2004). Though
sexual crimes are reportedly rare among psychotic patients (Nijman et al., 2003),
extreme violence including features of sadism, mutilation, sexual components or
multiple stabbings has been observed in about one-third of psychotic homicides
20
(Laajasalo & Häkkänen, 2006). Furthermore, homicides committed by mentally ill
persons tend to be particularly brutal in nature (Varma & Jha, 1966), and schizophrenic
murderers are reportedly more likely to use a sharp weapon, and to injure the victim’s
face in comparison to offenders with personality disorders, substance use disorders, and
offenders without a diagnosis (Häkkänen & Laajasalo, 2006). Finally, murders by
schizophrenic persons appear to take place more often during daytime hours on
weekdays (as opposed to weekends; Häkkänen & Laajasalo, 2006).
The affective correlates of homicide (i.e., the subjective emotional experience of
the offender before, during, and after the commission of the crime) have not been
systematically explored in murderers in general—though Varma and Jha (1966) report
murders committed by mentally ill offenders to be characterized by “complete emotional
indifference” and claims of amnesia for the crime. To date, formal comparative studies
of differential affective presentations among murderers with and without schizophrenia
have not been undertaken. Examining these motivational, behavioral, and affective
components of murder would be invaluable to the study of homicide in general and of
schizophrenic homicide in particular, and could be used in conjunction with
biopsychosocial data to elucidate the complex etiological processes at work in extreme
forms of schizophrenic violence. Additionally, they may suggest a form of homicidal
behavior specific to schizophrenia in particular, perhaps due to unique biological and
psychosocial risk factors or influences.
21
Schizophrenic Homicide: A Biopsychosocially Distinct Subgroup?
While schizophrenic individuals who murder may simply reflect the extreme
forms or additive effects of both conditions (see Schug & Raine, 2009), evidence
continues to mount which suggests that they may in fact be a unique and separate group
with distinct biological characteristics. Two recent studies have provided evidence for
the existence of such an antisocial schizophrenic subgroup. In a community sample,
Schug, Raine, and Wilcox (2007) found significantly reduced electrodermal orienting
(indexing, in part, prefrontal deficits) and increased criminality in individuals with both
schizophrenia spectrum personality disorders and antisocial personality disorder
compared to normal controls and those with either condition alone. Additionally, the
aforementioned series of meta-analyses (Schug & Raine, 2009) suggests a subgroup of
antisocial schizophrenic persons distinct from their non-schizophrenic counterparts by
widespread brain deficits and from their non-antisocial counterparts by specific deficits
in brain regions subserving general intellectual and memory functions. The question
remains, however, as to whether these individuals are set apart by any additional
biological and/or psychosocial characteristics.
Though these two studies utilized relatively broad-based definitions of
schizophrenia and antisociality, results have not to date been replicated with narrower
operationalizations of these constructs, to control for potential confounding effects of
differing definitions (see Schug & Raine, 2009). Replication using a specifically-defined
diagnostic group (i.e., schizophrenia) along with a specifically-defined antisocial
behavior (e.g., homicide) would be beneficial in that it would significantly reduce the
22
confounding effects of definitional heterogeneity, while potentially providing for
more robust effects as these constructs represent the extreme forms of their respective
definitional dimensions. Additionally, the possible translation of this distinct antisocial
schizophrenic biopsychosocial profile into unique forms of violent (i.e., homicidal)
behaviors specific to this group to date remains unexamined.
Specific Aims and Hypotheses
The overarching goal of the present study was to examine a specific group of
individuals—those characterized by both schizophrenia and homicide—to ascertain if
any biological, psychosocial, or forensic clinical factors distinguished them form those
with either condition alone, and provided further evidence for a distinct subtype of
homicidal schizophrenic persons. To that end, there were three distinct aims. The first
aim was to identify indicators of a potential biopsychosocial profile unique to
schizophrenic murderers, using four important comparisons: (1) comparisons of
schizophrenic murderers and schizophrenic persons who do not murder, to determine
which (if any) biological and/or psychosocial factors explain why only some individuals
with schizophrenia become homicidal; (2) comparisons of schizophrenic murderers and
murderers without schizophrenia, to determine if these factors were not characteristic of
those contributing to homicidal violence in general; (3) comparisons of schizophrenic
murderers and murderers with other psychiatric conditions, to see if these factors are
specific to homicidal violence in schizophrenia or merely reflecting the effects of mental
illness in general; and (4) comparisons with normal individuals.
23
The second aim was to determine—using forensic clinical data—if the
homicides of schizophrenic murderers were qualitatively different in any way (i.e.,
motivationally, behaviorally, or emotionally) from murderers without schizophrenia,
perhaps consistent with elements of a specific biopsychosocial profile for violent
schizophrenia. Finally, the third aim was to ascertain if revealed differences in biological
or psychosocial functioning, or any qualitative differences in homicidal behavior,
suggested an etiological pathway toward violence unique to schizophrenic murderers—
distinct from schizophrenic non-murderers, non-schizophrenic murderers, and even
murderers with other types of mental illness.
Specifically, the present study examined the following hypotheses:
1. Schizophrenic murderers would demonstrate widespread brain deficits as measured
by neuropsychological tests of general intellectual ability when compared to non-
schizophrenic murderers, individuals with schizophrenia who do not murder, and
controls.
2. Schizophrenic murderers would demonstrate frontal lobe deficits as measured by
neuropsychological tests of executive function when compared to non-schizophrenic
murderers, individuals with schizophrenia who do not murder, and controls.
3. Schizophrenic murderers would demonstrate multi-regional brain impairments as
measured by resting electroencephalogram (EEG) when compared to non-
schizophrenic murderers, individuals with schizophrenia who do not murder, and
controls. Specifically, these impairments would indicate widespread reduced cortical
24
activity; though differences will be particularly robust in the frontal and temporal
regions.
4. Schizophrenic murderers would be characterized by significantly more psychosocial
impairment and history of childhood psychosocial deprivation when compared to
non-schizophrenic murderers, individuals with schizophrenia who do not murder,
and controls.
5. Murderer groups would differ significantly on various forensic clinical aspects.
Specifically, the homicides committed by individuals with schizophrenia would
differ significantly in motivational, behavioral, and affective components from those
committed by murderers who do not suffer from schizophrenia.
25
Chapter 2: Method
Participants
Participants (162 men and women in Nanjing, China) were classified into five
diagnostic groups: individuals accused of homicide without psychosis (n = 31);
individuals accused of homicide who also suffer from schizophrenia or other psychoses
(n = 32); individuals accused of homicide who suffer from non-psychotic psychiatric
disorders (i.e., homicide psychiatric controls), including mental disorders due to brain
damage or vascular disease, epilepsy, hallucinosis due to alcohol use, alcohol
dependence, depression, mania, acute stress disorder, and twilight (n = 14); non-violent
patients with schizophrenia (n = 33); and normal controls (n = 47). Participants were
recruited from the Nanjing Brain Hospital at Nanjing Medical University—accused
murderers were detainees who were undergoing forensic psychiatric evaluation, non-
violent individuals with schizophrenia were hospital inpatients, and normal controls
were employees and community members, screened for history of mental illness.
Groups did not differ significantly in age, gender composition, or Hollingshead’s Two
Factor Index of Social Position (Miller, 1983). Group demographic characteristics are
listed in Table 1.
Participants took part in multiple test sessions in the Psychology, Forensic
Psychiatric, and Neurophysiology Departments of the Nanjing Brain Hospital. This
study was approved by the institutional review board at the University of Southern
California.
26
27
Screening for CCMD-3 and DSM-IV Disorders
Diagnostic interviews were conducted by a psychiatrist to assess the lifetime
presence of Axis I and Axis II psychopathology. For each diagnostic category, two
separate diagnoses were rendered based upon different classification systems—the
Chinese Classification of Mental Disorders Version 3 (CCMD-3; Chinese Society of
Psychiatry, 2001) and fourth edition of the Diagnostic and Statistical Manual of Mental
Disorders (DSM-IV; American Psychiatric Association, 1994). Each of two psychiatrists
verified the lifetime presence of the following disorders: psychotic disorders, mood
disorders, personality disorders (paranoid, schizoid, dissocial [antisocial in the DSM-
IV], impulsive, and schizotypal [DSM-IV only]), substance use disorders, and epilepsy.
Additionally, detailed item-level coding of DSM-IV criteria—using the Structured
Clinical Interview for DSM-IV Axis I Disorders (SCID I; First, Spitzer, Gibbon, &
Williams, 1997), and the SCID Axis II Personality Disorders (SCID II; First, Gibbon,
Spitzer, Williams, & Benjamin, 1997)—was conducted for dyssocial (antisocial)
personality disorder, conduct disorder, schizotypal personality disorder, and
schizophrenia. Coding for CCMD-3 and DSM-IV subtypes of schizophrenia (paranoid,
hebephrenic, catatonic, simple, undifferentiated, other) was also conducted.
Furthermore, data related to current antipsychotic medication use and history of
hospitalization for head injury was also collected during diagnostic interviews.
Neuropsychological Measures
Wechsler Adults Intelligence Scale: Revised in China (WAIS-RC; Gong, 1992).
The WAIS-RC is a variant of the WAIS-R (Wechsler, 1981) that has been used in
28
China. Full Scale IQ was estimated by prorating four WAIS-RC subtests—two from
the Verbal Scale (Similarities, Arithmetic), and two from the Performance Scale (Picture
Completion, Digit Symbol Coding). Raw subtest scores were converted to scaled scores
and subsequent Verbal, Performance, and Full Scale IQ scores based upon the area in
which participants reported being raised (i.e., city or country).
Color Trails Test (D’Elia, Satz, Uchiyama, & White, 1996). The Color Trails
Test (CTT) was administered as a culture-free analogue to the Trail Making Test (TMT)
from the Army Individualized Test Battery (1944). Instead of using the alphabet, the
CTT presents numbered colored circles. In CTT-1, a test of sustained visual attention
(similar to TMT-A), subjects are required to connect a series of numbered circles in
order (numbered 1-25, with even-numbered circles colored yellow and odd-numbered
ones colored pink) with a pencil. In CTT-2, a test of executive function (analogous to
TMT-B), subjects are presented with a different series of numbered circles (numbered 1-
25 in both yellow and pink separately) and are required to connect the circles in
numerical order while alternating between the two colors (1P-2Y-3P, etc). The CTT
retains the same psychometric properties as the TMT (Mitrushina, Boone, & D’Elia,
1999), and has demonstrated functional equivalency with the TMT in German
volunteers (Maj et al., 1993) and Chinese individuals in Hong Kong (Lee & Chan,
2000). For each trial, time to completion and total number of errors, near-misses, and
prompts are scored. The Interference Index is calculated by subtracting the CTT-1 raw
score from the CTT-2 raw score, and dividing by the CTT-1 raw score. All CTT scores
29
were converted to T scores using a normative data sample (D’Elia, Satz, Uchiyama, &
White, 1996).
Gambling Task (Bechara et al., 1994). A computerized version of the Gambling
Task (see Bechara, Damasio, Damasio, & Lee, 1999) was administered as an
orbitofrontal / ventromedial prefrontal test of executive function. Numerous
investigations (e.g., Lamar & Resnick, 2003; see also Bechara et al., 1999, and Ritter,
Meador-Woodruff, & Dalack, 2004, for reviews), including studies of psychopathy (see
Blair & Frith, 2000), have utilized the Gambling Task as an orbitofrontal / ventromedial
task. In this test, the participant attempts to win “money” by picking “cards” from four
decks (labeled A, B, C, and D) on the screen. The selected cards are tracked by the
computer, and the amount of money won/lost after each selection, along with the total,
are displayed. In accordance with previous studies using the Gambling Task (Mitchell,
Colledge, Leonard, & Blair, 2002), the task was divided into five blocks of 20 trials,
with the dependent variable being the number of disadvantageous selections (i.e., from
decks A and B) made in each block.
EEG Recording Procedures
Participants were tested in a temperature-controlled, light- and sound-attenuated
room. Participants were instructed to keep their eyes closed and to minimize all hand
and other body movements for the duration of the resting EEG recording. EEG was
recorded from 16 scalp sites according to the international 10/20 system: Fz, Cz, Pz, Oz,
Fp1, Fp2, F3, F4, F7, F8, T3, T4, P3, P4, O1, and O2. All electrodes were referred to
linked ear lobes, and a ground electrode was attached to the center of the forehead.
30
Resting EEG was recorded using a 16-channel electroencephalograph (Nicolet
Biomedical Bravo) with bandpass (0.1 – 70 Hz) and notch (50 Hz) filters, and a
sensitivity of 10 microvolts. The impedance was kept below 10 kΩ. Physiological
signals were acquired using a 16-bit 16 channel analog-to-digital converter with 1 KHz
maximum sampling rate, using a sampling frequency of 256 Hz. Resting EEG was
recorded during a three minute rest period at the beginning of a one-hour
psychophysiological testing session.
EEG data reduction. Spectral power was calculated from the following
frequency bands: delta (0.01- 4.00 Hz), theta (4.01 - 8.00 Hz), alpha1 (8.01 - 10.50 Hz),
alpha2 (10.51 - 13.00 Hz), beta1 (13.01 - 20.00 Hz) beta2 (20.01 - 30.00 Hz), and beta3
(30.01 - 50.00 Hz). Spectral power was calculated for each frequency band for each
electrode site, representing EEG recording over the entire three-minute rest period, free
of electrooculographic artifacts (i.e., eye blinks) and movement. Data from left
hemispheric sites (Fp
1
, F
3
, F
7
, P
3
, T
3
, O
1
) were averaged into one left hemispheric
measure. Data from right hemispheric sites (Fp
2
, F
4
, F
8
, P
4
, T
4
, O
2
) were averaged into
one right hemisphere measure. Data from medial prefrontal (Fp
1
and Fp
2
), dorsomedial
prefrontal (F
3
and F
4
), dorsolateral prefrontal (F
7
and F
8
), temporal (T
3
and T
4
), parietal
(P
3
and P
4
), and occipital (O
1
and O
2
) sites were averaged into individual lobe measures.
Psychosocial Functioning
Subjects were interviewed on psychosocial functioning measures covering
employment, independent living, and social (familial and non-familial) functioning.
Nine specific indices of psychosocial functioning were assessed. Categorical indices
31
included the presence/absence of recent homelessness, marital status (current and
history), and whether or not participants were currently living alone; and continuous
indices included the length of unemployment (in months) within the past year, total
number of couple relationships (i.e., boyfriends/girlfriends), number of close friends,
perceived emotional closeness to family members (not close/moderately close/very
close), and number of problems with the participant’s place and area of residence (i.e.,
overcrowding, damp, poor/no air-conditioning, neighbors, trash,
arguing/stealing/fighting, vermin, noise, structural problems, vandalism/graffiti,
gambling).
Childhood Psychosocial Deprivation
Participants were assessed via structured interview and self-report questionnaires
on 15 dimensions of childhood psychosocial deprivation (Ishikawa et al., 2001; Schug &
Fung, 2007). These included parental divorce, lacking one or both parents, parental
criminality, large family size, childhood verbal abuse, childhood physical abuse, low
parental SES, many different homes, parental verbal discord, parental physical discord,
parental mental health problems, parental physical health problems, parental alcohol
problems, parents on welfare, and childhood institutionalization. Categorical
dimensional variables derived from continuous variables (i.e., large family size, low
parental SES, many different homes) were converted using median splits of variable
data. Results on each dimension were summed into a composite childhood psychosocial
deprivation variable, and prorated sums were calculated in cases of missing item(s).
Additionally, birth order was assessed as a separate index of childhood psychosocial
32
deprivation, due to its unique ability to index child-rearing practices or the
psychosocial influences of having older or younger siblings (Cantor-Graae et al., 1997)
along with biological prenatal influences such as maternal viral infection (Sham,
Maclean, & Kendler, 1993), Rh incompatibility (Hollister, Laing, & Mednick, 1996), or
paternal germ line mutation (Tsuchiya et al., 2005).
Forensic Assessment Data
Motivation. Global ratings of homicide motivation were made at the end of the
interview based upon both case records and the forensic interview. These were coded
into a categorical homicide motivation variable (love/thuggee [see
below]/anger/retaliation/unknown to murderer).
Offense planning. Data related to the amount of homicide planning were coded
from case records and information yielded from the forensic interview. These data
reflected a spectrum of planning versus impulsivity (planned/mostly planned, some
impulsivity/mostly impulsive, some planning/impulsive) and were subsequently coded
for analyses as both a categorical planning variable (with the first three descriptors
indicating a presence of planning and the fourth an absence of planning) and a
continuous spectrum variable.
Sexual offense. Data related to the presence or absence of homicide concomitant
sexual offense were coded from case records and information yielded from the forensic
interview. These data were coded as a categorical sexual offense variable.
Victimology. Data related to homicide victimology were coded from case records
and information yielded from the forensic interview, which described the exact nature of
33
the relationship between murderer and victim (coded categorically as
stranger/acquaintance or friend/ family or relative/girlfriend or boyfriend, live-in partner,
spouse, or former spouse). Victim gender was also coded.
Emotional experience. Data related to offense-related emotional experience were
coded from case records and information yielded from the forensic interview. Data were
coded as descriptors reflecting negative affect/anxiety (afraid, angry, anxiety,
depression, depression, frustrated, heart beating, “in two minds,” out of control,
sweating, tension/dread, tired, upset), positive affect (aroused, happy, relieved,
satisfied), and no experience of emotion; and were collected to represent three distinct
time periods—before, during, and after the offense.
Statistical Analyses: Traditional and Modern Methods.
In some cases, where serious violations of the assumptions underlying traditional
statistical techniques were detected, additional modern methods were used. It is known
that conventional methods for comparing means can have very poor power (e.g.,
Wilcox, 2005). Comparing medians can reduce problems associated with methods for
comparing means, but a concern about using medians is that they trim too much of the
data, which again can result in relatively poor power. By trimming 20%, poor power due
to outliers, skewness, and variance can be reduced substantially, yet good power is still
achieved under standard assumptions (Wilcox, 2005). The employment of bootstrapping
techniques (see Wilcox, 2003) also seemed appropriate due to the small sizes of some of
the groups in the present study. Rather than assume normality to determine appropriate
34
critical values, bootstrap methods estimate appropriate critical values using the
available data (Efron & Tibshirani, 1993; Davison & Hinkley, 1997).
For all analyses involving group mean comparisons, boxplots were generated to
assess for violations of standard assumptions (i.e., significant outliers and/or skewness).
For the purposes of simplification, the reporting here of results from the aforementioned
modern methods indicates that boxplots had revealed the presence of significant outliers
and/or skewness in some cases, and that modern methods were subsequently employed
to augment conventional analyses (Schug et al., 2007).
35
Chapter 3: Results
Substance Use Disorders, Antipsychotics, and Head Injury
Participants were predominantly free of substance use disorders—approximately
one percent of the total sample was diagnosed with mental disorders due to substances
(one participant in the non-schizophrenic murderer group with mental disorder due to
cannabinoids use, and one participant in the homicide psychiatric control group with
mental disorder due to alcohol use).
Overall, 27.8% of the present sample was characterized by current usage of
antipsychotic medication (0.0% of normal controls, 100.0% of schizophrenic non-
murderers, 3.4% of non-schizophrenic murderers, 29.6% of schizophrenic murderers,
and 21.4% of homicide psychiatric controls). A Chi-Square analysis using all groups
revealed significant differences in proportions of current antipsychotics use among
groups ( χ
2
= 105.70, df = 4, p < .001). Subsequent Chi-Square analyses conducted on
individual group comparisons indicated that schizophrenic participants were
characterized by significantly greater proportions of current antipsychotics use than
normal controls ( χ
2
= 78.00, df = 1, p < .001), non-schizophrenic murderers ( χ
2
= 58.10,
df = 1, p < .001), schizophrenic murderers ( χ
2
= 33.98, df = 1, p < .001), and homicide
psychiatric controls ( χ
2
= 33.85, df = 1, p < .001); schizophrenic murderers by
significantly greater proportions of current antipsychotics use than normal controls ( χ
2
=
15.00, df = 1, p < .001) and non-schizophrenic murderers ( χ
2
= 7.11, df = 1, p = .008);
and homicide psychiatric controls by significantly greater proportions of current
antipsychotics use than normal controls ( χ
2
= 10.16, df = 1, p = .001), and by increased
36
proportions which approached significance in comparison to non-schizophrenic
murderers ( χ
2
= 3.62, df = 1, p = .057).
Antipsychotic medication names and diagnostic group frequencies are listed in
Table 2. Dosage level means and ranges for named antipsychotic medications were
consistent with normative levels reported in the literature (Ninan, 1989; Kutscher, 2008).
For unnamed antipsychotic medications, sample-wide recorded milligram dose levels (M
= 20.62, SD = 30.75) corresponded largely with low-milligram doses (i.e., 2-75 mg) of
high-potency neuroleptics (e.g., haloperidol)—as opposed to higher-milligram doses
(i.e., hundreds of milligrams) of low-potency neuroleptics (e.g., chlorpromazine)—
reported in the literature (Kutscher, 2008; Ninan, 1989). None of the available dosage
data immediately suggested high or megadose strategies (e.g., over 2000 mg of
chlorpromazine or 50 mg of haloperidol or equivalent; Ninan, 1989), and dose levels
across medication types appeared to be comparable, given the cross-medication dosage
equivalencies reported in the literature (Kutscher, 2008; Ninan, 1989).
Additionally, 13.2% of the present sample reported a history of hospitalization
for head injury (10.9% of normal controls, 6.1% of schizophrenic non-murderers, 29.2%
of murderers, 7.4% of schizophrenic murderers, and 21.4% of homicide psychiatric
controls). A Chi-Square analysis using all groups revealed differences which approached
significance in proportions of head injury history among groups ( χ
2
= 8.65, df = 4, p =
.071). Subsequent Chi-Square analyses conducted on individual group comparisons
indicated that non-schizophrenic murderers were characterized by significantly greater
proportions of head injury history than schizophrenic non-murderers ( χ
2
= 5.58, df = 1, p
37
Table 2. Antipsychotic Medications
Diagnostic groups
Medication name
S
(n = 33)
H
(n = 1)
SH
(n = 8)
HPC
(n = 3)
Antischizophrenic (no details)
16
(48.5%)
0
(0.0%)
5
(62.5%)
1
(33.3%)
Chlorpromazine
3
(9.1%)
1
(100.0%)
0
(0.0%)
1
(33.3%)
Clozapine
6
(18.2%)
0
(0.0%)
1
(12.5%)
0
(0.0%)
Olanzapine
1
(3.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
Perphenazine
0
(0.0%)
0
(0.0%)
0
(0.0%)
1
(33.3%)
Risperidone
7
(21.2%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
Sulpiride
0
(0.0%)
0
(0.0%)
1
(12.5%)
0
(0.0%)
Traditional Chinese drug
0
(0.0%)
0
(0.0%)
1
(12.5%)
0
(0.0%)
Note. S = Schizophrenia. H = Homicide. SH = Schizophrenia-Homicide. HPC =
Homicide psychiatric controls.
38
= .018) and schizophrenic murderers ( χ
2
= 4.14, df = 1, p = .042); and by increased
proportions of head injury which approached significance in comparison to normal
controls ( χ
2
= 3.72, df = 1, p = .054). However, a one-way ANOVA and a subsequent
percentile bootstrap method for 20% trimmed means (Benjamini-Hochberg method)
both revealed no significant group differences in the length of hospitalization for brain
injury.
Neuropsychology
IQ. One-way ANOVAs were employed to assess group differences in Verbal,
Performance, and Full Scale IQ scores. Results are listed in Table 3. Post hoc
(Bonferroni) analyses revealed that schizophrenic non-murderers, non-schizophrenic
murderers, and schizophrenic murderers were characterized by significantly reduced
scores in comparison to normal controls for Verbal and Full Scale IQ scores; and that all
groups were characterized by significantly reduced scores in comparison to normal
controls for Performance IQ scores. A subsequent percentile bootstrap method for 20%
trimmed means (Benjamini-Hochberg method) indicated that all groups were
characterized by significantly reduced scores in comparison to normal controls for
Verbal, Performance, and Full Scale IQ (see Table 3). To control for potential effects of
head injury upon IQ scores, subsequent ANOVAs and modern methods were employed
on IQ scores with those participants reporting a history of head injury removed from
analyses. Results were identical to those listed in Table 3.
Furthermore, given the potential for malingered symptomatology in forensic
populations (Hall, Thompson, & Poirier, 2007; Maloney, 1985), FSIQ scores were
39
40
subsequently assessed for validity. This was conducted by employing correlational
analyses to verify the presence among groups of significant positive correlations
between FSIQ scores and educational attainment, which is known to account for
sizeable percentages of performance variance on the WAIS-R (Lezak, Howieson,
Loring, Hannay, & Fischer, 2004). Results indicated significant, positive correlations
between FSIQ scores and years of education for normal controls and schizophrenic non-
murderers, and nonsignificant positive correlations for all homicide groups (i.e., non-
schizophrenic murderers, schizophrenic murderers, and homicide psychiatric controls).
These results remained consistent with head-injured participants removed from
correlational analyses—though correlations differed slightly in magnitude and became
significant for homicide psychiatric controls (see Table 3).
CTT. A series of one-way ANOVAs was employed to assess group differences
across all CTT variable T scores (see Table 4). Results indicated no significant group
differences. A subsequent bootstrap method of trimmed mean comparison (method SR;
Wilcox, 2005) revealed that schizophrenic murderers were characterized by significantly
increased CTT-2 number error T scores (i.e., committed fewer CTT-2 number errors) in
comparison to non-schizophrenic murderers (p = .005). No other significant group
differences were observed using modern methods, though method SR revealed that
schizophrenic murderers were characterized by increased CTT-2 near-misses T scores
which approached significance in comparison to schizophrenic non-murderers (p = .012,
p-critical = .010).
41
42
To control for potential effects of head injury upon CTT T scores, subsequent
ANOVAs and modern methods were employed on all CTT variable T scores with those
participants reporting a history of head injury removed from analyses. Results from both
conventional and modern analyses revealed no significant group performance
differences on all CTT variables, including CTT2 number errors and near misses T
scores. To further assess potential brain injury effects, independent-samples T tests were
employed to compare CTT T scores among participants across the entire sample
reporting head injury history versus those who did not. Results indicated no significant
group differences, and a Yuen-Welch method for comparing trimmed means revealed
congruent results. An additional independent-samples T tests was employed to compare
CTT2 number error T scores among head-injury versus head-injury-free participants
within the non-schizophrenic murderer group specifically. Results indicated no
significant group differences, and a Yuen-Welch method for comparing trimmed means
revealed congruent results.
Gambling Task. For analysis, the 100 trials were divided into five blocks of 20
trials each. For each block, the number of disadvantageous selections (decks A and B)
was calculated (see Table 5 and Figure 1). A repeated-measures multiple analysis of
variance (MANOVA) with one between-subject factor (GROUP with five levels) and
one within-subject factor (BLOCK with five levels) was employed to first assess for
overall differences in group and block performances, and for group-by-block interaction.
Violations of sphericity were detected using Mauchly’s test of sphericity, thus
significance levels were adjusted. Results indicated no significant group or block
43
Table 5. Gambling Task Results
Diagnostic groups
Block
N
M (SD)
S
M (SD)
H
M (SD)
SH
M (SD)
HPC
M (SD)
Block 1 10.12
(2.39)
9.52
(4.12)
11.42
(4.06)
9.96
(2.18)
8.40
(3.10)
Block 2 9.69
(2.93)
9.21
(3.75)
10.62
(4.43)
9.88
(2.58)
10.80
(3.91)
Block 3 10.14
(3.74)
9.21
(4.06)
11.54
(3.43)
8.77
(3.95)
10.90
(4.07)
Block 4 9.76
(3.83)
9.61
(4.44)
9.15
(4.35)
10.15
(3.95)
11.40
(4.45)
Block 5 9.93
(4.22)
8.91
(3.98)
10.27
(4.82)
11.08
(4.32)
8.60
(2.32)
Note. N = normal controls. S = schizophrenia. H = homicide. SH = schizophrenia-
homicide. HPC = homicide psychiatric controls.
44
Figure 1. Gambling Task Results
8
9
10
11
12
Block 1 Block 2 Block 3 Block 4 Block 5
Blocks
Disadvantageous Deck Selections
N
S
H
SH
HPC
45
differences, though a marginally-significant group-block interaction was revealed
(F(12.80) = 1.75, p = .050). However, this interaction only approached significance
when using the more-conservative Greenhouse-Geisser correction (F(12.10) = 1.75, p =
.054). Subsequent one-way ANOVAs conducted on scores for each block revealed
group differences which approached significance only for block 3 scores (F(4) = 2.22, p
= .070). Results from modern methods confirmed those of conventional analyses—no
significant group differences among individual block scores were revealed.
To control for potential effects of head injury upon Gambling Task performance,
the aforementioned procedures were repeated with participants reporting a history of
head injury removed from analyses. Similar to the previous results, a repeated-measures
ANOVA revealed no significant group or block differences; though a significant group-
block interaction was not indicated when applying Greenhouse-Geisser correction. To
further assess for potential brain injury effects, these procedures were employed to
compare Gambling Task performance among participants reporting head injury history
versus those who did not. Results indicated no significant differences in Gambling Task
performance between head-injured and non-head-injured participants.
EEG
Tables 6-12 list resting EEG amplitudes at individual electrode sites in each
power band. To avoid Type I error in conducting multiple analyses of group differences
across individual electrode sites and power bands, a series of repeated-measures
MANOVAs with one between-subject factor (GROUP with five levels) and two within-
subject factors (HEMISPHERE with two levels and LOBE with six levels) was
46
Table 6. Resting EEG Amplitudes at Individual Electrode Sites: Delta Band (0.01- 4.00 Hz)
Diagnostic groups
Electrode
site
N
(n = 46)
M (SD)
S
(n = 30)
M (SD)
H
(n = 29)
M (SD)
SH
(n = 30)
M (SD)
HPC
(n = 14)
M (SD)
F
z
7.14
(10.78)
10.78
(7.59)
8.31
(3.83)
10.67
(15.71)
6.61
(1.68)
C
z
7.95
(3.55)
10.47
(6.97)
7.74
(4.03)
10.70
(17.55)
7.40
(2.89)
P
z
6.83
(2.78)
11.43
(9.02)
7.58
(4.76)
10.40
(16.11)
7.25
(2.19)
O
z
6.45
(9.24)
9.38
(9.50)
6.00
(4.67)
9.05
(15.91)
5.29
(1.88)
Fp
1
10.75
(6.02)
15.05
(10.45)
12.03
(6.91)
14.67
(17.41)
9.45
(3.26)
Fp
2
9.96
(6.42)
15.30
(9.86)
10.92
(5.34)
14.43
(16.51)
10.34
(4.07)
F
3
7.64
(2.84)
9.85
(6.16)
8.38
(4.27)
10.52
(15.98)
6.32
(1.11)
F
4
7.12
(4.11)
10.95
(8.27)
8.07
(4.61)
10.59
(13.87)
7.00
(2.40)
F
7
8.13
(4.48)
9.10
(6.07)
8.58
(4.80)
10.84
(15.85)
6.24
(1.92)
F
8
7.36
(8.29)
10.97
(8.29)
7.12
(3.84)
10.23
(13.22)
6.86
(2.68)
T
3
5.17
(6.92)
5.37
(3.67)
5.21
(3.46)
7.58
(16.81)
5.04
(3.07)
T
4
4.44
(3.59)
7.29
(5.36)
4.93
(5.51)
7.18
(14.71)
4.75
(3.36)
P
3
6.24
(2.79)
9.83
(10.33)
6.27
(4.28)
9.12
(15.13)
6.70
(2.47)
P
4
6.83
(4.39)
10.60
(8.81)
7.06
(5.51)
9.88
(16.60)
6.23
(2.59)
O
1
5.36
(5.17)
8.72
(9.88)
5.34
(4.96)
8.68
(16.19)
4.73
(2.39)
O
2
7.23
(16.56)
9.10
(10.05)
5.96
(5.35)
8.97
(16.17)
4.79
(1.81)
Note. N = normal controls. S = schizophrenia. H = homicide. SM = schizophrenia-homicide.
HPC = homicide psychiatric controls.
47
Table 7. Resting EEG Amplitudes at Individual Electrode Sites: Theta Band (4.01- 8.00 Hz)
Diagnostic groups
Electrode
site
N
(n = 46)
M (SD)
S
(n = 31)
M (SD)
H
(n = 29)
M (SD)
SH
(n = 30)
M (SD)
HPC
(n = 14)
M (SD)
F
z
6.99
(4.05)
16.28
(14.22)
7.46
(4.17)
10.97
(12.65)
7.50
(2.10)
C
z
7.12
(3.66)
13.69
(10.68)
6.48
(3.92)
10.27
(12.19)
7.43
(2.36)
P
z
5.77
(3.00)
12.32
(10.59)
5.74
(3.71)
9.65
(13.48)
6.25
(2.66)
O
z
4.29
(2.31)
10.05
(10.40)
4.06
(2.94)
8.64
(14.54)
4.79
(2.34)
Fp
1
5.48
(2.64)
12.78
(11.05)
6.22
(2.94)
8.56
(10.03)
6.30
(1.46)
Fp
2
5.37
(2.79)
13.58
(13.29)
6.18
(3.05)
8.51
(9.93)
6.85
(1.83)
F
3
6.10
(3.14)
12.74
(9.50)
6.47
(3.86)
9.30
(10.89)
6.30
(1.75)
F
4
6.13
(3.57)
14.10
(13.83)
6.38
(3.30)
9.30
(10.74)
7.08
(1.71)
F
7
4.28
(2.54)
7.28
(4.97)
4.47
(2.33)
6.02
(9.34)
4.69
(1.81)
F
8
4.08
(2.12)
8.79
(8.30)
4.43
(2.08)
6.07
(8.96)
4.66
(0.99)
T
3
3.13
(2.99)
4.77
(3.43)
3.34
(2.19)
4.59
(8.30)
3.24
(1.12)
T
4
3.21
(3.08)
5.67
(4.51)
2.91
(1.70)
4.65
(8.44)
2.93
(0.87)
P
3
5.28
(3.44)
10.67
(8.99)
4.82
(3.33)
8.33
(12.55)
5.33
(2.14)
P
4
5.66
(3.95)
11.91
(10.49)
5.26
(4.01)
8.63
(11.78)
5.05
(1.69)
O
1
3.82
(2.20)
9.00
(10.15)
3.49
(2.74)
8.30
(14.68)
4.53
(2.55)
O
2
3.97
(2.33)
10.09
(12.32)
3.85
(3.49)
8.62
(15.79)
4.37
(2.60)
Note. N = normal controls. S = schizophrenia. H = homicide. SH = schizophrenia-homicide.
HPC = homicide psychiatric controls.
48
Table 8. Resting EEG Amplitudes at Individual Electrode Sites: Alpha1 Band (8.01 - 10.50
Hz)
Diagnostic groups
Electrode
site
N
(n = 46)
M (SD)
S
(n = 31)
M (SD)
H
(n = 29)
M (SD)
SH
(n = 30)
M (SD)
HPC
(n = 14)
M (SD)
F
z
10.47
(8.75)
13.82
(10.63)
6.77
(6.48)
11.73
(14.96)
7.63
(6.49)
C
z
11.36
(9.10)
13.09
(9.62)
7.25
(7.79)
10.65
(12.62)
8.25
(7.60)
P
z
13.94
(19.78)
14.17
(12.76)
10.90
(14.90)
13.92
(24.81)
10.54
(10.98)
O
z
9.91
(10.67)
12.69
(12.70)
7.74
(9.10)
11.92
(19.60)
9.78
(10.89)
Fp
1
7.23
(6.53)
8.86
(6.50)
4.80
(4.51)
8.42
(11.40)
6.42
(6.55)
Fp
2
7.35
(6.67)
9.29
(7.03)
4.83
(4.67)
8.28
(10.64)
6.64
(6.46)
F
3
8.58
(7.20)
10.30
(6.90)
5.61
(5.24)
9.84
(12.97)
7.23
(6.87)
F
4
8.76
(7.18)
11.40
(9.04)
5.61
(4.94)
9.64
(12.36)
7.92
(6.85)
F
7
4.84
(4.28)
5.32
(3.23)
3.28
(3.17)
5.51
(7.43)
4.16
(3.59)
F
8
5.34
(4.90)
6.35
(4.74)
3.43
(2.92)
5.55
(7.44)
4.43
(3.71)
T
3
2.71
(2.28)
3.61
(2.42)
2.42
(2.45)
2.77
(2.81)
2.65
(2.39)
T
4
2.98
(2.20)
4.16
(2.98)
2.15
(1.80)
3.09
(3.43)
2.74
(2.29)
P
3
12.51
(21.21)
12.91
(12.46)
9.37
(12.77)
12.62
(20.95)
10.41
(11.49)
P
4
15.40
(21.46)
13.87
(14.15)
10.32
(12.45)
13.83
(19.38)
10.78
(10.93)
O
1
11.57
(21.19)
12.49
(14.36)
6.76
(7.70)
11.96
(20.38)
9.11
(10.26)
O
2
12.38
(17.38)
12.21
(12.80)
7.17
(9.30)
10.83
(18.06)
10.35
(13.29)
Note. N = normal controls. S = schizophrenia. H = homicide. SH = schizophrenia-homicide.
HPC = homicide psychiatric controls.
49
Table 9. Resting EEG Amplitudes at Individual Electrode Sites: Alpha2 Band (10.51 - 13.00
Hz)
Diagnostic groups
Electrode
site
N
(n = 46)
M (SD)
S
(n = 31)
M (SD)
H
(n = 29)
M (SD)
SH
(n = 30)
M (SD)
HPC
(n = 14)
M (SD)
F
z
7.73
(7.19)
6.96
(5.55)
5.84
(5.58)
6.73
(9.71)
4.21
(3.95)
C
z
10.59
(9.94)
8.61
(7.28)
7.21
(7.82)
7.29
(9.67)
5.30
(4.97)
P
z
16.04
(20.72)
13.28
(12.00)
14.65
(17.74)
13.94
(31.03)
8.52
(7.38)
O
z
13.73
(17.98)
11.64
(11.68)
10.89
(12.74)
12.36
(25.04)
7.29
(6.23)
Fp
1
5.37
(5.44)
4.88
(4.02)
4.19
(4.10)
4.92
(7.11)
3.23
(3.03)
Fp
2
5.47
(5.54)
5.03
(4.12)
4.21
(4.03)
4.93
(7.28)
3.26
(3.06)
F
3
6.50
(6.23)
5.51
(4.15)
4.89
(4.77)
5.65
(7.86)
3.74
(3.63)
F
4
6.56
(5.95)
5.85
(4.85)
4.81
(4.13)
5.87
(9.21)
3.89
(3.62)
F
7
3.66
(3.74)
3.03
(2.28)
2.89
(3.08)
3.33
(4.90)
2.29
(1.97)
F
8
4.01
(3.82)
3.57
(2.92)
2.93
(2.49)
3.52
(5.27)
2.26
(1.98)
T
3
2.21
(1.80)
2.37
(1.82)
2.02
(2.06)
1.92
(2.15)
1.66
(1.16)
T
4
2.52
(1.82)
2.66
(2.05)
1.93
(1.54)
2.11
(2.39)
1.66
(1.18)
P
3
14.42
(19.72)
12.20
(13.59)
12.75
(16.23)
12.66
(25.69)
8.79
(8.95)
P
4
19.12
(29.51)
13.04
(14.23)
14.96
(17.60)
13.62
(21.06)
10.34
(9.25)
O
1
12.59
(19.48)
11.37
(12.59)
9.41
(10.81)
12.09
(24.38)
6.75
(6.32)
O
2
13.88
(18.06)
11.11
(12.26)
10.70
(15.74)
12.07
(23.92)
7.25
(6.30)
Note. N = normal controls. S = schizophrenia. H = homicide. SH = schizophrenia-homicide.
HPC = homicide psychiatric controls.
50
Table 10. Resting EEG Amplitudes at Individual Electrode Sites: Beta1 Band (13.01 - 20.00
Hz)
Diagnostic groups
Electrode
site
N
(n = 45)
M (SD)
S
(n = 32)
M (SD)
H
(n = 29)
M (SD)
SH
(n = 30)
M (SD)
HPC
(n = 14)
M (SD)
F
z
4.35
(4.02)
5.13
(2.77)
3.56
(2.22)
3.85
(5.24)
3.92
(1.88)
C
z
6.48
(5.72)
6.73
(4.48)
4.51
(2.72)
4.78
(6.35)
5.07
(2.91)
P
z
8.88
(8.98)
9.75
(8.72)
8.66
(7.32)
7.40
(14.35)
7.88
(6.00)
O
z
6.73
(5.75)
7.47
(6.77)
6.75
(5.88)
6.40
(10.98)
5.93
(3.07)
Fp
1
3.06
(2.52)
3.34
(1.99)
2.68
(1.62)
3.09
(4.41)
2.85
(0.96)
Fp
2
3.02
(2.38)
3.44
(2.10)
2.70
(1.60)
2.93
(3.80)
2.90
(1.11)
F
3
3.91
(4.12)
4.40
(2.24)
3.28
(2.11)
3.59
(5.27)
3.31
(1.26)
F
4
3.93
(3.64)
4.48
(2.58)
3.17
(2.00)
3.36
(4.02)
4.03
(2.11)
F
7
2.25
(1.84)
2.46
(1.24)
2.16
(1.47)
2.29
(3.08)
2.50
(1.08)
F
8
2.41
(1.83)
2.74
(1.52)
2.18
(1.42)
2.25
(2.77)
2.62
(1.44)
T
3
2.25
(1.81)
2.24
(1.32)
1.88
(1.11)
2.11
(2.10)
2.61
(1.52)
T
4
2.21
(1.51)
2.66
(1.68)
1.91
(1.17)
2.02
(2.20)
2.46
(1.65)
P
3
10.87
(18.03)
8.55
(8.09)
8.13
(8.53)
6.44
(11.45)
7.19
(4.32)
P
4
10.79
(17.81)
8.63
(7.60)
9.00
(8.13)
6.82
(9.69)
8.12
(2.17)
O
1
5.99
(5.40)
6.91
(6.20)
5.93
(5.04)
5.97
(10.14)
5.86
(3.34)
O
2
6.24
(6.07)
6.93
(7.44)
6.51
(7.15)
6.39
(10.58)
5.22
(2.94)
Note. N = normal controls. S = schizophrenia. H = homicide. SH = schizophrenia-homicide.
HPC = homicide psychiatric controls.
51
Table 11. Resting EEG Amplitudes at Individual Electrode Sites: Beta2 Band (20.01 - 30.00
Hz)
Diagnostic groups
Electrode
site
N
(n = 46)
M (SD)
S
(n = 31)
M (SD)
H
(n = 29)
M (SD)
SH
(n = 30)
M (SD)
HPC
(n = 14)
M (SD)
F
z
2.00
(1.28)
3.16
(3.12)
1.65
(0.87)
2.18
(2.44)
3.02
(2.37)
C
z
2.38
(1.45)
3.26
(2.23)
1.90
(1.10)
2.33
(2.45)
3.39
(2.09)
P
z
2.03
(1.57)
2.64
(1.83)
2.02
(1.24)
2.28
(3.19)
3.43
(2.21)
O
z
2.09
(1.59)
2.30
(1.43)
1.85
(1.25)
2.37
(3.36)
2.66
(1.51)
Fp
1
1.78
(1.37)
2.10
(2.17)
1.28
(0.59)
2.24
(4.07)
3.12
(2.88)
Fp
2
1.59
(0.94)
2.17
(2.58)
1.30
(0.68)
1.90
(2.58)
2.39
(1.73)
F
3
1.84
(1.24)
2.65
(2.45)
1.54
(0.81)
2.38
(3.90)
3.18
(2.99)
F
4
1.85
(1.18)
2.60
(2.35)
1.54
(0.86)
1.97
(2.25)
3.63
(3.06)
F
7
1.14
(0.66)
1.63
(1.78)
1.13
(0.64)
1.84
(3.36)
2.46
(3.18)
F
8
1.29
(0.67)
1.80
(1.99)
1.18
(0.79)
1.40
(1.67)
2.61
(2.83)
T
3
1.18
(0.68)
1.24
(0.69)
1.24
(0.98)
1.42
(1.31)
2.17
(2.47)
T
4
1.31
(0.93)
1.46
(1.10)
1.24
(1.21)
1.19
(1.08)
2.45
(3.45)
P
3
2.39
(1.83)
2.50
(1.72)
1.97
(1.26)
2.23
(3.17)
3.37
(2.79)
P
4
2.70
(2.84)
2.43
(1.65)
2.28
(1.55)
2.25
(2.95)
3.07
(1.87)
O
1
2.00
(1.85)
2.26
(1.47)
1.71
(1.14)
2.44
(3.57)
2.42
(1.31)
O
2
2.04
(1.72)
2.14
(1.64)
1.93
(1.76)
2.36
(3.33)
2.49
(1.43)
Note. N = normal controls. S = schizophrenia. H = homicide. SH = schizophrenia-homicide.
HPC = homicide psychiatric controls.
52
Table 12. Resting EEG Amplitudes at Individual Electrode Sites: Beta3 Band (30.01 - 50.00
Hz)
Diagnostic groups
Electrode
site
N
(n = 46)
M (SD)
S
(n = 31)
M (SD)
H
(n = 29)
M (SD)
SH
(n = 30)
M (SD)
HPC
(n = 14)
M (SD)
F
z
2.76
(3.36)
1.58
(1.46)
2.61
(3.14)
3.73
(5.35)
5.04
(10.06)
C
z
5.23
(20.26)
1.79
(2.48)
2.71
(3.50)
4.83
(6.67)
4.66
(9.69)
P
z
3.47
(9.74)
1.18
(0.73)
2.38
(3.20)
3.90
(5.53)
4.18
(8.54)
O
z
2.96
(3.71)
1.99
(1.61)
3.83
(5.51)
5.62
(7.91)
5.53
(10.45)
Fp
1
3.16
(3.80)
1.25
(0.79)
2.73
(3.36)
5.21
(10.54)
5.62
(11.83)
Fp
2
3.18
(4.12)
1.40
(1.17)
3.31
(4.05)
4.40
(6.48)
4.94
(9.53)
F
3
2.50
(3.35)
1.18
(0.69)
2.55
(3.36)
5.07
(8.43)
4.51
(8.98)
F
4
2.33
(3.06)
1.05
(0.52)
2.35
(2.97)
3.04
(3.59)
4.53
(8.12)
F
7
2.25
(3.01)
1.20
(1.43)
3.13
(4.02)
4.97
(7.96)
4.68
(9.70)
F
8
3.46
(8.38)
1.06
(0.92)
4.13
(8.62)
3.48
(4.31)
4.15
(9.06)
T
3
2.87
(3.42)
1.32
(1.15)
2.85
(3.73)
4.73
(6.88)
5.38
(9.44)
T
4
2.68
(3.09)
2.12
(2.82)
3.00
(4.09)
3.11
(3.16)
5.76
(9.59)
P
3
2.53
(3.27)
1.22
(0.79)
2.54
(3.54)
5.75
(8.18)
4.70
(9.74)
P
4
2.53
(3.39)
1.12
(0.66)
4.38
(8.84)
3.76
(4.89)
4.66
(7.25)
O
1
2.81
(3.16)
2.30
(2.27)
4.60
(9.56)
5.92
(9.09)
5.14
(10.37)
O
2
3.01
(3.67)
2.18
(2.17)
3.67
(4.62)
5.97
(8.37)
6.56
(11.73)
Note. N = normal controls. S = schizophrenia. H = homicide. SH = schizophrenia-homicide.
HPC = homicide psychiatric controls.
53
employed to first assess for group-by-hemisphere, group-by-lobe, and group-by-
hemisphere-by-lobe interactions in each band. Significance levels were adjusted using
the more-conservative Greenhouse-Geisser correction if the assumption of sphericity
was violated. If significant interactions were detected, secondary one-way ANOVAs
along with post-hoc multiple comparisons (Bonferroni) were subsequently conducted on
individual hemisphere, lobe, or hemisphere-lobe site groupings. Results from
MANOVAs indicated no significant interactions in the delta, alpha1, alpha2, or beta1
bands.
Theta. A repeated measures MANOVA revealed significant group-lobe (Pillai’s
Trace F(20) = 1.92, p = .010; Greenhouse-Geisser correction F(8.08) = 2.17, p = .029)
and group-hemisphere (Pillai’s Trace F(4) = 2.81, p = .028; Greenhouse-Geisser
correction F(4) = 2.81, p = .028) interactions in the theta band. For individual group
hemispheric sites, one-way ANOVAs revealed significant group differences for both the
left (F(4) = 4.26, p = .003) and right (F(4) = 5.16, p = .001) hemispheric sites. Post hoc
(Bonferroni) analyses revealed schizophrenic non-murderers were characterized by
significantly increased amplitudes in comparison to normal controls and non-
schizophrenic murderers for both left (schizophrenic non-murderers – normal controls p
= .004, schizophrenic non-murderers – non-schizophrenic murderers p = .020) and right
(schizophrenic non-murderers – normal controls p = .001, schizophrenic non-murderers
– non-schizophrenic murderers p = .005) hemispheric sites. A percentile bootstrap
method for 20% trimmed means (Benjamini-Hochberg method) indicated that
schizophrenic non-murderers were characterized by significantly increased amplitudes
54
in comparison to all other groups for left (schizophrenic non-murderers – normal
controls p < .001, schizophrenic non-murderers – non-schizophrenic murderers p = .002,
schizophrenic non-murderers – schizophrenic murderers p = .011, schizophrenic non-
murderers – homicide psychiatric control p = .011) and right hemispheric sites
(schizophrenic non-murderers – normal controls p < .001, schizophrenic non-murderers
– non-schizophrenic murderers p = .002, schizophrenic non-murderers – schizophrenic
murderers p = .016, schizophrenic non-murderers – homicide psychiatric control p =
.010). Additionally, homicide psychiatric controls were characterized by significantly
increased amplitudes in comparison to normal controls (p = .014) for right hemispheric
sites only.
For individual group lobe sites, one-way ANOVAs revealed significant group
differences for the medial prefrontal (F(4) = 5.91, p < .001), dorsomedial prefrontal
(F(4) = 5.33, p < .001), dorsolateral prefrontal (F(4) = 2.95, p = .022), parietal (F(4) =
4.16, p = .003), and occipital lobe sites (F(4) = 3.19, p = .015). Post hoc (Bonferroni)
analyses revealed schizophrenic non-murderers were characterized by significantly
increased amplitudes in comparison to normal controls and non-schizophrenic murderers
for medial prefrontal (schizophrenic non-murderers – normal controls p < .001,
schizophrenic non-murderers – non-schizophrenic murderers p = .003), dorsomedial
prefrontal (schizophrenic non-murderers – normal controls p < .001, schizophrenic non-
murderers – non-schizophrenic murderers p = .004), and parietal lobe sites
(schizophrenic non-murderers – normal controls p = .008, schizophrenic non-murderers
– non-schizophrenic murderers p = .012). Schizophrenic non-murderers were also
55
characterized by significantly increased amplitudes in comparison to normal controls
for dorsolateral prefrontal lobe sites (p = .020). Additionally, schizophrenic non-
murderers were characterized by increased amplitudes which approached significance in
comparison to homicide psychiatric controls for medial prefrontal (p = .058) and
dorsomedial prefrontal lobe sites (p = .060).
A percentile bootstrap method for 20% trimmed means (Benjamini-Hochberg
method) indicated that for medial prefrontal lobe sites, schizophrenic non-murderers
were characterized by significantly increased amplitudes in comparison to normal
controls, non-schizophrenic murderers, and schizophrenic murderers (schizophrenic
non-murderers – normal controls p < .001, schizophrenic non-murderers – non-
schizophrenic murderers p = .008, schizophrenic non-murderers – schizophrenic
murderers p = .010); and homicide psychiatric controls were characterized by
significantly increased amplitudes in comparison to normal controls (p = .006). For
dorsomedial prefrontal lobe sites, schizophrenic non-murderers were characterized by
significantly increased amplitudes in comparison to all other groups (schizophrenic non-
murderers – normal controls p < .001, schizophrenic non-murderers – non-schizophrenic
murderers p = .002, schizophrenic non-murderers – schizophrenic murderers p = .015,
schizophrenic non-murderers – homicide psychiatric control p = .017). For dorsolateral
prefrontal lobe sites, schizophrenic non-murderers were characterized by significantly
increased amplitudes in comparison to normal controls, non-schizophrenic murderers,
and schizophrenic murderers (schizophrenic non-murderers – normal controls p = .001,
schizophrenic non-murderers – non-schizophrenic murderers p = .009, schizophrenic
56
non-murderers – schizophrenic murderers p = .008). For temporal lobe sites,
schizophrenic non-murderers were characterized by significantly increased amplitudes
in comparison to normal controls (p = .001). For parietal lobe sites, schizophrenic non-
murderers were characterized by significantly increased amplitudes in comparison to all
other groups (schizophrenic non-murderers – normal controls p = .001, schizophrenic
non-murderers – non-schizophrenic murderers p = .001, schizophrenic non-murderers –
schizophrenic murderers p = .018, schizophrenic non-murderers – homicide psychiatric
control p = .005). For occipital lobe sites, schizophrenic non-murderers were
characterized by significantly increased amplitudes in comparison to normal controls
and non-schizophrenic murderers (schizophrenic non-murderers – normal controls p <
.001, schizophrenic non-murderers – non-schizophrenic murderers p = .002).
Beta3. A repeated measures MANOVA revealed a significant group-hemisphere
interaction (Pillai’s Trace F(4) = 3.99, p = .004; Greenhouse-Geisser correction F(4) =
3.99, p = .004) in the beta3 band. Subsequent one-way ANOVAs revealed significant
group differences for only left hemispheric sites (F(4) = 2.98, p = .021). Post hoc
(Bonferroni) analyses revealed schizophrenic murderers were characterized by
significantly increased amplitudes in comparison to schizophrenic non-murderers for left
hemispheric sites (p = .025). A percentile bootstrap method for 20% trimmed means
(Benjamini-Hochberg method) yielded consistent results (p = .005).
Antipsychotic medication use, head injury, and EEG. Subsequent analyses were
conducted to assess for potential effects of antipsychotic medication use and head injury
history upon group differences in EEG functioning. As antipsychotic medication status
57
and head injury data were not available for all participants (i.e., n = 148 or 91.4% of
the sample for antipsychotics use, and n = 144 or 88.9% of the sample for head injury),
separate analyses controlling these factors were not employed in order to avoid the
potential obfuscation of results due to reduced statistical power. Antipsychotic
medication data were additionally problematic as positive medication status
characterized the entire schizophrenic non-murderer group, and did not—for the most
part (i.e., 70.4%) characterize schizophrenic murderers. As such antipsychotics use
became inextricably confounded with diagnostic group membership status for
schizophrenic non-murderers, and analyses controlling for antipsychotics usage could
not be effectively performed.
To address this very important issue, group comparisons between participants
using and not using antipsychotic medication across the entire sample were performed,
with the notion that significant amplitude differences between antipsychotic-free and
antipsychotic-medicated participants across the entire sample would indicate that
medication effects (as opposed to group effects) were driving the aforementioned group
amplitude differences. However, if no significant differences were observed, then EEG
results would likely be attributable more so to diagnostic group differences. These
subsequent analyses (along with similar analyses comparing head-injured and non-head-
injured participants to address the potential effects of head injury history upon EEG
amplitudes) were conducted in the theta and beta3 power bands for hemispheric, lobe, or
individual electrode sites where significant group differences were revealed above.
58
A series of independent-samples T tests along with modern methods (i.e.,
analyses utilizing a percentile bootstrap method comparing modified M-estimators;
Wilcox, 2005) were employed to assess theta amplitude group differences between
antipsychotic-free and antipsychotic-using participants for left hemispheric, right
hemispheric, and all individual group lobe sites; and beta3 amplitude group differences
between antipsychotic-free and antipsychotic-using participants for left hemispheric
sites. Results are listed in Table 13, and indicated significant group differences only for
parietal and occipital lobe sites in the theta band (though group differences approached
significance for left hemispheric, right hemispheric, and temporal lobe sites in this
power band). Resting EEG amplitudes did not differ significantly between
antipsychotic-free and antipsychotic-using participants for left hemispheric sites in the
beta3 band. A similar series of analyses comparing head-injured and non-head injured
participants across the entire sample revealed no significant group differences in theta or
beta3 amplitudes for the aforementioned sites (see Table 14).
Psychosocial Functioning
Unemployment. A one-way ANOVA was subsequently conducted to assess
group differences in length of unemployment (in months) within the past year. Results
indicated significant group differences (see Table 15). Post hoc (Bonferroni) analyses
revealed schizophrenic non-murderers were characterized by significantly longer
unemployment than both normal controls (p < .001) and non-schizophrenic murderers (p
= .005). A percentile bootstrap method for 20% trimmed means (Benjamini-Hochberg
method) yielded consistent results (ps < .001 and = .008, respectively).
59
60
61
Homelessness. A series of Chi-Square analyses was employed to assess
differences among groups in the presence of homelessness within the past year (see
Table 15). Chi-Square analyses using all groups revealed no significant group
differences ( χ
2
= 7.13, df = 4, p = .129). Subsequent Chi-Square analyses conducted on
individual group comparisons indicated that schizophrenic murderers were characterized
by significantly increased proportions of recent homelessness relative to normal controls
( χ
2
= 5.59, df = 1, p = .018); and schizophrenic murderers were characterized by
increased proportions of recent homelessness which approached significance relative to
schizophrenic non-murderers ( χ
2
= 3.10, df = 1, p = .078).
Number of couple relationships. A one-way ANOVA was conducted to assess
group differences in the number of couple relationships. Results indicated no significant
group differences (see Table 15). A percentile bootstrap method for 20% trimmed means
(Benjamini-Hochberg method) confirmed there were no significant group differences.
Number of close friends. A one-way ANOVA was employed to assess group
differences in number of close friends. Results indicated significant group differences
(see Table 15), though post hoc (Bonferroni) analyses revealed these differences were
not significant when controlling for Type I error. A percentile bootstrap method for 20%
trimmed means (Benjamini-Hochberg method) indicated that normal controls were
characterized by significantly more close friends than non-schizophrenic murderers (p <
.001) and homicide psychiatric controls (p = .004), and that schizophrenic non-
murderers were characterized by significantly more close friends than non-schizophrenic
murderers (p = .012).
62
Emotional closeness to family. A one-way ANOVA was conducted to assess
group differences in perceived emotional closeness to family members. Results
indicated significant group differences (see Table 15). Post hoc (Bonferroni) analyses
revealed that normal controls reported significantly increased perceived emotional
closeness to family members in comparison to non-schizophrenic murderers (p = .003),
schizophrenic murderers (p = .035), and homicide psychiatric controls (p = .006). A
percentile bootstrap method for 20% trimmed means (Benjamini-Hochberg method)
indicated that normal controls reported significantly increased perceived emotional
closeness to family members in comparison to non-schizophrenic murderers (p = .003),
schizophrenic murderers (p = .008), and homicide psychiatric controls (p = .002).
Problems with place or area of residence. A one-way ANOVA was conducted
to assess group differences in the number of problems with area or place of residence.
Results indicated no significant group differences (see Table 15). A percentile bootstrap
method for 20% trimmed means (Benjamini-Hochberg method) confirmed there were no
significant group differences.
Marital status. A series of Chi-Square analyses was employed to assess
differences among groups in current marital status (see Table 15). Chi-Square analyses
using all groups revealed significant group differences in proportions of currently
married participants ( χ
2
= 16.07, df = 4, p = .003). Subsequent Chi-Square analyses
conducted on individual group comparisons indicated that normal controls were
characterized by significantly increased proportions of currently married participants in
comparison to schizophrenic non-murderers ( χ
2
= 10.37, df = 1, p = .001) and non-
63
64
65
schizophrenic murderers ( χ
2
= 6.63, df = 1, p = .010), as were schizophrenic
murderers ( χ
2
= 8.41, df = 1, p = .004 for schizophrenic non-murderers, and χ
2
= 5.51, df
= 1, p = .019 for non-schizophrenic murderers).
An additional series of Chi-Square analyses was employed to assess differences
among groups in marital history (see Table 15). Chi-Square analyses using all groups
revealed group differences which approached significance in proportions of participants
who had never been married ( χ
2
= 9.09, df = 4, p = .059). Subsequent Chi-Square
analyses conducted on individual group comparisons indicated that schizophrenic non-
murderers were characterized by significantly increased proportions of participants with
no marital history in comparison to normal controls ( χ
2
= 6.93, df = 1, p = .008) and
schizophrenic murderers ( χ
2
= 5.47, df = 1, p = .019).
Living alone. A series of Chi-Square analyses was employed to assess
differences among groups in proportions of individuals currently living alone (see Table
15). Chi-Square analyses using all groups revealed significant group differences in
proportions of participants who were living alone ( χ
2
= 18.86, df = 4, p = .001).
Subsequent Chi-Square analyses conducted on individual group comparisons indicated
that normal controls were characterized by significantly increased proportions of
participants currently living alone in comparison to schizophrenic non-murderers ( χ
2
=
16.64, df = 1, p < .001) and homicide psychiatric controls ( χ
2
= 4.43, df = 1, p = .035).
Additionally, both non-schizophrenic and schizophrenic murderers were characterized
by significantly increased proportions of participants currently living alone in
66
comparison to schizophrenic non-murderers ( χ
2
= 4.09, df = 1, p = .043, and χ
2
= 8.80,
df = 1, p = .003, respectively).
Childhood Psychosocial Deprivation
A one-way ANOVA was employed to assess group differences in total number
of childhood psychosocial deprivation factors. Results are listed in Table 15, and
revealed group differences which approached significance. A percentile bootstrap
method for 20% trimmed means (Benjamini-Hochberg method) indicated that
schizophrenic murderers were characterized by significantly higher total numbers of
different childhood psychosocial deprivation factors in comparison to normal controls (p
= .005).
Birth order. A one-way ANOVA was employed to assess group differences in
birth order among diagnostic groups (see Table 15). Results indicated significant group
differences (F(4) = 4.27, p = .003). Post hoc (Bonferroni) analyses revealed
schizophrenic murderers were characterized by significantly higher birth order in
comparison to schizophrenic non-murderers (p = .001), and that normal controls were
characterized by higher birth orders which approached significance in comparison to
schizophrenic non-murderers (p = .068). A percentile bootstrap method for 20%
trimmed means (Benjamini-Hochberg method) indicated that schizophrenic murderers
were characterized by significantly higher birth order in comparison to both
schizophrenic non-murderers (p < .001) and non-schizophrenic murderers (p = .014),
and that that normal controls were characterized by higher birth order which approached
significance in comparison to schizophrenic non-murderers (p = .010, p critical = .010).
67
Forensic Clinical Correlates
Motivation. A series of Chi-Square analyses was employed to assess differences
among murderer groups in homicides motivated by love, thuggee ( a historical and
culture-specific term for robbery-motivated homicide—see Wagner, 2007—likely
reflecting here homicide associated with property crime and possible cult involvement),
anger, retaliation, or for reasons unknown to the murderer (see Table 16). Chi-Square
analyses using all murderer groups revealed no significant group differences in homicide
motivation ( χ
2
= 13.57, df = 8, p = .094). Subsequent Chi-Square analyses conducted on
individual group comparisons indicated that schizophrenic murderers were characterized
by significantly increased proportions of unknown motivations in comparison to non-
schizophrenic murderers, while those of non-schizophrenic murderers appeared to be
motivated more by anger ( χ
2
= 11.40, df = 4, p = .022). No significant group differences
in motivation were revealed between either non-schizophrenic murderers or
schizophrenic murderers and homicide psychiatric controls ( χ
2
= 5.79, df = 4, p = .215,
and χ
2
= 4.50, df = 3, p = .212, respectively).
Offense planning. A series of Chi-Square analyses was employed to assess
differences among murderer groups in the presence of homicide planning (see Table 16).
Chi-Square analyses using all groups revealed no group differences ( χ
2
= 4.17, df = 2, p
= .124). Subsequent Chi-Square analyses conducted on individual group comparisons
indicated that non-schizophrenic murderers were characterized by significantly
increased proportions of homicide planning in comparison to schizophrenic murderers
( χ
2
= 4.13, df = 1, p = .042). No significant differences were revealed between
68
Table 16. Offense Characteristics and Concomitants
Homicide groups
H SH HPC
Motivation
Love
1
(3.3%)
0
(0.0%)
0
(0.0%)
Thuggee
2
(6.7%)
0
(0.0%)
1
(7.1%)
Anger
12
(40.0%)
6
(19.4%)
5
(35.7%)
Retaliation
11
(36.7%)
10
(32.3%)
2
(14.3%)
Unknown to
murderer
4
(13.3%)
15
(48.4%)
6
(42.9%)
Offense planning
19
(61.3%)
11
(35.5%)
5
(45.5%)
Concomitant sexual
offense
5
(16.7%)
0
(0.0%)
1
(8.3%)
Note. H = Homicide. SH = Schizophrenia-Homicide. HPC = Homicide psychiatric
controls.
69
schizophrenic murderers or non-schizophrenic murderers and homicide psychiatric
controls ( χ
2
= 0.34, df = 1, p = .559, and χ
2
= 0.83, df = 1, p = .362, respectively).
Concomitant sexual offenses. A series of Chi-Square analyses was employed to
assess differences among groups in the presence of homicides with concomitant sexual
offenses among murderer groups (see Table 16). Chi-Square analyses using all groups
revealed group differences which approached significance in proportions of concomitant
sexual offenses ( χ
2
= 5.77, df = 4, p = .056). Subsequent Chi-Square analyses conducted
on individual group comparisons indicated that non-schizophrenic murderers were
characterized by significantly increased proportions of concomitant sexual offenses in
comparison to schizophrenic murderers ( χ
2
= 5.80, df = 1, p = .016). No significant
differences were revealed between schizophrenic murderers or non-schizophrenic
murderers and homicide psychiatric controls ( χ
2
= 2.73, df = 1, p = .099, and χ
2
= 0.49,
df = 1, p = .486, respectively).
Victimology. Data indicated that seven participants (three non-schizophrenic
murderers and four schizophrenic murderers) had multiple victims. For the purposes of
simplification, the victimological characteristics of the first victim recorded were used
for the present analyses. A series of Chi-Square analyses was employed to assess
differences among groups in offender-victim relationships (see Table 17). Chi-Square
analyses using all groups revealed group differences which approached significance in
types of offender-victim relationships ( χ
2
= 11.36, df = 6, p = .078). Subsequent Chi-
Square analyses conducted on individual group comparisons indicated that
schizophrenic murderers were characterized by significantly increased proportions of
70
Table 17. Victimology
Homicide groups
H SH HPC
Relationship
Stranger
4
(13.3%)
3
(9.7%)
1
(7.7%)
Acquaintance/friend
12
(40.0%)
12
(38.7%)
7
(53.8%)
Family/relative
1
(3.3%)
10
(32.3%)
2
(15.4%)
Romantic partner/spouse
13
(43.3%)
6
(19.4%)
3
(23.1%)
Gender
Male
19
(63.3%)
18
(56.2%)
8
(61.5%)
Female
11
(36.7%)
14
(43.8%)
5
(38.5%)
Gender relationship
Same gender
11
(36.7%)
17
(53.1%)
7
(53.8%)
Opposite gender
19
(63.3%)
15
(46.9%)
6
(46.2%)
Note. H = Homicide. SH = Schizophrenia-Homicide. HPC = Homicide psychiatric
controls.
71
family/relative victims compared to non-schizophrenic murderers, whereas non-
schizophrenic murderers were characterized by significantly increased proportions of
romantic partner/spouse victims ( χ
2
= 10.07, df = 3, p = .018). No significant differences
were revealed between schizophrenic murderers or non-schizophrenic murderers and
homicide psychiatric controls ( χ
2
= 1.54, df = 3, p = .672, and χ
2
= 3.53, df = 3, p = .317,
respectively).
An additional series of Chi-Square analyses was employed to assess differences
among groups in victim gender (see Table 17). Chi-Square analyses using all groups
revealed no significant group differences ( χ
2
= 0.34, df = 2, p = .844), and subsequent
Chi-Square analyses conducted on individual group comparisons yielded similar results.
Further Chi-Square analyses were employed to assess differences among groups
in offender-victim gender relationship (i.e., the likelihood of an offender victimizing an
individual of the same or opposite gender; see Table 17). Chi-Square analyses using all
groups revealed no significant group differences ( χ
2
= 2.01, df = 2, p = .366), and
subsequent Chi-Square analyses conducted on individual group comparisons yielded
similar results.
Emotional experience. Affective descriptors characterizing the period of time
before, during, and after the offense for each murderer group are depicted graphically in
Figure 2. Participants among groups did not endorse the full range of possible descriptor
choices across offense time periods, and only the select group of reported descriptors are
included here. In this figure, descriptors reflecting negative affect are represented in
patterns of red, positive affect in patterns of green, and neutral (or no) affect in yellow.
72
73
74
75
76
Chapter 4: Discussion
Hypothesis 1: Intellectual Functioning
The present study’s first aim was to identify biological and psychosocial
indicators of a deficit profile possibly unique to a subgroup of schizophrenic persons
who commit murder. The hypothesis that schizophrenic murderers would be
characterized by significantly reduced intellectual functioning in comparison to
schizophrenic non-murderers, non-schizophrenic murderers, and normal controls
received only partial support. While schizophrenic murderers did demonstrate
significantly reduced Verbal, Performance, and Full Scale IQ scores in comparison to
normal controls, these IQ score reductions were not present when compared to
schizophrenic non-murderers, non-schizophrenic murderers, or homicide psychiatric
controls. In fact, all other groups demonstrated comparable and significant Verbal,
Performance, and Full Scale IQ reductions in comparison to normal controls. Group
differences remained largely identical when controlling for history of head injury.
Furthermore, IQ scores correlated significantly with educational attainment for
schizophrenic non-murderers and normal controls, while those for homicide groups did
not. This may suggest malingering on the part of these individuals—especially for the
non-schizophrenic murderer group, where motivation for faking bad is understandable
and even likely given the possibility for significant secondary gain (i.e., avoiding
execution) in their particular situation. On balance, IQ-educational attainment
correlations were weak but positive for these groups, and non-significance may
represent not malingering but merely insufficiently powered correlational analyses. As
77
formal measures of malingering were not administered in the present study, this
determination cannot definitively be made here, and neuropsychological data from these
groups should be interpreted with some degree of caution.
The present findings are consistent with extant schizophrenia and antisocial
behavior neuropsychological literatures which report general intellectual deficits in these
populations (see above). Though all schizophrenia and homicide groups demonstrated
similar IQ deficits, it is important to consider that the etiological pathways to these
deficits may differ among groups. However, findings here may suggest that while both
schizophrenia and violent behavior separately are characterized by significant IQ
reductions, it is not general intellectual dysfunction per se—or even the additive effects
of both conditions—that contributes to violent behavior specifically within
schizophrenia (or, similarly, within mental illness in general). Alternatively, widespread
reductions in cognitive function (indexed by IQ measures) may produce deleterious—
yet differential—effects upon individuals with and without schizophrenia, or may
interact with other diagnostically-specific factors, to contribute to violent or nonviolent
outcomes among these groups.
In summary, schizophrenic murderers did not demonstrate the hypothesized
deficits in IQ functioning relative to all other groups. Rather, all schizophrenia and
homicide groups were characterized by similar general intellectual dysfunction in
comparison to normal controls. Ultimately, IQ performance did not differentiate
schizophrenic murderers from schizophrenic non-murderers; and no evidence was
78
revealed in this area which would indicate why only some individuals with
schizophrenia become violent.
Hypothesis 2: Executive Functioning
The hypothesis that schizophrenic murderers would be characterized by
significantly reduced performance on executive functioning measures (i.e., the CTT and
Gambling Task) in comparison to schizophrenic non-murderers, non-schizophrenic
murderers, and normal controls was not supported by the present findings. In fact,
though groups overall did not largely differ on executive functioning measures, results
suggested schizophrenic murderers demonstrated better performance than some other
groups on very specific aspects of these measures. On the CTT, groups did not differ on
commonly-reported performance variables (i.e., time to completion, interference index),
though schizophrenic murderers—in comparison to non-schizophrenic murderers—
committed significantly fewer number errors on the CTT-2. In fact, the mean CTT-2
number errors T scores of non-schizophrenic murderers fell in the below average range
of Heaton and colleagues’ clinical interpretation guidelines (D’Elia et al., 1995), while
those for schizophrenic murderers were in the average range. Additionally,
schizophrenic murderers’ CTT-2 performance was characterized by fewer near misses
than their schizophrenic non-murderer counterparts, though group differences only
approached significance.
Various explanations may account for these findings. First, they may in fact be
spurious—as the SD for schizophrenic murderers’ number errors T scores was very
small relative to other groups. Second, increased executive functioning in schizophrenic
79
murderers compared to non-schizophrenic murderers may simply reflect the possible
malingered neuropsychological performance discussed earlier. Third, the
aforementioned elevated head injury history in murderers may explain these findings, as
they were eliminated when head-injury-free participants were analyzed separately. This
indicates that either (1) the presence of head injury reduced CTT2 number error
performance in murderers and was responsible for the original group differences, or (2)
group sizes in the head-injury-free analyses were reduced and analyses were
subsequently underpowered. The latter is supported by the fact that participants with and
without head injury history in the overall sample did not differ significantly in
performance on any CTT variables. Either way, though evidence indicates that the
influence of head injury history upon CTT2 number errors within the murderer group
may be relatively minor, it cannot ultimately be discounted completely. Finally, results
may indicate actual group differences on specific CTT variable performance between
schizophrenic murderers and their non-schizophrenic and non-murderer counterparts.
On the Gambling Task, a marginally-significant interaction was detected
indicating group differences only during the middle portion of the test (i.e., Block 3),
wherein murderers appeared to make more disadvantageous selections than
schizophrenic murderers. Figure 1 indicates that across all five blocks, schizophrenic
non-murderers’ performance appeared very similar to that of normal controls (i.e.,
relatively stable). In fact, these groups’ selections from disadvantageous decks across
blocks were similar in number and pattern to those of psychopaths reported in other
studies, and normal controls did not demonstrate the characteristic decline in high-risk
80
selections seen in some previous studies (e.g., Mitchell et al., 2002); though they were
largely consistent with that of normal controls reported in other studies (e.g., Schmitt et
al., 1999). Other groups’ performances were characterized by distinct patterns of marked
fluctuations in disadvantageous deck selections across blocks. Non-schizophrenic
murderers were characterized by disadvantageous deck selections that decreased relative
to Block 1 selections across all five blocks. Schizophrenic murderers were characterized
by disadvantageous selections that decreased relative to Block 1 during Block 3 and
increased over Blocks 4 and 5—a pattern somewhat similar to cross-block performance
seen in low-anxious individuals in other studies (Schmitt et al., 1999). Homicide
psychiatric controls were characterized by disadvantageous selections that increased
relative to Block 1 over Blocks 2-4 and decreased back to Block 1 levels during Block 5.
Though group differences were not significant for individual block analyses, results
suggest schizophrenic murderers may be characterized by patterns of disadvantageous
deck selection that vary over time in comparison to nonschizophrenic murderers, and
that indicate less-risky decision-making than their non-schizophrenic murderer
counterparts during specific portions of the test. If and how this translates into an overall
performance deficit, however, is somewhat unclear. Nevertheless, Gambling Task
performance across all schizophrenia and homicide groups should ultimately be
considered normative for this sample, as none differed significantly in performance from
normal controls.
Mixed executive functioning results in the present study are consistent in many
ways with previous findings. First, they are partially in line with Barber’s (2004)
81
findings of increased performance on two executive functioning tests, along with
other neuropsychological measures, in schizophrenic murderers relative to schizophrenic
non-murderers. Second, the failure to find more robust group differences here on the
Gambling Task reflects earlier findings within the schizophrenia neuropsychological
literature. Heinrichs and Zakzanis (1998), in a meta-analysis of 204 neurocognitive
studies of schizophrenia, found neurocognitive deficits in 60-80% of schizophrenia
patients (leaving an additional 20-40% who do not demonstrate such deficits); and
Palmer and colleagues (1997) found that approximately one-quarter of schizophrenia
patients continue to function within normal neurocognitive limits. In fact, it appears that
no single pattern of deficits is unique or common to all schizophrenic individuals (Salva,
Moore, & Palmer, 2008). Third, the present results are partially in line with Schug and
Raine’s (2009) recent series of meta-analyses that did not find broadly-defined EF
deficits in antisocial schizophrenic persons relative to schizophrenic nonantisocials, but
did find deficits in broadly-defined EF and DLPFC functioning (which both included
Trails B) and specific EF (i.e., WCST) in comparison to non-schizophrenic antisocials.
However, in the latter schizophrenic/non-schizophrenic antisocial comparisons, ESs for
narrowly-defined EF (which also included Trails B) and OFC functioning (which
included the Gambling Task) were not significant. Mixed results—seen both here and in
previous studies—may reflect heterogeneity of cognitive functioning in schizophrenia,
heterogeneity of definitions of executive functioning, or both.
Additional considerations may also be important. First, the specific nature of
executive functioning deficits possibly indicated here by CTT performance might be
82
unclear, as the CTT error, prompt, and near-misses scores are to date considered
experimental—though thought to represent qualitative features of cognitive impairment
(D’Elia et al., 1995). CTT number errors may reflect set-changing ability, and may
speak to dorsolateral prefrontal functioning. However, these types of errors may also
simply reflect an inability to recognize Arabic numerals—understandable difficulties for
an individual whose primary language is Chinese. In fact, the impact of cultural factors
upon these neuropsychological tests of executive functioning is an important second
consideration. It may be that the CTT as a culture-free analogue to the Trailmaking Test
is simply not powerful enough to detect significant group differences in this population.
Furthermore, the cross-cultural validity of the Gambling Task has yet to be empirically
validated.
However, if findings in fact represent true group differences, better executive
performance in schizophrenic relative to non-schizophrenic murderers may reflect
compensatory frontal or prefrontal mechanisms that allow the schizophrenic murderer to
effectively commit homicide despite the characteristic cognitive disorganization of
schizophrenia. That is to say, within murderers in general, the brains of those with
schizophrenia may be burdened with an additional cognitive workload when planning,
orchestrating, and perpetrating a homicidal act (as would the brains of schizophrenic
murderers relative to their schizophrenic non-murderer counterparts—suggested by the
findings which approached significance listed above), and might require an ability to
enlist additional executive resources to successfully meet these increased cognitive
demands.
83
In summary, contrary to predictions, a small amount of evidence here may
suggest better performance in very specific areas of executive functioning (i.e.,
particular aspects of cognitive set-shifting ability and avoidance of high-risk decision-
making) in schizophrenic murderers relative to schizophrenic non-murderers and non-
schizophrenic murderers, though results may ultimately be inconclusive due to the
possible influences of small sample sizes, head injury, and other factors. Such evidence
may suggest that executive functioning differentiates schizophrenic murderers from
schizophrenic non-murderers not in the form of a relative performance deficit but rather
as a comparative performance strength—and may not explain so much why only some
schizophrenic individuals become violent, but rather how they are able to become
violent in spite of the debilitating cognitive effects of their illness.
Hypothesis 3: EEG
The hypothesis that schizophrenic murderers would be characterized by
significant EEG deficits in comparison to schizophrenic non-murderers, non-
schizophrenic murderers, and normal controls received only partial support. In the theta
band, it was schizophrenic non-murderers that were characterized by significant EEG
deficits relative to other groups ( i.e., significantly increased theta amplitudes compared
to all other groups for left and right hemispheric sites and dorsomedial prefrontal and
parietal lobe sites; compared to normal controls, murderers, and schizophrenic
murderers for medial prefrontal and dorsolateral prefrontal lobe sites; compared to
normal controls and murderers for occipital lobe sites; and compared to normal controls
at temporal lobe sites). These results are largely consistent with previous findings, and
84
likely reflect the EEG slowing and cortical hypoactivity commonly reported in
schizophrenia patients (see above). Schizophrenic murderers were also characterized by
increased theta amplitudes compared to normal controls (in fact, all groups were slightly
elevated in theta activity across most electrode sites relative to normal controls—see
Table 6), though these differences were not significant. This may reflect differences
among violent and nonviolent schizophrenic participants in medication prevalence (see
below) or medication efficacy; or may indicate a violent form of schizophrenia
characterized by less EEG slowing than that seen in nonviolent schizophrenic persons.
In the beta3 band, however, results were more supportive of Hypotheses 3,
though EEG deficits revealed in these analyses require some interpretation. Significant
differences in left hemispheric beta3 amplitudes specifically between violent and
nonviolent schizophrenic groups may be explained by reduced beta3 activity in
schizophrenic non-murderers, increased beta3 activity in schizophrenic murderers, or
both. It is here that normative context becomes important. For example, even though left
hemispheric mean beta3 amplitudes for schizophrenic murderers were increased nearly
twofold in comparison to healthy controls (M = 5.28, SD = 6.83, and M = 2.69, SD =
3.17, respectively), and mean differences between schizophrenic murderers and healthy
controls in this region were over twice those of schizophrenic non-murderers and
healthy controls (i.e., 2.59 versus -1.28), neither of these group differences were
significant—indicating that both groups demonstrated what could be considered
normative beta3 activity in this region. On balance, though neither group was
characterized by statistically abnormal left hemispheric beta3 activity per se, it is how
85
the two groups differ relative to each other in this case—rather than how either differ
from the norm—that may provide evidence as to why only some schizophrenic
individuals become violent. Given the positioning of both schizophrenic groups’ mean
beta3 amplitudes at opposite ends of a “normative” spectrum, and the positioning of
each relative to the mean beta3 amplitudes of healthy controls, the characterization of
these group differences as increased beta3 activity in schizophrenic murderers appears
plausible.
Findings of increased beta3 activity in schizophrenic murderers relative to their
schizophrenic non-murderer counterparts are also in accordance with previous findings.
For example, Saletu, Kufferle, and Anderer (1986) found increased baseline fast beta
activity—particularly in the left hemisphere—and less alpha activity in 20
exacerbated/acute schizophrenics in comparison to 20 nonpatient controls, and
interpreted these findings as suggesting left hemispheric dysfunction in the form of
overarousal. Other studies have also reported similar asymmetrical left hemispheric fast
beta activity increases in schizophrenic individuals (Shagass, 1991)—though violent
behavior was not addressed in their samples. The beta3 group differences revealed in the
present study may be a key extension of previous findings, and may suggest a
relationship between left hemispheric hyperarousal and the mechanisms underlying
violence or aggression in schizophrenia.
Additionally, beta3 findings may be explained within the context of anomalous
gamma band activity. In the present study, the beta3 power band frequency range (30 –
50 Hz) overlapped with gamma frequencies (i.e., 40 Hz; Gallinat, Winterer, Herrmann,
86
& Senkowski, 2004; Gordon, Williams, Haig, Wright, & Meares, 2001; Hong et al.,
2004) and bandwidths (e.g., 28 - 48 Hz; Basar-Eroglu et al., 2007) traditionally reported
in the schizophrenia literature. Gamma activity has been associated with numerous
cognitive functions, such as memory, attention, object representation, and feature
binding; and—due to its intraindividual stability—may represent a personal trait variable
(Lenz et al., 2008). Schizophrenia patients have demonstrated altered gamma activity on
tasks involving these cognitive abilities (Basar-Eroglu et al., 2007; Lenz et al., 2008).
Patients suffering predominantly from negative symptoms have displayed reduced
gamma activity, whereas positive symptomatology has been associated with augmented
gamma activity (Lenz et al., 2008); and it has been demonstrated that the enhancement
or reduction of gamma in schizophrenic patients results from an increase or decrease in
EEG phase-locking (thought to represent high-frequency neural synchronization
required in specific cognitive processes), respectively (Spencer et al., 2003; Spencer et
al., 2004). Thus, increased left hemispheric gamma activity here may represent
hyperarousal deficits with deleterious effects in specific domains of cognitive
functioning, which may characterize violent but not nonviolent schizophrenic persons.
Left hemispheric hyperarousal may be interpreted within the context of EEG
arousal asymmetries associated with emotion regulation, though the direction of arousal
asymmetry in schizophrenic murderers relative to schizophrenic non-murderers here
(i.e., left > right) is opposite to those reported in some antisocial populations (i.e., right >
left; Ishikawa & Raine, 2002), and likely does not represent emotional regulation
deficits. Alternatively, left hemispheric hyperarousal in schizophrenic murderers relative
87
to schizophrenic non-murderers may reflect over-activity in left hemispheric brain
centers responsible for logical or sequential reasoning, rational or analytical thought, or
the ability to be objective or look at parts rather than wholes which characterize left
hemispheric function (e.g., analytic, bit-by-bit processing; Lezak et al., 2004). Such
over-activity might explain schizophrenic paranoid ideation (which has itself been
associated with violence; Bjørkly, 2002) as an over-processing of particular
environmental cues (i.e., those interpreted as threatening) at the expense of
understanding important overarching contextual factors; or perhaps an over-
interpretation, analysis, or rationalization of available information into perceptions
which are congruent with a delusional framework. Reasoning (i.e., thinking with a
conscious intent to reach a conclusion) about content-independent situations appears to
be mediated by left hemispheric regions, whereas reasoning influenced by information
based on previous beliefs, values, or goals is mediated by right hemispheric and bilateral
ventromedial regions (Lezak et al., 2004). This latter right-hemispheric reasoning could,
intuitively, be based more upon memory function, which appears reduced in antisocial
schizophrenic persons (Schug & Raine, 2009). Though memory function was not
assessed here, and no differences were seen in the Gambling Task (which is thought to
index ventromedial functioning), overactive left hemispheric reasoning processes in the
absence of intact right hemispheric reasoning processes (undermined by deficits in
context-relevant or even “emotional” types of memory) could explain why a
schizophrenic individual would focus upon context-independent cues and perceive them
as threatening while failing to call upon important right hemispheric belief, value, and
88
goal-oriented processing resources; and thus be more likely to act out violently
relative to a schizophrenic individual with normative or even deficient left-hemispheric
reasoning abilities.
Furthermore, this proposal would be in line with region-specific gamma phase-
locking anomalies found to be associated with schizophrenia symptomatology. For
example, Spencer et al. (2004) found visual Gestalt stimulus-evoked gamma band
phase-locking in the occipital lobe to be significantly and positively correlated with
conceptual disorganization, visual hallucinations, thought withdrawal delusions, and
global thought disorder in schizophrenia patients. Though phase-locking was not
assessed in the present study, increased resting gamma amplitudes could be related to
altered gamma phase-locking. Such activity occurring in the left hemisphere might
facilitate excessive context-independent feature-binding, which could predispose a
schizophrenic individual to violence via an integration of context-independent
environmental cues which are subsequently perceived as confusing or even threatening.
Limitations to this proposal exist. For example, IQ data from the present study
do not immediately suggest specific left hemispheric dysfunction—in the form of over-
activation or otherwise—which differentiates among schizophrenic groups (i.e., Verbal
IQ deficits or Verbal-Performance IQ discrepancy differences). On balance, the
proposed left-hemispheric dysfunction evidenced by EEG functioning here may not be
directly indexed by traditional intelligence measures—particularly the limited number of
prorated WAIS-RC subscales from which IQ scores here were generated. Ultimately, a
left-hemispheric over-processing hypothesis specific to violent schizophrenia would
89
require further testing, and limited EEG evidence here can only provide an impetus
for such a line of enquiry. Nonetheless, the present EEG findings do suggest
differentiation based upon EEG functioning between violent and nonviolent
schizophrenic persons, and may—if nothing else—provide further empirical support for
EEG subtypes of schizophrenia which have been discussed in the literature (Shagass,
1991).
Antipsychotics, head injury, and EEG findings. The potential influences of
antipsychotics use and head injury history upon EEG findings in the present study
require some discussion. Regarding head injury, separate analyses indicated that a
history of such injury did not affect resting EEG amplitudes over the entire sample in the
regions and power bands where group differences were revealed. Head injury data may
have been limited in validity as it was self-reported, and injury to different areas of the
brain could produce differential EEG abnormalities. However, while the influences of
head injury must always be considered and cannot be entirely ruled out here, it is likely
that these influences did not largely contribute to resting EEG amplitude group
differences in the present study.
The relationship between antipsychotics use and the present EEG findings is
somewhat more complex. Several different types and classes of antipsychotic
medications were used in this sample, though these medications overall appeared to be
applied in typical dose levels which were relatively similar in dosage equivalencies
across medications. As such, the within-subjects variance in medicated participants of
medication effects due to medication type and dosage was likely minimized, though it
90
cannot be ruled out entirely—particularly due to the heterogeneity of response to
antipsychotics reported in the literature (Garver, Holcomb, & Christensen, 2000), the
limitations of reported dosage equivalencies, and the lack of established correlations
between dosage and clinical response (Borison, Pathiraja, Haverstock, Gowda, &
Diamond, 1995).
Increased theta amplitudes in schizophrenic participants relative to all other
groups and reduced beta3 activity in schizophrenic non-murderers relative to
schizophrenic murderers may simply reflect the effects of antipsychotic medication. For
example, various traditional and second-generation antipsychotics are known to increase
slow-wave activity in schizophrenic patients (Freudenreich, Weiner, & McEvoy, 1997;
Shagass, 1991; Wichniak, Szafrański, Wierzbicka, Waliniowaska, & Jernajczyk, 2006),
and limited data suggests that pre-treatment and treatment-associated EEG abnormalities
may be associated with clinical improvement in psychotic patients using particular
antipsychotic agents (Centorrino et al., 2002; Pillay et al., 1996; Stevens, 1995). In fact,
Stevens (1995) has argued clozapine-induced EEG slowing is evidence for its
therapeautic activity and the basis for its clinical efficacy. These effects are paralleled in
electroconvulsive therapy (ECT) studies, in which ECT-induced electrophysiologic
slowing reflects its therapeautic effects and is associated with a more-positive clinical
outcome (Fink, 2002; Fink & Kahn, 1957).
However, effects of antipsychotic medication upon sample-wide EEG theta
amplitudes are complex and somewhat unclear. While antipsychotics appeared to
significantly increase theta amplitudes in some brain regions (i.e., parietal and occipital
91
lobe sites), their effects were of marginal significance in other areas (i.e., overall left
and right hemispheric sites, and temporal lobe sites) and had no significant effect in
prefrontal region lobe sites. Additionally, left hemispheric gamma activity was not
significantly affected by antipsychotic medication use sample-wide. Thus, it appears that
the EEG slowing observed in the present study may be due in part to the effects of
antipsychotic medication, though these effects may be limited to specific brain regions.
In the prefrontal area, observed slowing may be more attributable to diagnostic group
differences (though medication effects cannot be entirely ruled out). Group differences
in theta amplitudes between schizophrenic murderers and schizophrenic non-murderers
in prefrontal regions could have significant meaning—suggesting that the former are not
characterized by the hypofrontality observed in the latter. This may indicate a functional
strength in schizophrenic murderers relative to their schizophrenic non-murderer
counterparts, and may speak to increased executive abilities (goal formulation, planning,
purposive action, and self-monitoring/correction) which would facilitate homicidal
behavior in these individuals—particularly in conjunction with the context-independent
over-processing proposed above. It would also be in line with the marginally increased
performance seen here in one aspect of executive functioning (CTT near misses) by
schizophrenic murderers relative to schizophrenic non-murderers. However, given the
potential complexity of antipsychotic medication effects upon EEG here, along with the
limited amount and questionable nature of corroborating data here, this explanation must
be offered with caution.
92
Antipsychotic medication use in the present study may also have its own
intrinsic meaning. For example, it may reflect a greater pre-medication degree of illness
severity or reduced level of functioning—wherein those schizophrenic individuals with
the most severe symptomatology were administered antipsychotic medication. To that
end, schizophrenic murderers could represent a group of higher-functioning
schizophrenic individuals in the sense that they were proportionally less medicated with
antipsychotics. The use of clozapine illustrates this point. In general, only a small
proportion of patients (5% or less) are treated with clozapine in clinical settings, and its
use is reserved for only the most severe forms of schizophrenia (Sajatovic,
Madhusoodanan, & Fuller, 2008). The fact that 25 percent of the schizophrenic non-
murderers in the present study were characterized by clozapine use (as opposed to only
one schizophrenic murderer) may indicate that they were as a group suffering from more
severe forms of schizophrenia. On balance, antipsychotic medication status may be a
derivative of the schizophrenic individual’s antisociality. Schizophrenic murderers may
overall be comparatively more antisocial and perhaps characterized by poorer
medication compliance than schizophrenic non-murderers. Or, schizophrenic murderers
may be predominantly antipsychotic-nonmedicated due to limited access to these
medications within the criminal justice system. In all, the potential meaning of
antipsychotic medication use here may have important implications beyond its
immediate effects upon EEG functioning.
Antipsychotic medication data possessed some significant limitations in the
present study. First, information regarding medication use status was not available for all
93
participants; and medication class, name, and/or dosages could not be ascertained in
some cases. Additionally, data related to illness stage and duration of antipsychotic
administration were not available here—an important consideration given that EEGs in
schizophrenic individuals may change secondary to antipsychotic administration or
illness progression (Merlo, Kleinlogel, & Koukkou, 1998; Nagase, Okubo, & Toru,
1996). To this end, schizophrenic murderers may have been in earlier stages of illness—
perhaps characterized by sudden psychotic episode onset or prodromal phases which
passed unnoticed by intimates—and EEG slowing associated with more long-term
illness or antipsychotic medication use may not yet have developed in this violent
schizophrenic group. Finally, the confounding of medication status with group
membership (in the case of schizophrenic non-murderers) precluded any attempts to
examine differential effects of antipsychotic medication among groups. Though the
present findings may suggest schizophrenia subtype-specific EEG profiles representing
differential clinical responses to antipsychotic medication among schizophrenia groups
(see Merlo et al., 1998), this assumption cannot definitively be made as while all
schizophrenic non-murderers were characterized by antipsychotic medication use, only
30 percent of schizophrenic murderers were medicated with antipsychotics. Future
methodologies involving equitable proportions of both medicated and non-medicated
schizophrenic murderers and non-murderers would allow for comparisons that might
help address this issue.
Ultimately, while the complete extent of antipsychotic medication influences
cannot effectively be known here, it appears that the group differences in resting EEG
94
amplitudes were not due entirely to medication effects for many brain regions in the
present study—though the effects appeared more prominent in some areas. Future
studies incorporating both medicated and unmedicated groups (analyzed separately)
with more comprehensive medication data could help to disentangle these medication
influences, and further the understanding of antipsychotic medication effects upon EEG
functioning in schizophrenic individuals who do and do not become violent.
In summary, resting EEG amplitudes may suggest functional cortical deficits
which characterize schizophrenic non-murderers and schizophrenic murderers
separately. Relative to each other, schizophrenic persons who did not murder
demonstrated diffuse EEG slowing (as they were in comparison to all other groups),
whereas those who did were characterized by increased left-hemispheric fast-wave
activity—suggesting hyperarousal deficits in this region which may predispose to
violent behavior.
Hypothesis 4: Psychosocial Functioning and Childhood Psychosocial Deprivation
The hypothesis that schizophrenic murderers would be characterized by
significantly more deficits in psychosocial functioning and more childhood psychosocial
deprivation in comparison to schizophrenic non-murderers, non-schizophrenic
murderers, and normal controls received only partial support. While groups did not
differ on some indices of psychosocial functioning (i.e., couple relationship history,
residential difficulties), other mixed findings were revealed. First, while schizophrenic
non-murderers and non-schizophrenic murderers were characterized by longer periods of
recent unemployment compared to normal controls—schizophrenic murderers were not
95
(in fact, the mean length of recent unemployment for schizophrenic murderers was
approximately centered between those of normal controls and schizophrenic non-
murderers, whereas that of non-schizophrenic murderers was more proximal to
schizophrenic non-murderers—see Table 15). However, further examination revealed
that percentages of individuals currently employed in skilled positions varied across
groups (i.e., 66.7% of normal controls, 70.0% of schizophrenic non-murderers, 52.4% of
non-schizophrenic murderers, 50.0% of schizophrenic murderers, and 35.7% of
homicide psychiatric controls); and while schizophrenic murderers were not
characterized by abnormally extended periods of recent unemployment, they were more
likely to be employed in unskilled work. This might suggest a group of schizophrenic
murderers characterized by reduced occupational functioning, wherein unskilled work
represents employment that is less demanding upon cognitive and interpersonal
resources and better-suited for individuals with deficits in these areas. On balance,
however, it may speak to a group of schizophrenic murderers who are better equipped
psychosocially on some level—at least with the abilities necessary to maintain
consistent though perhaps more menial employment—than their schizophrenic non-
murderer and non-schizophrenic murderer counterparts. As such, abnormally extended
periods of recent unemployment may constitute a key psychosocial difference between
schizophrenic individuals who do and do not become violent, as well as between violent
individuals with and without schizophrenia.
Second, psychosocial functioning group differences were observed in
proportions of recent homelessness. Schizophrenic murderers were characterized by
96
more recent homelessness in comparison to both normal controls and schizophrenic
non-murderers (though group differences with the latter only approached significance).
These findings may reflect the additive effects of both schizophrenia and homicide
conditions, as both demonstrated increased (though not significantly increased) rates of
homelessness relative to normal controls (see Table 15), or may suggest a risk factor or
psychosocial etiological mechanism specific to a homicidal form of schizophrenia.
Homelessness is a noted problem among the severely and persistently mentally
ill—including those with schizophrenia—which can impede access to treatment,
contribute to deterioration in social functioning, and attenuate social bonds and family
support (Felix, Herman, & Susser, 2008). Yet why increased rates of homelessness here
would characterize schizophrenic murderers more so than other mentally ill groups
(schizophrenic non-murderers or homicide psychiatric controls—though rates for the
latter were similar to those of schizophrenic murderers) is an interesting question. One
explanation may be that homelessness here represents a psychosocial stressor which
leads to the onset of schizophrenia in vulnerable individuals (Bebbington & Kuipers,
2008)—particularly in those with pre-existing proneness to schizophrenia, violence, or
both. Unfortunately, no data related to the temporal relationship of both schizophrenia
and homelessness onset was available in the present study to support this proposal.
Another explanation may be that homelessness here is a by-product of a more
psychosocially debilitating form of schizophrenia in schizophrenic murderers relative to
non-murderers, or a form of schizophrenia characterized by or interacting with more
potent risk factors for psychosocial impairment. For example, childhood adversity
97
(family separations, abuse, and neglect) has been associated with homelessness in the
severely mentally ill (Felix et al., 2008)—findings which are consistent with the
aforementioned results indicating abnormally elevated levels of childhood psychosocial
deprivation in schizophrenic murderers here. Yet another explanation may be that
homelessness is an exacerbating psychosocial factor which specifically pushes
schizophrenic individuals “over the edge” into violence. For example, it may suggest a
financial motive for homicides in schizophrenic murderers. However, this does not
appear to be supported by data in other study areas—for though schizophrenic murderers
were characterized by lower SES relative to non-schizophrenic murderers (see Table 1),
significant group differences in SES were not observed; and none of the schizophrenic
murderers reported financial motivations for their homicides (see Table 16). Finally,
lack of current family support may also be associated with homelessness among severely
mentally ill persons (Felix et al., 2008), and homelessness here would be consistent with
a failure or lack of social support networks which may distinguish schizophrenic
murderers from schizophrenic non-murderers (see below). Though all of these
explanations are plausible and suggest more indirect influences of homelessness upon
schizophrenic violence, none directly explains why homelessness—in and of itself—
would necessarily cause violence in schizophrenic persons. Still, they suggest a
psychosocial factor—abnormally elevated recent homelessness—which may be unique
to a homicidal subgroup of schizophrenic persons.
Third, groups differed on psychosocial functioning indices of social support
network viability. Both normal controls and schizophrenic non-murderers were
98
characterized by more close friends than murderers, whereas schizophrenic murderers
were not. Normal controls were also characterized by more close friends than homicide
psychiatric controls; and by more family emotional closeness than all murderer groups
(murderers, schizophrenic murderers, homicide psychiatric controls). Thus, while
schizophrenic murderers did not demonstrate indicators of social support network
impairment in some areas relative to other groups (i.e., nonfamilial interpersonal deficits
relative to normal controls and schizophrenic non-murderer counterparts like their non-
schizophrenic counterparts), they, like other types of murderers, demonstrated what
could be impaired support networks in the form of reduced emotional closeness with
family members. In schizophrenic murderers, such impairment could represent familial
support network failure secondary to interpersonal difficulties occurring at illness onset,
or may suggest a pre-schizophrenic incapacity for emotional closeness. For example,
familial emotional closeness may in this case indirectly index emotional attachment,
which has been found to be impaired in some antisocial populations (Bowlby, 1946).
The fact that all homicide groups were characterized by a similar lack of familial
emotional closeness supports this latter explanation (that reduced emotional closeness
indexing emotional attachment deficits predisposes to violence in general rather than in
schizophrenia in particular); though the former proposal cannot be ruled out entirely as
etiological pathways to violence may differ among homicide groups. Thus, while
reduced social support network viability appeared to differentiate violent from
nonviolent schizophrenic persons in some respects, these differences were more
99
characteristic of homicidal violence as a whole, and not necessarily specific to a
violent schizophrenic subgroup.
Fourth, groups differed in psychosocial functioning in terms of current marital
status and marital history. Both normal controls and schizophrenic murderers were more
likely to be married than schizophrenic non-murderers and non-schizophrenic
murderers—first evidence in the present study for a subgroup of violent schizophrenic
persons simultaneously distinct on the same factor from both their schizophrenic non-
murderer and non-schizophrenic murderer counterparts; and schizophrenic non-
murderers were more likely to have never been married than both schizophrenic
murderers and normal controls. This may suggest a group of schizophrenic murderers
who are perhaps better equipped interpersonally to navigate intimate marital
relationships than their non-murderer schizophrenic counterparts. However, whether
marital status—in the absence of other contextual data—should ultimately be considered
a protective factor (i.e., increased social support) or a risk factor (i.e., increased
opportunity for intimate partner victimization) for schizophrenic violence may be
difficult to ascertain here. Either way, marital status appears to be a psychosocial factor
which simultaneously differentiates schizophrenic murderers from their schizophrenic
non-murderer and non-schizophrenic murderer counterparts, and may be first study
evidence for a psychosocially distinct violent schizophrenic subgroup.
Finally, groups differed in psychosocial functioning in terms of current
independent/solitary living arrangements. Normal controls, non-schizophrenic
murderers, and schizophrenic murderers were all more likely to be living alone than
100
schizophrenic non-murderers. Additionally, normal controls were more likely to be
living alone than homicide psychiatric controls. Living alone may reflect higher
psychosocial functioning and ability to live independently (i.e., schizophrenic murderers
more likely to live alone than schizophrenic non-murderers may indicate that
schizophrenic murderers require less assistance with daily living needs). On balance, it
also may reflect a psychosocial deficit—indicating an individual whose psychosocial
support network has failed, or an individual who is psychosocially ill-equipped to live
with others. Or, it may represent a risk factor for more severe forms of illness, as social
isolation can facilitate the exacerbation of prodromal or developing psychotic
symptomatology (e.g., delusional beliefs; Bebbington & Kuipers, 2008). However, as
the pattern of living alone in schizophrenic murderers highly resembled that of normal
controls, it can ultimately be assumed that this way of life is somewhat normative in this
particular area of China. As such, it could represent an area of normative psychosocial
functioning, superior to that of schizophrenic non-murderers and similar to (though
somewhat better than) that of non-schizophrenic murderers. Regardless,
independent/solitary living appears to be another psychosocial factor which
distinguishes schizophrenic murderers at least from their non-schizophrenic
counterparts, and may offer insight as to why only some schizophrenic individuals
become violent.
Interestingly, high percentages of both being currently married and living alone
were separately reported among normal controls and schizophrenic murderers, and this
at first glance may appear somewhat contradictory. However, subsequent Chi Square
101
analyses among currently married individuals revealed that many participants across
groups reported both being currently married and living alone (73.3% or n = 22 of
normal controls, 55.6% or n = 5 of schizophrenic non-murderers, 50.0% or n = 4 of non-
schizophrenic murderers, 77.8% or n = 14 of schizophrenic murderers, and 33.3% or n =
2 of homicide psychiatric controls), and that groups overall did not differ significantly in
proportions of currently married individuals living alone when analyzed together ( χ
2
=
6.05, df = 4, p = .195). Additionally, groups largely did not differ when analyzed
individually—only schizophrenic murderers were characterized by significantly higher
proportions of individuals currently married and living alone in comparison to homicide
psychiatric controls ( χ
2
= 4.00, df = 1, p = .046); though normal controls were
characterized by higher proportions which approached significance relative to this group
( χ
2
= 3.60, df = 1, p = .058). However, group sizes were comparatively small for
homicide psychiatric controls (i.e., n = 4 currently married and living together, n = 2
currently married and living alone), and these results should be interpreted with caution.
In summary, living alone while married was not an anomalous characteristic in the
present sample, and groups on the whole did not differ in this trend.
The question remains as to if living alone while married constitutes a
psychosocially normative or impaired characteristic. On one hand, it may represent a
culturally-specific normative tendency for married individuals to reside apart for
employment reasons (i.e., employment in separate cities). On balance, however, it might
also reflect separation due to marital problems, or be due to group-specific situations,
such as being currently married but incarcerated (i.e., in murderer groups)—essentially
102
representing psychosocial impairment (it is also consistent with the elevated rates of
homelessness in schizophrenic murderers discussed above). Ultimately the specific
explanation for this phenomenon cannot be ascertained here, as data related to reasons
for living alone while married were not collected. However, as this type of living
situation was highly prevalent among normal controls, it will be considered here a
normative practice among married individuals in this sample.
Childhood psychosocial deprivation. In contrast, findings related to childhood
psychosocial deprivation were much more supportive of Hypothesis 4. First, though all
groups demonstrated elevated levels (in varying degrees) of total childhood psychosocial
deprivation factors relative to normal controls, schizophrenic murderers were the only
group in which these levels were significantly increased. Similar to findings here related
to homelessness, this pattern of group differences may reflect the additive effects of both
schizophrenia and homicide, or may suggest an etiological factor specific to
schizophrenic individuals who commit murder.
These findings are largely consistent with previous work in this area. For
example, studies have shown increased rates of childhood maltreatment history among
severely mentally ill adults—rates which far exceed those seen in the general
population; and others have reported an association between increased symptomatology
and histories of abuse among psychiatric patients (Schenkel, Spaulding, DiLillo, &
Silverstein, 2005). Failure to replicate findings of significantly increased childhood
psychosocial deprivation across all schizophrenia and homicide groups may reflect a
broader scope of childhood psychosocial deprivation used here relative to previous
103
studies (e.g., childhood maltreatment in Schenkel et al., 2005). On balance, it may
reflect a risk factor which finds expression in only the most severe manifestations of
both violence and mental illnesses such as schizophrenia (i.e., violent individuals with
schizophrenia, or schizophrenic persons characterized by violence). However, most
studies of childhood maltreatment in mentally ill individuals in general and
schizophrenia in particular (including Schenkel et al., 2005) have not addressed violence
per se (except for suicidality—see Schenkel et al., 2005) or attempted to differentiate
violent from nonviolent persons using childhood maltreatment measures. As a
neurodevelopmental disorder, it would make sense that the underpinnings of
schizophrenia lie in the earlier years of life—where the additive effects of environmental
or psychosocial adversity could propel schizophrenia development onto a more severe
trajectory. Ultimately, findings here suggest that abnormally elevated levels of
childhood psychosocial deprivation is a psychosocial etiological factor which may be—
in and of itself or perhaps in increased quantities due to additive effects of two
conditions—specific to a violent subgroup of schizophrenic individuals.
Second, schizophrenic murderers were characterized by significantly higher birth
order than both schizophrenic non-murderers and non-schizophrenic murderers—further
evidence for a distinct subgroup of violent schizophrenic persons—and suggesting
differences which may be biological or psychosocial in nature. From a psychosocial
perspective, higher birth order may reflect large family size, which has been associated
with a greater risk for delinquency, crime, and violence—possibly explained by
economic factors (economic deprivation, overcrowding in homes, and poor living
104
conditions), overtaxed parent care resources, the “contagion” hypothesis (i.e., the
larger the family size, the more likely it is that one of the siblings will be delinquent, and
the more likely that another sibling will “catch” delinquency), or the suggestion that
criminals are more likely to have large families (the offspring of which are more likely
to become criminal or delinquent either through genetic transmission of or familial
consequences of parental criminality; Raine, 1993). In the present study, however,
subsequent correlational analyses revealed small, nonsignificant negative correlations
between birth order and family size, and these correlations approached significance only
for the schizophrenic murderer group (r = -.396, p = .062). Additionally, Chi Square
analysis revealed that groups did not differ overall with respect to the sample median
split of family size; indicating that family size per se did not likely contribute to a
differential risk for violence among groups. On balance, subsequent traditional and
modern methods revealed that schizophrenic non-murderers were characterized by
significantly higher parental SES than both normal controls and schizophrenic murderers
(ps = .001 and < .001, respectively). This could suggest, in part, a psychosocial
explanation for findings of higher birth order (i.e., indexing economic deprivation) in
schizophrenic murderers relative to non-murderers, though low parental SES cannot
definitively be attributed to large family size here. In fact, subsequent correlational
analyses revealed only small, nonsignificant positive correlations between parental SES
and family size (i.e., larger family size associated with lower SES) for all sample
participants together—correlations which reversed direction (yet remained
nonsignificant) for all groups other than normal controls when analyzed separately. No
105
data related to sibling criminality was available in the present sample, so the
contagion hypothesis could not be tested. Furthermore, participants were not largely
characterized by parental criminality (the highest percentage—9.4 %—was observed in
schizophrenic non-murderers), and Chi Square analysis revealed that groups overall did
not differ with respect to parental criminality ( χ
2
= 5.33, df = 4, p = .255)—evidence
against genetic transmission or familial consequences of parental criminality. Thus, in
aggregate, evidence suggests possible but not definitive psychosocial explanations for
birth order findings here.
Alternatively, from a biological perspective, findings of higher birth order in
schizophrenic murderers may indicate the presence of neurodevelopmental risk factors
for schizophrenia such as maternal prenatal viral infection (reflecting increased infection
risk due to having young immuno-deficient children in the household), in utero
nutritional deficiency (Sham, Maclean, & Kendler, 1993), or maternal-fetal Rh
incompatibility (Gaughran et al., 2007; Hollister et al., 1996). Additionally, higher birth
order may reflect advanced paternal age at birth—thought to represent a risk factor for
schizophrenia (Tsuchiya et al., 2005)—and be explained by de novo mutations or
abnormal mylenation of paternally imprinted genes (Malaspina et al., 2005).
Corroborating evidence here may support the latter explanation. Though data
related to paternal age at birth were not collected in the present study, maternal birth age
data were—and these were subsequently examined as a proxy measure of paternal birth
age (given the natural and strong positive correlations between the two; Tsuchiya et al.,
2008). A one-way ANOVA was conducted to assess group differences in mother’s age
106
at birth among diagnostic groups (M = 29.35, SD = 6.82 for normal controls; M =
26.23, SD = 3.96 for schizophrenic non-murderers; M = 28.50, SD = 5.22 for non-
schizophrenic murderers; M = 31.96, SD = 6.46 for schizophrenic murderers; and M =
26.79, SD = 4.23 for homicide psychiatric controls). Results indicated significant group
differences (F(4) = 3.79, p = .006). Post hoc (Bonferroni) analyses revealed that age of
mothers at birth for schizophrenic murderers was significantly higher in comparison to
that of schizophrenic non-murderers (p = .004). A subsequent percentile bootstrap
method for 20% trimmed means (Benjamini-Hochberg method) confirmed that age of
mothers at birth for schizophrenic murderers was significantly higher in comparison to
that of schizophrenic non-murderers (p = .001). This may (via correlation) suggest
paternal age at birth which is significantly advanced in schizophrenic murderers relative
to schizophrenic non-murderers. Additionally, maternal birth ages ranged from 22 – 46
years in the schizophrenic murderer group—consistent with maternal birth ages
associated with elevated risk for schizophrenia (i.e., < 35 years in some studies, 28 – 51
years in others; Dalén, 1988). Corresponding paternal birth ages may have been even
higher, though even if identical to maternal birth ages they would overlap to an extent
with the reported range of paternal birth age associated with increased schizophrenia risk
(i.e., 29 – 50 years and above; Malaspina et al., 2005; Tsuchiya et al., 2005). Together,
these may constitute supportive evidence for a biological explanation of birth order
findings, though this suggestion must be made with caution. Perhaps in schizophrenic
murderers these paternal germ line mutations are even more severe than in schizophrenic
non-murderers, and subsequently manifest in a form of schizophrenia characterized by
107
homicidal violence. Regardless of explanation, significantly higher birth order in
schizophrenic murderers may be key evidence for a biopsychosocially distinct subgroup,
separable from both their schizophrenic non-murderer and non-schizophrenic murderer
counterparts, and may speak to important etiological mechanisms underlying this
group’s particular violent form of illness.
China’s one-child mandate. One important and unique sociocultural factor to
consider here is the possible impact of China’s birth quota or one child per couple policy
(OCP), established in 1979 (Zhang & Goza, 2006). Violations of this policy may result
in heavy financial and/or household registration penalties (Li, 1995), which could have
been directly (or indirectly) related to indices of childhood psychosocial deprivation
here. Though the term “one child per couple” is misleading—as the policy allows for
local variations in policy interpretation and for two or more children in some families
(Yang, 2007)—it would be important here to ascertain if childhood psychosocial
deprivation was attributable in any way to OCP violation. Data related to participants’
births representing OCP violations was not directly collected, however, so birth dates
(occurring in or after the year 1979) cross-referenced with birth order (i.e., second child
or greater) provided a proxy measure of those individuals whose birth may have been in
violation of the mandate—with the understanding that it did not as a measure consider
differences in regional OCP implementation or other OCP exceptions (e.g., girl-
exception or exceptions based upon child health status; Yang, 2007). An independent-
samples T-test employed on the entire study sample revealed that though participants
whose births may have been in violation of the OCP were characterized by increased
108
total number childhood psychosocial deprivation factors in comparison to those who
were not (M = 5.15, SD = 2.13, and M = 4.17, SD = 2.28, respectively), differences only
approached significance (t(146) = -1.85, p = .066). Modern methods confirmed these
differences were nonsignificant (ps = .197 and .325 using a Yuen-Welch test and a
percentile bootstrap method comparing modified M-estimators, respectively: Wilcox,
2005). Furthermore, a Chi-Square analysis revealed no significant differences among
group proportions of individuals whose births were in potential violation of the OCP ( χ
2
= 2.61, df = 4, p = .625). Thus, though OCP birth violation may have affected childhood
psychosocial deprivation in the overall sample to some extent, it likely did not affect
inter-group differences in childhood psychosocial deprivation.
In summary, several indices of psychosocial functioning suggest a psychosocial
landscape which distinguishes schizophrenic murderers from other groups, with
functional deficits in some distinct psychosocial and interpersonal arenas (lack of family
emotional closeness—a characteristic shared by all murderers—and recent
homelessness) yet relatively normative functioning in others (employment consistency,
marital status and history, and independent living). This contrasted sharply with their
schizophrenic non-murderer counterparts, who were characterized by more
unemployment, dependent living, and normative viability of social support networks
(i.e, close friendships and familial emotional closeness). Schizophrenic murderers were
also unique in being the only group characterized by abnormally elevated levels of
childhood psychosocial deprivation, and by significantly higher birth order than both
their schizophrenic non-murderer and non-schizophrenic murderer counterparts.
109
Whether or not these patterns of psychosocial functioning reflect precursors, risk
factors, or products of the illnesses cannot be definitively determined at this point;
though the latter finding may suggest a biopsychosocial etiological mechanism which is
unique to a violent schizophrenic subgroup.
Hypothesis 5: Forensic Clinical Correlates
The second aim of the present study was to assess forensic clinical data for
indicators of qualitative differences in the homicides of schizophrenic murderers, which
perhaps reflected their unique biopsychosocial profile. The hypothesis that groups would
differ significantly in various forensic clinical aspects received only partial support.
First, the homicides of schizophrenic murderers appeared to differ qualitatively in
motivational aspects relative to other murderer groups, in ways which were congruent
with their distinct pattern of biopsychosocial functioning. For example, non-
schizophrenic murderers appeared to be motivated most by anger, whereas
schizophrenic murderers were more often unable to articulate their own homicidal
motivations. This latter finding is consistent with the apparently motiveless homicides in
schizophrenic murderers reported by Varma and Jha (1966), and could reflect to some
extent the characteristic confusion and disorganization of schizophrenia. It may also,
however, reflect the biological deficits—in the form of left-hemispheric hyperarousal—
observed in schizophrenic murderers here, wherein overactive context-independent
processing mechanisms facilitated in these individuals the construction of homicidal
motivation which was meaningful in the pre-offense period but made no subjective
sense in the post-offense period.
110
Additionally, neither schizophrenic murderers nor homicide psychiatric
controls reported being motivated by love, nor were any schizophrenic murderers
motivated by financial gain. For schizophrenic murderers, the former again may be
consistent biologically with findings here of left hemispheric overarousal (i.e., a focus
upon rational rather than emotional processing) and psychosocially with findings of lack
of familial emotional closeness—possibly indexing attachment deficits. The latter,
however, may speak to the enigmatic cognitive disorganization or randomness of
schizophrenia, wherein schizophrenic murderers here were characterized by lower SES
and more recent homelessness but not financial homicidal motivation.
Second, the homicides of schizophrenic murderers were characterized by
qualitative differences in behavioral aspects compared to other types of murderers, in
ways which were possibly reflective of their unique biopsychosocial profile. For
example, schizophrenic murderers’ homicides involved less offense planning (also
consistent with Varma & Jha, 1966, though somewhat inconsistent with CTT2 number
errors results here which may reflect better planning abilities in schizophrenic
murderers) and concomitant sexual offenses in comparison to those of non-
schizophrenic murderers, and no differences between either group and homicide
psychiatric controls were revealed. The absence of sexual homicides among
schizophrenic murderers is consistent with previous research which discusses the rarity
of sexual crimes among schizophrenic persons.
Additionally, victimology appeared to differ in some respects among groups, and
for schizophrenic murderers these differences again reflected their unique
111
biopsychosocial characteristics. Victims of non-schizophrenic murderers were more
likely to be romantic partners or spouses compared to schizophrenic murderers, whereas
schizophrenic murderers were more likely to kill family members and relatives. In fact,
strangers were the least likely victims of schizophrenic murderers in the present study
(see Table 17). These results are consistent with previous research which has found
violent schizophrenic individuals to more likely assault intimates (Häkkänen &
Laajasalo, 2006; Joyal et al., 2004; Nijman et al., 2003; Varma & Jha, 1966)—and
possibly explained by the fact that is these individuals who are likely to be available
victims as they assume caretaker roles in the schizophrenic person’s life. On balance,
schizophrenic murderers were characterized by reduced family emotional closeness and
independent living (even when married), which may be evidence against victimization
of a family member simply due to their proximity as caretakers. It may be that this lack
of emotional closeness fostered paranoia and subsequent aggression toward these known
victims. In contrast, it is also interesting that family members and relatives were the least
likely victims of non-schizophrenic murderers. No differences were observed in
victimology between either group and homicide psychiatric controls, however.
Additionally, no group differences in victim gender were observed among murderer
groups, nor were there significant differences in offender-victim gender relationships.
Non-schizophrenic murderers were characterized by more victims of the opposite gender
(though this was not a significant difference), which is in line with the finding that these
murderers were more likely to kill romantic partners or spouses.
112
Third, the homicides of schizophrenic murderers appeared qualitatively
distinct in emotional aspects relative to murderers without schizophrenia, in ways which
were consistent with their unique biopsychosocial profile. All murderer groups
described a variety of emotions occurring in time periods before, during, and after
homicide offenses. The chronological order of these descriptors suggests that the
homicides were dynamic emotional experiences which differed among groups in terms
of specific emotions as well as overall affect valence shift (i.e., from positive to
negative, or vice versa). Regarding prevalence of specific descriptors, highest
percentages were seen within non-schizophrenic murderers for anger before the offense
(46.7%), followed by anger during the offense (46.7%) and fear after the offense
(34.5%; in fact, this prevalent post-offense fear in non-schizophrenic murderers was the
highest percentage of fear across all groups and time periods—perhaps reflecting they
were the most concerned with or capable of understanding the legal ramifications of
their offense). Interestingly, and in contrast to other murderer groups, no non-
schizophrenic murderers reported feeling out of control at any time during the offense.
Within schizophrenic murderers, highest percentages were seen for anger before
(37.0%), anger during (35.5%), and absence of emotion after the offense (23.3%).
Within homicide psychiatric controls, highest percentages were seen for arousal before
(28.6%), absence of emotion (35.7%), and anger after the offense (35.7%--the highest
percentage of post-offense anger observed across murderer groups).
Interestingly, each group was characterized by the highest percentage of absence
of emotion across groups at different offense time periods. Schizophrenic murderers
113
reported the highest percentages of absence of emotion before the offense (33.3%),
homicide psychiatric controls during the offense (35.7%), and non-schizophrenic
murderers after the offense (27.6%).
Differential shifts in affect valence were also observed among murderer groups.
Within non-schizophrenic murderers, overall prevalence of negative affect decreased
from 73.3 to 55.1 percent, overall prevalence of positive affect increased from 10.0 to
17.2 percent, and overall prevalence of absence of emotion increased from 16.7 to 27.6
percent across offense time periods. Within schizophrenic murderers, overall prevalence
of negative affect decreased slightly from 45.5 to 39.9 percent, overall prevalence of
positive affect increased from 11.1 to 23.3 percent (an increase of over two-fold), and
overall prevalence of absence of emotion decreased from 33.3 to 23.3 percent across
offense time periods. In fact, schizophrenic murderers as a group demonstrated the
highest total percentage of post-offense positive affect and the highest overall
proportional increase in positive affect across time periods. Within homicide psychiatric
controls, overall prevalence of negative affect increased from 49.9 to 71.3 percent,
overall prevalence of positive affect decreased from 28.6 to 7.1 percent (a decrease of
approximately four-fold), and overall prevalence of absence of emotion remained
constant.
The differential proportions and types of emotions endorsed in each offense time
period along with the differential shifts in affect valence across offense time periods
within murderer groups may reflect distinctive biological, psychological, or motivational
factors among groups. Biologically, differential shifts in affect valence across groups
114
may—in schizophrenic murderers—reflect the left hemispheric
hyperarousal/overactivity proposed here (which could affect emotional processing in
general, or the integration, interpretation and meaning of a traumatic event in particular),
or may—for all groups—represent underlying structural or functional deficits in other
brain regions (e.g., limbic responses to violent or traumatic experiences) that may
characterize each group distinctly. Psychologically, the endorsement of absence of
emotion may reflect emotional numbing in response to trauma or flattened affect
characteristic of schizophrenia. Motivationally, the proportions of anger among non-
schizophrenic murderers were the highest across groups for time periods before and
during offenses, and are consistent with aforementioned findings of comparatively
higher percentages of anger motivation in this group. However, at this point these
findings can only be considered exploratory—as statistical analyses could not be
employed due to lack of statistical power due to small group sizes across descriptor
choices. Ultimately, the subjective emotional experience accompanying homicidal
violence (particularly in the time periods before, during, and after the homicide) in
schizophrenic murderers may be a direct translation or derivative of their unique
biopsychosocial profile, and may provide valuable insight into the nature of the violence
characterizing this subgroup of schizophrenic persons.
In summary, forensic clinical data related to the motivational, behavioral, and
affective concomitants of homicide suggest qualitative differences in the murders
perpetrated by individuals with and without schizophrenia, and by individuals with other
psychiatric conditions. For schizophrenic murderers, these differences were largely
115
consistent with the patterns of biological and psychosocial deficits which
distinguished them from other diagnostic groups. As such, they may provide first
evidence for the possible translation of biopsychosocial functioning into a distinct
homicidal behavioral and emotional experience as it occurs within this violent
schizophrenic subtype, as well as a unique vantage point for elucidating its underlying
etiological mechanisms.
A Biopsychosocial Interactional Pathway to Schizophrenic Violence
The present study’s third aim was to examine group differences in
biopsychosocial functioning and homicidal affective/behavioral factors for any
indication of a possible etiological pathway toward homicidal violence unique to
murderers with schizophrenia—distinct from their non-murderer and non-schizophrenic
counterparts. In aggregate, the aforementioned evidence suggests such a biopsychosocial
interactional trajectory toward violence in schizophrenia, which could be proposed here
as follows: Paternal germ line mutations, due to advanced paternal birth age, may have
contributed to neurodevelopmental abnormalities leading to left hemispheric dysfunction
in the schizophrenic murderer group—as the risk pathway of de novo mutations in
paternal age-related schizophrenia is thought to be in part through reductions in more-
globalized cognition (i.e., intelligence; Malaspina et al., 2005), which would be
consistent with the broad-based (as opposed to more-localized) left hemispheric EEG
dysfunction seen in schizophrenic murderers here. These left-hemispheric hyperarousal
deficits may have in schizophrenic murderers subsequently interacted with or been
exacerbated by psychosocial stressors or psychosocial dysfunction specific to their
116
group; wherein a negative physical environment (i.e., homelessness) or emotional
environment (i.e., lack of familial closeness, married but living alone)—in combination
with childhood adversity (i.e., childhood psychosocial deprivation)—may provide an
increased amount of negative environmental and experiential cues available for context-
independent processing and contribute to an increased vulnerability to confusion, threat
perception, and subsequent violence in a schizophrenic individual.
Importantly, the biopsychosocial interactional trajectory toward schizophrenic
violence proposed here is only one interpretation of the present data, which itself may be
limited to the unique characteristics of this sample. Future studies could validate this
particular trajectory in other samples, or perhaps use it as a template for hypothesis
testing in different populations. On balance, future researchers could incorporate other
measures of biological and psychosocial functioning to develop and propose their own
empirically-based trajectories to be used as frameworks for understanding violent
behavior in schizophrenic persons. Ultimately, violence in schizophrenia is a highly-
complex phenomenon, and deciphering the enigmatic relationship between these two
conditions will likely require empirical efforts which incorporate multifaceted, biosocial
interactional research approaches.
Schizophrenic Murderers: Distinct Subgroup or a “Perfect Storm”?
Taken together, findings from the present study provide some additional
evidence for a subgroup of schizophrenic murderers that is distinct—either biologically,
psychosocially, or both—from their non-murderer and non-schizophrenic counterparts.
However, the existence of a distinct subgroup was not definitively suggested here. Only
117
two variables—current marital status and birth order—differentiated schizophrenic
murderers from both schizophrenic non-murderers and non-schizophrenic murderers.
Other findings merely indicated how schizophrenic murderers differed from either
schizophrenic non-murderers or non-schizophrenic murderers separately (i.e., one group
but not the other)—which may suggest group differences attributable to the other
condition in general (i.e., either schizophrenia or homicide), or perhaps the additive
effects of both conditions, rather than a unique and separable subgroup. As such, the
present study offered only minimal support for the subgroup hypothesis.
An alternative explanation may be that findings here represent a unique
combination of biological and psychosocial risk factors which, when combined with
schizophrenia, predispose a schizophrenic individual to violence. This biopsychosocial
“perfect storm” could explain why only some individuals with schizophrenia become
violent. More than simply the additive effects of a number of general risk factors for
violence, this hypothesis would attribute schizophrenic violence to a unique combination
of schizophrenia with general intellectual dysfunction, intact executive functioning,
abnormally elevated childhood psychosocial deprivation, higher birth order, lack of
family emotional closeness, some current/recent psychosocial stressors (i.e.,
homelessness) but relatively normative psychosocial functioning in other areas (which
may facilitate systemic failure of social support networks, due to perceptions by others
that caregiving is unwarranted), non-medicated status or using antipsychotics which are
not effective, along with left hemispheric hyperarousal which contributes to over-
118
analysis of content-independent information and subsequent vulnerability to threat
perception and violence.
Homicide Psychiatric Controls
A separate group of murderers characterized by psychiatric conditions other than
schizophrenia was included in the present study for a specific and important purpose—to
ascertain if any of the effects observed in schizophrenic murderers could be attributed
more so to violence within a broader range of mental illness rather than schizophrenia in
particular. One prominent and recurrent theme among study findings suggests that this
may have been so—the fact that no group differences between schizophrenic murderers
and homicide psychiatric controls were revealed in any area of biological or
psychosocial functioning. This may indicate that the two groups were largely similar in
these respects, and that any such differences observed between schizophrenic murderers
and other groups may have been due to the additive effects of mental illness in general
rather than those of schizophrenia in particular, or of a biopsychosocially distinct
homicidal schizophrenic subgroup. As such, the inclusion of a separate homicide
psychiatric control group may have effectively served its purpose.
However, evidence for this general mental illness explanation appears mixed.
For example, support may be suggested in the analysis of one categorical variable—
recent homelessness—where similar percentages among these two groups were
indicated; and possibly inferred in another—EEG beta3 activity—where the two groups
are characterized by similar increased mean amplitudes. In the latter, however,
amplitude means of homicide psychiatric controls were accompanied by comparatively
119
larger SDs, and boxplots of these data indicate that means were in fact artificially
inflated by outliers. Additionally, group differences replicated across non-schizophrenic
murderer-schizophrenic murderer comparisons and non-schizophrenic murderer-
psychiatric control comparisons would also suggest influences more attributable to
general mental illness than schizophrenia in particular—though no effects of this sort
were observed here. One other way to corroborate schizophrenic murderer-homicide
psychiatric control comparison findings might have been to utilize an additional group
of nonviolent individuals with mental illnesses other than schizophrenia, to see if effects
could be replicated with homicide removed from the equation (i.e., schizophrenic non-
murderer – psychiatric non-murderer control comparisons). Though this was not
possible in the present study, future research could incorporate such a control group to
provide further clarification.
On balance, characteristics among murderers unique to those with schizophrenia
(as opposed to general mental illness) may have been indirectly revealed in comparisons
which yielded normative or more-normative functioning in schizophrenic murderers but
not homicide psychiatric controls (i.e., independent living arrangements, number of
close friends). Additionally, failure to find group differences in other areas may actually
reflect aberrant homicide psychiatric control group data distributions with unusually
large SDs or outliers which could not be compensated for even with the use of modern
statistical methods (a significant possibility given the diagnostic heterogeneity of this
group); or the more conservative level of Type I error correction required in mean
comparison analyses involving five groups rather than four. In fact, given the range of
120
psychopathology among these participants, the question remains as to whether or not
homicide psychiatric controls in the present study even comprised a valid diagnostic
group. Perhaps the inclusion of a homicidal control group characterized by more
homogeneity in mental illness (i.e., a single class of disorders such as mood disorders)
would have provided further clarification. Ultimately, the factors here which
differentiate schizophrenic individuals who murder from those who do not cannot be
entirely parsed from the influences of mental illness in general, and this very important
research question should be among the key methodological considerations in future
examinations of violent schizophrenic persons.
Limitations and Strengths of the Present Study
The primary methodological limitations of the proposed study include its
reliance upon EEG, which is characterized by relatively poor spatial resolution and
inferiority to brain imaging techniques (Raine, 1993). Also, the study’s relatively small
diagnostic group sizes may have affected affect statistical power and generalizability.
Additionally, the present study’s reliance upon self-report data may be another
limitation, as the veracity of this type of data may be questionable, and it may be
susceptible to memory effects. This could be an especially important consideration for
those measures of psychosocial deprivation which relied heavily upon recall of early
childhood experiences; and may be of particular relevance to the schizophrenic murderer
group given the comparative memory deficits observed in antisocial schizophrenic
persons (Schug & Raine, 2009). Moreover, the limitations of cultural and language
factors which may have affected assessment in psychodiagnostic, neuropsychological,
121
psychosocial, and forensic realms must be considered. This was at least partially
addressed by the use of the Chinese psychodiagnostic system (the CCMD-3), the
Chinese translation of the WAIS-R (the WAIS-RC), and a culture-free version of the
Trailmaking Test (the CTT). However, transcultural applications of standardized
questionnaires and psychological testing instruments must be done cautiously, and the
problems of equivalency and validity must be considered before generalizing
conclusions cross-culturally (Tseng, Griffith, Ruiz, & Buchanan, 2007). As such, all
assessment-derived findings here must be considered within this important cultural
context. Furthermore, interpretation of group differences between schizophrenic non-
murderers and schizophrenic murderers must be made with caution, given the
confounding of medication and hospital inpatient status with schizophrenic non-
murderer group membership, and the possible differences in current and history of
psychiatric symptomatology which could not be addressed here.
Lastly, the concepts of homicide as a crime and the homicide offender as an
individual may comprise a culturally-specific context for interpretation of the present
findings, and cultural differences in the meanings of and attitudes toward violence must
also be considered. For example, the most-recent statistics published by the World
Health Organization (Krug, Dahlberg, Mercy, Zwi, & Lozano, 2002) indicate that the
overall homicide rate in the United States is nearly four times that of China (6.9 versus
1.8 deaths per 100,000 population, respectively), which suggests that China is
comparatively a much more nonviolent society than the United States. This reduced
violence may be attributable to collectivist Asian culture (Lim & Chang, 2009), or the
122
deterrent effects of China’s efficient and severe Yanda policy and death penalty
practice—which are responsible for what is thought to be upwards of 10,000 executions
per year in the People’s Republic of China (Trevaskes, 2008). Nonetheless, relative to
what appears to be a reduced baseline for violence, the crime of homicide in China may
represent a particularly brutal and violent behavior compared to other countries (e.g., the
United States) where this crime is more common. Thus, the generalizability of the
present study’s findings on homicide offenders to nations with much higher rates of
homicide and violence may be somewhat limited.
These limitations, however, are offset by several methodological strengths.
These include a design incorporating five diagnostic groups allowing for comparisons of
murderers and non-murderers, with and without schizophrenia, and for comparisons
with healthy controls to allow for the assessment of comparative group deficits within
the context of normative functioning. Additionally, the inclusion of a homicide
psychiatric control group allows for interpretations of potential observed differences in
schizophrenic and non-schizophrenic murderers beyond those ascribable to mental
illness in general. Furthermore, the study’s EEG hypotheses are more theory-driven,
which addresses a recently-stated need in the literature on crime and violence (Raine,
1993). Together, these elements serve to significantly strengthen the present study’s
methodology, relevance, and generalizability of findings in other ways, which largely
compensate for perceived study weaknesses.
123
Implications
The identification of a distinct homicidal schizophrenia subtype may have
important implications in research, treatment, forensic, and social realms. First, research
endeavors of either schizophrenia or homicidal behavior separately must take into
account the potential presence and confounding influence of the other disorder—that is,
investigations of schizophrenia populations should consider violent behavior or
propensities for violence as a moderating factor, while studies of homicide should
consider intervening schizophrenia symptomatology. Failure to do so may obscure or
even mask potential findings and lead to spuriously heterogeneous results across studies.
For example, disparate results achieved in past studies of electrodermal functioning in
schizophrenia (see Schug et al., 2007) could be clarified by considerating violent
antisociality as a moderator.
Second, identification of specific areas of biopsychosocial dysfunction in
schizophrenic murderers may facilitate the development of treatment approaches
effective beyond those employed in their schizophrenic non-murderer and non-
schizophrenic murderer counterparts. Pharmacological treatments of both schizophrenia
and violence separately are complex on their own (Bilder, 2006; Martell, 1992;
Moncrieff, 2003) and even more challenging when both conditions are present
simultaneously (Volavka, 2006). A greater understanding of the biopsychosocial nature
of violent schizophrenia might lead to the development of pharmacological approaches
specifically tailored for this group—perhaps targeting symptoms more-related to left
hemispheric over-processing or specific psychosocial stressors or backgrounds. Non-
124
pharmacological interventions targeting criminogenic needs (Gendreau, Goggin,
French, & Smith, 2006)—proven effective in non-psychotic offender populations
currently being adapted and tested with schizophrenic offenders (Hodgins, Tiihonen, &
Ross, 2005)—may benefit from incorporating approaches designed to address cognitive
impairment (see Gold, 2004), focusing upon left hemispheric hyperarousal deficits along
with the rehabilitation of psychosocial functioning, while perhaps taking advantage of
intact executive abilities. Furthermore, a greater understanding of the etiological
processes at work in violent schizophrenia may lead to the identification of the
biological and psychosocial precursors of this condition—information which could be
incorporated into intervention programs for children and adolescents designed to reduce
violent behavior before the onset of schizophrenia (Hodgins, 2004).
Finally, the identification of a biologically and psychosocially distinct violent
subtype of schizophrenia may have important social implications. Identifying a separate
group of individuals with schizophrenia who are prone to homicidal violence may serve
to reduce the stigma attached to schizophrenia in general, which is significantly
amplified by sensational (albeit rare) cases of violence perpetrated by schizophrenia
individuals. This stigma—which undermines efforts to obtain important community
resources for schizophrenia individuals such as treatment and housing (Hodgins,
2004)—arguably constitutes a grave social injustice against individuals with
schizophrenia who are not violent; and does little to advance the cause of those who are.
It is hoped that the results of the present study may contribute to a better understanding
125
of the violence observed in schizophrenia, and assist in moving society toward a
greater acceptance of individuals with this illness in general.
Conclusions
In summary, no single type of deficit—biological or psychosocial—appears to
distinguish schizophrenic murderers simultaneously from both their non-schizophrenic
and non-murderer counterparts. Rather, in very specific ways, schizophrenic murderers
are different from schizophrenic non-murderers, non-schizophrenic murderers, and
normal controls—both together and individually. In some areas, schizophrenic
murderers are comparatively worse of; while in others, contrary to expectations, they
appear quantitatively and qualitatively more similar to normal controls than their
schizophrenic non-murderer counterparts. Biologically, schizophrenic murderers do not
demonstrate characteristic EEG slowing or traditional responses to antipsychotic
medications and demonstrate better executive functioning abilities in very specific areas
compared to schizophrenic non-murderers. However, they do demonstrate increased left-
hemispheric fast-wave activity, indicative of overarousal—which may reflect cognitive
reasoning processes hypersensitive to threat perception and prone to violence.
Psychosocially, they are in some ways better-equipped functionally than non-
schizophrenic murderers and schizophrenic non-murderers, while simultaneously
characterized negatively by recent psychosocial stressors such as homelessness, along
with psychosocially-deprived childhoods and higher birth order. Additionally, the
homicide-related behaviors of schizophrenic murderers were qualified in some ways
consistently with the aforementioned patterns of biological and psychosocial deficits, as
126
well as the cognitive, affective, and interpersonal deficits known to accompany
schizophrenic symptomatology (i.e., lack of offense planning and concomitant sexual
offenses, comparatively higher prevalence of irrational motivation for murder, unusual
offense-related affective experiences). All of these factors, separately, have
demonstrated empirical relationships with violent behavior; and group differences found
here are consistent with what is known about an increased propensity for violence and
antisocial behavior in general. When inter-laced, however, they may represent an
etiological pathway or a biopsychosocial risk factor profile that is unique to
schizophrenic persons who become violent; and which may have important implications
in research, treatment, forensic applications, and social perception.
127
Bibliography
American Psychological Association. (1994). Diagnostic and statistical manual of
mental disorders (4
th
ed.). Washington, DC: Author.
Army Individualizad Test Battery (1944). Manual of directions and scoring.
Washington DC: War Department, Adjutant General’s Office.
Arrigo, B. A., & Purcell, C. E. (2001). Explaining paraphilias and lust murder: Toward
an integrated model. International Journal of Offender Therapy and
Comparative Criminology, 45(1), 6-31.
Arseneault, L., Moffit, T. E., Caspi, A., & Taylor, P. J. (2000). Mental disorder and
violence on a total birth cohort: Results from the Dunedin study. Archives of
General Psychiatry, 57, 979-986.
Aylward, E., Walker, E., & Bettes, B. (1984). Intelligence in schizophrenia: Meta-
analysis of the research. Schizophrenia Bulletin, 10(3), 430-459.
Barber, F. (1994). An investigation of the neuropsychological correlates of extreme
violence in schizophrenia patients. Unpublished master’s thesis, University of
Surrey, Guildford, Surrey, United Kingdom.
Barkataki, I., Kumari, V., Das, M., Hill, M., Morris, R., O’Connell, P., Taylor, P., &
Sharma, T. (2005). A neuropsychological investigation into violence and mental
illness. Schizophrenia Research, 74, 1-13.
Basar-Eroglu, C., Brand, A., Hildebrandt, H., Kedzior, K. K., Mathes, B., & Schmiedt,
C. (2007). Working memory related gamma oscillations in schizophrenia
patients. International Journal of Psychophysiology, 64, 39-45.
Basile, L. F. H., Yacubian, J., Ferreira, B. L. C., Valim, A. C., & Gataz, W. F.
(2004). Topographic abnormality of slow cortical potentials in schizophrenia.
Brazilian Journal of Medical and Biological Research, 37, 97-109.
Bebbington, P., & Kuipers, E. (2008). Psychosocial factors. In K. T. Mueser & D. V.
Jeste (Eds.), Clinical handbook of schizophrenia (pp. 74-81).New York: The
Guilford Press.
Bechara, A., Damasio, H., Damasio, A. R., & Lee, G. P. (1999). Different contributions
of the human amygdala and ventromedial prefrontal cortex to decision-making.
The Journal of Neuroscience, 19(13), 5473-5481.
128
Bender, L., & Curran, F. (1940). Children and adolescents who kill. Criminal
Psychopathology, 1(4), 297-322.
Bilder, R. M. (2006). Schizophrenia. In P. J. Snyder, P. D. Nussbaum, & D. L.
Robins (Eds.), Clinical neuropsychology (2
nd
ed.), pp. 155-182. Washington DC:
American Psychological Association.
Bjørkly, S. (2002). Psychotic symptoms and violence toward others—a literature
review of some preliminary findings: Part 1. Delusions. Aggression and violent
behavior, 7, 617-631.
Bilder, R. M. (2006). Schizophrenia. In P. J. Snyder, P. D. Nussbaum, & D. L.
Robins (Eds.), Clinical neuropsychology (2
nd
ed.), pp. 155-182. Washington DC:
American Psychological Association.
Blair, J., & Frith, U. (2000). Neurocognitive explanations of the antisocial
personality disorders. Criminal Behaviour and Mental Health, 10, S66-S81.
Blake, P. Y., Pincus, J. H., & Buckner, C. (1995). Neurologic abnormalities in
murderers. Neurology, 45, 1641-1647.
Bleuler, E. (1911). Dementia Praecox or the group of schizophrenias. Reprinted 1950.
Zinkin, J. (Trans. And Ed.). New York: International University Press.
Borison, R. L., Pathiraja, A., Haverstock, S., Gowda, S., & Diamond, B. I. (1995).
Standard and novel antipsychotic drugs in schizophrenia. In C. L. Shriqui & H.
A. Nasrallah (Eds.), Contemporary issues in the treatment of schizophrenia (pp.
247-293). Washington, DC: American Psychiatric Press, Inc.
Bowlby, J. (1946). Forty-four juvenile thieves: their characters and home-life.
Oxford, England: Baillière, Tindall & Cox.
Brennan, P. A., & Alden, A. (2006). Schizophrenia and violence: The overlap. In A.
Raine (Ed.), Crime and schizophrenia: Causes and cures (pp. 15-27). New York:
Nova Science Publishers, Inc.
Brennan, P. A., Mednick, S. A., & Hodgins, S. (2000). Major mental disorders and
riminal violence in a Danish birth cohort. Archives of General Psyciatry, 57,
494-500.
Cantor-Graae, E., McNeil, T. F., Sjöström, K., Nordström, L. G., & Rosenlund, T.
(1997). Maternal demographic correlates of increased history of obstetric
complications in schizophrenia. Journal of Psychiatric Research, 31(3), 347-357.
129
Centorrino, F., Price, B. H., Tuttle, M., Bahk, W. M., Hennen, J., Albert, M. J., &
Baldessarini, R. J. (2002). EEG abnormalities during treatment with typical and
atypical antipsychotics. American Journal of Psychiatry, 159(1), 109-115.
Chinese Society of Psychiatry (2001). The Chinese Classification and Diagnostic
Criteria of Mental Disorders Version 3 (CCMD-3). Jinan, Chinese Society of
Psychiatry.
Clementz, B. A., Sponheim, S. R., Iacono, W. G., & Beiser, M. (1994). Resting EEG in
first-episode schizophrenia patients, bipolar psychosis patients, and their first-
degree relatives. Psychophysiology, 31(5), 486-494.
Cole, K. E., Fisher, G., Cole, S. S. (1968). Women who kill. Archives of General
Psychiatry, 19, 1-8.
Cornell, D. C., Roberts, M., & Oram, G. (1997). The Rey-Osterrieth Complex Figure
Test as a neuropsychological measure in criminal offenders. Archives of Clinical
Neuropsychology, 12(1), 47-56.
Dalén, P. (1988). Schizophrenia, season of birth, and maternal age. British Journal of
Psychiatry, 153, 727-733.
Davison, A. C., & Hinkley, D. V. (1997). Bootstrap methods and their application.
Cambridge: Cambridge University Press.
D’Elia, L. F., Satz, P., Uchiyama, C. L., & White, T. (1996). Color trails test:
Professional manual. Odessa, FL: Psychological Assessment Resources, Inc.
Deiker, T. E. (1973). WAIS characteristics of indicted male murderers. Psychological
Reports, 32, 1066.
DeWolfe, A. S., & Ryan, J. J. (1984). Wechsler Performance IQ > Verbal IQ index in
a forensic sample: A reconsideration. Journal of Clinical Psychology, 40(1),
291-294.
Dickinson, D., Ramsey, M. E., & Gold, J. M. (2007). Overlooking the obvious: A
meta-analytic comparison of Digit Symbol Coding tasks and other cognitive
measures in schizophrenia. Archives of General Psychiatry, 64, 532-542.
Dolan, M., & Park, I. (2002). The neuropsychology of antisocial personality disorder.
Psychological Medicine, 32(3), 417-427.
Driver, M. V., West, L. R., & Faulk, M. (1974). Clinical and EEG studies of prisoners
charged with murder. British Journal of Psychiatry, 125, 583-587.
130
Eriksson, Å (2006). Risk factors for criminal offending among men with
schizophrenia. Unpublished master’s thesis, Karolinska Institutet, Stockholm,
Sweden.
Eronen, M., Hakola, P., & Tiihonen, J. (1996). Mental disorders and homicidal behavior
in Finland. Archives of General Psychiatry, 53(6), 497-501.
Evans, J. R., & Park, N. S. (1997). Quantitative EEG findings among men convicted of
murder. Journal of Neurotherapy, 2(2), 31-37.
Efron, B., & Tibshirani, R. J. (1993). An introduction to the bootstrap. New York:
Chapman and Hall.
Farrington, D. P. (2006). Family background and psychopathy. In C. J. Patrick (Ed.),
Handbook of psychopathy (pp. 229-250). New York: The Guilford Press.
Felix, A., Herman, D., & Susser, E. (2008). Housing instability and homelessness. In K.
T. Mueser & D. V. Jeste (Eds.), Clinical handbook of schizophrenia (pp. 411-
423).New York: The Guilford Press.
Fink, M. (2002). EEG changes with antipsychotic drugs. American Journal of
Psychiatry, 159, 1439.
Fink, M., & Kahn, R. L. (1957). Relation of EEG delta activity to behavioral response
to electroshock: Quantitative serial studies. Archives of General Psychiatry, 39,
1189-1191.
First, M. B., Gibbon, M., Spitzer, R. L., Williams, J. B. W., & Benjamin, L. S. (1997).
User’s guide for the Structured Clinical Interview for DSM-IV Axis II
Personality Disorders (SCID-II). Washington, DC: American Psychiatric Press.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1997). User’s guide for
the Structured Clinical Interview for DSM-IV Axis I Disorders—Clinician
Version (SCID-CV). Washington, DC: American Psychiatric Press.
Fisher, S. S. (1999). Juvenile males who murder: A descriptive study. Unpublished
doctoral dissertation, California School of Professional Psychology, Fresno.
Fox, J. A., & Zawitz, M. W. (2007). Homicide trends in the U.S.: Trends by gender.
Retrieved May 1, 2009, from U.S. Department of Justice, Office of Justice
Programs, Bureau of Justice Statistics Web site:
http://www.ojp.usdoj.gov/bjs/homicide/gender.htm
131
Freudenreich, O., Weiner, R. D., & McEvoy, J. P. (1997). Clozapine-induced
electroencephalogram changes as a function of clozapine serum levels.
Biological Psychiatry, 42, 132-137.
Gallinat, J., Winterer, G., Herrmann, C. S., & Senkowski, D. (2004). Reduced
oscillatory gamma-band responses in unmedicated schizophrenic patients
indicate impaired frontal network processing. Clinical Neurophysiology, 115,
1863-1874.
Garver, D. L., Holcomb, J. A., & Christensen, J. D. (2000). Heterogeneity of response to
antipsychotics from multiple disorders in the schizophrenia spectrum. Journal of
Clinical Psychiatry, 61(12), 964-972.
Gatzke-Kopp, L. M., Raine, A., Buchsbaum, M., & LaCasse, L. (2001). Temporal lobe
deficits in murderers: EEG findings undetected by PET. Journal of
Neuropsychiatry and Clinical Neuroscience, 13(4), 486-491.
Gaughran, F., Blizard, R., Mohan, R., Zammit, S., & Owen, M. (2007). Birth order and
the severity of illness in schizophrenia. Psychiatry Research, 150, 205-210.
Gendreau, P., Goggin, C., French, S., & Smith, P. (2006). Practicing psychology in
correctional settings. In I. B. Weiner & A. K. Hess (Eds.), The handbook of
forensic psychology (3
rd
ed., pp. 722-750). Hoboken, New Jersey: John Wiley &
Sons, Inc.
Gold, J. M. (2004). Cognitive deficits as treatment targets in schizophrenia.
Schizophrenia Research, 72, 21-28.
Gong, Y. X. (1992). Manual for the Wechsler Adults Intelligence Scale: Revised in
China (2nd ed.). Changsha, Hunan, China: Hunan Medical College.
Green, J., Leon-Barth, C., Venus, S., & Lucey, T. (2001). Murder and the EEG. The
Forensic Examiner, 10(1-2), 32-34.
Häkkänen, H., & Laajasalo, T. (2006). Homicide crime scene behaviors in a Finnish
sample of mentally ill offenders. Homicide Studies, 10(1), 33-54.
Hall, H. V., Thompson, J. S., & Poirier, J. G. (2007). Detecting deception in
neuropsychological cases: Toward an applied model. The Forensic Examiner,
16(3), 7-15.
Harbort, S., & Mokros, A. (2001). Serial murderers in Germany from 1945 to 1995: A
descriptive study. Homicide Studies, 5(4), 311-334.
132
Heilbrun, Jr., A. B. (1982). Cognitive models of criminal violence based upon
intelligence and psychopathy levels. Journal of Consulting and Clinical
Psychology, 47(3), 509-516.
Heinrichs, R. W., & Zakzanis, K. K. (1998). Neurocognitive deficit in schizophrenia:
A quantitative review of the literature. Neuropsychology, 12(3), 426-445.
Hill, D. H., & Pond, D. A. (1952). Reflections on one hundred capital cases submitted
to electroencephalography. Journal of Mental Sciences, 98, 23-43.
Hirose, S. (1979). Depression and homicide: A psychiatric and forensic study of four
cases. Acta Psychiatrica Scandinavica, 59, 211-217.
Hodgins, S. (2004). Criminal and antisocial behaviours and schizophrenia: A neglected
topic. In W. F. Gattaz & H. Häfner (Eds.), Search for the causes of
schizophrenia—Volume V (pp. 315-341). New York: Springer-Verlag.
Hodgins, S., Tiihonen, J., & Ross, D. (2005). The consequences of Conduct Disorder
for males who develop schizophrenia: Associations with criminality, aggressive
behavior, substance use, and psychiatric services. Schizophrenia Research, 78,
323-335.
Holcomb, W. R., Adams, N. A., & Ponder, H. M. (1985). The development and cross-
validation of an MMPI typology of murderers. Journal of Personality
Assessment, 49(3), 240-244.
Hollister, J. M., Laing, P., & Mednick, S. A. (1996). Rhesus incompatibility as a risk
factor for schizophrenia in male adults. Archives of General Psychiatry, 53(1),
19-24.
Hong, L. E., Symmerfelt, A., McMahon, R., Adami, H., Francis, G., Elliot, A.,
Buchanan, R. W., & Thaker, G. K. (2004). Evoked gamma band synchronization
and the liability for schizophrenia. Schizophrenia Research, 70, 293-302.
Hugdahl, K. (1995). Psychophysiology: The mind-body perspective. Cambridge,
Massachusetts: Harvard University Press.
Hughes, C., Kumari, V., Soni, W., Das, M., Binneman, B., Drozd, S., O’Neil, S.,
Mathew, V., & Sharma, T. (2002). Longitudinal study of symptoms and
cognitive function in chronic schizophrenia. Schizophrenia Research, 59, 137-
146.
133
Iacono, W. G. (1982). Bilateral electrodermal habituation-dishabituation and resting
EEG in remitted schizophrenics. Journal of Nervous and Mental Disease,
170(2), 91-101.
Ishikawa, S. S., & Raine, A. (2002). Psychophysiological correlates of antisocial
behavior: A central control hypothesis. In J. Glicksohn (Ed.), The neurobiology
of criminal behavior 8 (pp. 187-229). Norwell: Kluwer Academic Publishers.
Ishikawa, S. S., Raine, A., Lencz, T., Bihrle, S., & Lacasse, L. (2001). Autonomic stress
reactivity and executive functions in successful and unsuccessful criminal
psychopaths from the community. Journal of Abnormal Psychology, 110(3).
423-432.
Itil, T. M., Saletu, B., & Davis, S. (1972). EEG findings in chronic schizophrenics based
on digital computer period analysis and analog power spectra. Biological
Psychiatry, 5, 1-13.
Itil, T. M., Saletu, B., Davis, S., & Allen, M. (1974). Stability studies in
schizophrenics and normals using computer-analyzed EEG. Biological
Psychiatry, 8, 321-335.
Jamison, T. E. (2006). The homicidal narcissist. Unpublished doctoral dissertation,
University of Tennessee, Knoxville.
Joyal, C. C., Putkonen, A., Paavola, P., & Tiihonen, J. (2004). Characteristics and
circumstances of homicidal acts committed by offenders with schizophrenia.
Psychological Medicine, 34, 433-442.
Joyce, E. M., & Roiser, J. P. (2007). Cognitive heterogeneity in schizophrenia.
Current Opinion in Psychiatry, 20, 268-272.
Junginger, J. (2006). “Stereotypic” delusional offending. Behavioral Sciences and the
Law, 24, 295-311.
Kahn, M. W. (1967). Correlates of Rorschach reality adherence in the assessment of
murderers who plead insanity. Journal of Projective Techniques and Personality
Assessment, 31(4), 44-47.
Kahn, M. W. (1968). Superior Performance IQ of murderers as a function of overt act or
diagnosis. Journal of Social Psychology, 76, 113-116.
Kraepelin, E. (1919). Dementia Praecox and Paraphrenia. Livingstone, Edinburgh.
134
Krug, E. G., Dahlberg, L. L., Mercy, J. A., Zwi, A. B., & Lozano, R. (Eds.). (2002).
World report on violence and health. Geneva: World Health Organization.
Kumar, R., & Sharma, R. G. (1991). Pattern of memory deficits in paranoid and non-
paranoid schizophrenia. Indian Journal of Clinical Psychology, 18, 55-57.
Kumari, V., Aasen, I., Taylor, P., ffytche, D. E., Das, M., Barkataki, I., Goswami, S.,
O’Connell, P., Howlett, M., Williams, S. C. R., Sharma, T. (2006). Neural
dysfunction and violence in schizophrenia: An fMRI investigation.
Schizophrenia Research, 84, 144-164.
Laajasalo, T., & Häkkänen, H. (2006). Excessive violence and psychotic
symptomatology among homicide offenders with schizophrenia. Criminal
Behaviour and Mental Health, 16, 242-253.
Lahey, B. B., Hammer, D., Crumrine, P. L., & Forehand, R. L. (1980). Birth order X sex
interactions in child behavior problems. Developmental Psychology, 16(6), 608-
615.
Lamar, M., & Resnick, S. M. (2004). Aging and prefrontal functions: Dissociating
orbitofrontal and dorsolateral abilities. Neurobiology of Aging, 25, 553-558.
Langevin, R. (2003). A study of the psychosexual characteristics of sex killers: Can
we identify them before it is too late? International Journal of Offender Therapy
and Comparative Criminology, 47(4), 366-382.
Langevin, R., Ben-Aron, M. H., Wright, P., Marchese, V., & Handy, L. (1988). The sex
killer. Annals of Sex Research, 1, 263-301.
Langevin, R., Ben-Aron, M., Wortzman, G., Dickey, R., & Handy, L. (1987). Brain
damage, diagnosis, and substance abuse among violent offenders. Behavioral
Sciences and the Law, 5(1), 77-94.
Lee, T. M. C., & Chan, C. C. H. (2000). Are Trail Making and Color Trails Tests of
equivalent constructs? Journal of Clinical and Experimental Neuropsychology,
22(4), 529-534.
Lehrman, P. R. (1939). Some unconscious determinants in homicide. The Psychiatric
Quarterly, 13, 605-621.
Lenz, D., Krauel, K., Schadow, J., Baving, L., Duzel, E., & Herrmann, C. S. (2008).
Enhanced gamma-band activity in ADHD patients lacks correlation with
memory performance found in healthy children. Brain Research, 1235, 117-132.
135
Lewis, D. O., Lovely, R., Yeager, C., Ferguson, G., Friedman, M., Sloane, G.,
Friedman, H., & Pincus, J. H. (1988). Intrinsic and environmental characteristics
of juvenile murderers. Journal of the American Academy of Child and
Adolescent Psychiatry, 27(5), 582-587.
Lewis, D. O., Moy, E., Jackson, L. D., Aaronson, R., Restifo, N., Serra, S., & Simos, A.
(1985). Biopsychosocial characteristics of children who later murder: A
propesctive study. American Journal of Psychiatry, 142, 161-1167.
Lezak, M. D., Howieson, D. B., Loring, D. W., Hannay, H. J., & Fischer, J. S. (2004).
Neuropsychological assessment (4
th
ed.). New York: Oxford University Press.
Li, J. (1995). China’s one-child policy: How and how well has it worked? A case study
of Hebei province, 1979-88. Population and Development Review, 21(3), 563-
585.
Lim, L. L., & Chang, W. C. (2009). Role of collective self-esteem on youth violence in a
collective culture. International Journal of Psychology, 44(1), 71-78.
Lindberg, N., Tani, P., Virkkunen, M., Porkka-Heiskanen, T., Appelberg, B.,
Naukkarinen, H., & Salmi, T. (2005). Quantitative electroencephalographic
measures in homicidal men with antisocial personality disorder. Psychiatry
Research, 136, 7-15.
Lombroso, C. (1876). Criminal man. Milan: Hoepli.
Maj, M., D’Elia, L. D., Satz, P., Janssen, R., Zaudig, M., Uchiyama, C., Starace, F.,
Galderisi, S., & Chervinsky, A (1993). Evaluation of two new
neuropsychological tests designed to minimize cultural bias in the assessment of
HIV-1 Seropositive persons: A WHO study. Archives of Clinical
Neuropsychology, 8, 123-135.
Malaspina, D. A., Reichenberg, A. B., Weiser, M. C., Fennig, S. D., Davidson, M. B.,
Harlap, S. E., Wolitzky, R. A., Rabinowitz, J. F., Susser, E. E., Knobbler, H. Y.
C. (2005). Paternal age and intelligence: Implications for age-related genomic
changes in male germ cells. Psychiatric Genetics, 15(2), 117-125.
Maloney, M. P. (1985). A clinician’s guide to forensic psychological assessment.
New York: The Free Press.
Martell, D. A. (1992). Estimating the prevalence of organic brain dysfunction in
maximum-security forensic psychiatric patients. Journal of Forensic Sciences,
37(3), 878-893.
136
McCord, W., & McCord, J. (1964). The psychopath: An essay on the criminal mind.
Princeton, NJ: Van Nostrand.
Mednick, S. A., Volavka, J., Gabrielli, W. F., & Itil, T. (1982). EEG as a predictor of
antisocial behavior. Criminology, 19, 219-231.
Meloy, J. R., Gacono, C. B., & Kenney, L. (1994). A Rorschach investigation of
sexual homicide. Journal of Personality Assessment, 62(1), 58-67.
Merlo, M. C. G., Kleinlogel, H., & Koukkou, M. (1998). Differences in the EEG
profiles of early and late responders to antipsychotic treatment in first-episode,
drug-naïve psychotic patients. Schizophrenia Research, 30, 221-228.
Merrin, E. L., & Floyd, T. C. (1996). Negative symptoms and EEG alpha in
schizophrenia: A replication. Schizophrenia Research, 19, 151-161.
Mientus, S., Gallinat, J., Wuebben, Y., Pascual-Marqui, R. D., Mulert, C., Frick, K.,
Dorn, H., Herrmann, W. M., & Winterer, G. (2002). Cortical hypoactivation
during resting EEG in schizophrenics but not in depressives and schizotypal
subjects as revealed by low resolution electromagnetic tomography (LORETA).
Psychiatry Research Neuroimaging, 116, 95-111.
Miller, D. C. (1983). Handbook of research design and social measurement (4
th
ed.).
New York: Longman.
Miller, R. (1989). Schizophrenia as a progressive disorder: Relations to EEG, CT,
neuropathological and other evidence. Progress in Neurobiology, 33, 17-44.
Mitchell, D. G. V., Colledge, E., Leonard, A., & Blair, R. J. R. (2002). Risky
decisions and response reversal: Is there evidence of orbitofrontal cortex
dysfunction in psychopathic individuals? Neuropsychologia, 40, 2013-2022.
Mitrushina, M. N., Boone, K. B., & D’Elia, L. F. (1999). Handbook of normative data
for neuropsychological assessment. NY: Oxford University Press.
Miyauchi, T., Tanaka, K., Hagimoto, H., Miura, T., Kishimoto, H., & Matsushita,
M. (1990). Computerized EEG in schizophrenic patients. Biological Psychiatry,
28, 488-494.
Mockler, D., Riordan, J., & Sharma, T. (1997). Memory and intellectual deficits do
not decline with age in schizophrenia. Schizophrenia Research, 26, 1-7.
137
Moncrieff, J. (2003). Clozapine v. conventional antipsychotic drugs for treatment-
resistant schizophrenia: A re-examination. British Journal of Psychiatry, 183,
161-166.
Moran, P., & Hodgins, S. (2004). The correlates of comorbid antisocial personality
disorder in schizophrenia. Schizophrenia Bulletin, 30(4), 791-802.
Morgan, A. B., & Lilienfeld, S. O. (2000). A meta-analytic review of the relation
between antisocial behavior and neuropsychological measures of executive
function. Clinical Psychology Review, 20(1), 113-136.
Mundy-Castle, A. C. (1955). The EEG in twenty-two cases of murder or attempted
murder. Appendix on possible significance of alphoid rhythms. Journal of the
National Institute for Personnel Research, 6, 103-120.
Myers, W. C., Scott, K., Burgess, A. W., & Burgess, A. G. (1995). Psychopathology,
biopsychosocial factors, crime characteristics, and classification of 25 homicidal
youths. Journal of the American Academy of Child and Adolescent Psychiatry,
34(11), 1483-1489.
Nagase, Y., Okubo, Y., & Toru, M. (1996). Electroencephalography in schizophrenic
patients: Comparison between neuroleptic-naïve state and after treatment.
Biological Psychiatry, 40, 452-456.
Nestor, P. G., Haycock, J., Doiron, S., Kelly, J., & Kelly, D. (1995). Lethal violence and
psychosis: A clinical profile. Bulletin of the American Academy of Psychiatry
and the Law, 23(3), 331-341
Nestor, P. G., Kimble, M., Berman, I., & Haycock, J. (2002). Psychosis, psychopathy,
and homicide: A preliminary neuropsychological inquiry. American Journal of
Psychiatry, 159, 138-140.
Nijman, H., Cima, M., & Merckelbach, H. (2003). Nature and antecedents of psychotic
patients’ crimes. Journal of Forensic Psychiatry & Psychology, 14(3), 542-553.
Ninan, P. T. (1989). Pharmacological management of schizophrenia. In A. S. Bellack
(Ed.), A clinical guide for the treatment of schizophrenia (pp. 23-42). New York:
Plenum Press.
Osuji, I. J., McGarrahan, A., Mihalakos, P., Garver, D., Kingsbury, S., & Cullum, C.
M. (2007). Neuropsychological functioning in MRI-derived subgroups of
schizophrenia. Schizophrenia Research, 92, 189-196.
138
Pagan, D., & Smith, S. S. (1979). Homicide: A medico-legal study of thirty cases.
Bulletin of the American Academy of Psychiatry and the Law, 7(3), 275-285.
Palmer, B. W., Heaton, R. K., Paulsen, J. S., Kuck, J., Braff, D., Harris, M. J., et al.
(1997). Is it possible to be schizophrenic yet neuropsychologically normal?
Neuropsychology, 11, 437-446.
Patterson, R. M. (1942). Psychiatric study of juveniles involved in homicide.
American Journal of Orthopsychiatry, 13, 125-130.
Pillay, S. S., Stoll, A. L., Weiss, M. K., Tohen, M., Zarate, Jr., C. A., Banov, M. D., &
Cole, J. O. (1996). EEG abnormalities before clozapine therapy predict a good
clinical response to clozapine. Annals of Clinical Psychiatry, 8(1), 1-5.
Prouteau, A., Verdoux, H., Briand, C., Lesage, A., Lalonde, P., Nicole, L., Reinharz,
D., & Stip, E. (2005). Cognitive predictors of psychosocial functioning outcome
in schizophrenia: A follow-up study of subjects participating in a rehabilitation
program. Schizophrenia Research, 77, 343-353.
Raine, A. (1993). The psychopathology of crime: Criminal behavior as a clinical
disorder. San Diego, California: Academic Press.
Raine, A. (2006). Pursuing a second generation of research on crime and
schizophrenia. In A. Raine (Ed.), Crime and schizophrenia: Causes and cures
(pp. 3-12). New York: Nova Science Publishers, Inc.
Raine, A., Buchsbaum, M. S., & LaCasse, L. (1997). Brain abnormalities in
murderers indicated by positron emission tomography. Biological Psychiatry, 42,
495-508.
Raine, A., Meloy, J. R., Bihrle, S., Stoddard, J., LaCasse, L., & Buchsbaum, M. S.
(1998). Reduced prefrontal and increased subcortical brain functioning assessed
using positron emission tomography in predatory and affective murderers.
Behavioral Sciences and the Law, 16, 319-332.
Raine, A., Reynolds, C., Venables, P., Mednick, S. A., & Farrington, D. P. (1998).
Fearlessness, stimulation-seeking, and large body size at age 3 years as early
predispositions to childhood aggression at age 11 years. Archives of General
Psychiatry, 55, 745-751.
Raine, A., Stoddard, J., Bihrle, S., & Buchsbaum, M. (1998). Prefrontal glucose
deficits in murderers lacking psychosocial deprivation. Neuropsychiatry,
Neuropsychology, and Behavioral Neurology, 11(1), 1-7.
139
Ressler, R. K., Burgess, A. W., & Douglas, J. E. (1988). Sexual homicide: Patterns
and motives. New York: The Free Press.
Ritter, L. M., Meador-Woodruff, J. H., & Dalack, G. W. (2004). Neurocognitive
measures of prefrontal cortical dysfunction in schizophrenia. Schizophrenia
Research, 68, 65-73.
Rosanoff, A. J. (1943). Thirty condemned men. American Journal of Psychiatry, 99,
484-495.
Rund, B. R, Melle, I., Friis, S., Johannessen, J. O., Larsen, T. K., Midbøe, L. J.,
Opjordsmoen, S., Simonsen, E., Vaglum, P., & McGlashan, T. (2007). The
course of neurocognitive functioning in first-episode psychosis and its relation to
premorbid adjustment, duration of untreated psychosis, and relapse.
Schizophrenia Research, 91, 132-140.
Rymer, C. A. (1942). The insanity plea in murder. American Journal of Psychiatry,
98, 690-697.
Sajatovic, M., Madhusoodanan, S., & Fuller, M. A. (2008). Clozapine. In K. T. Mueser
& D. V. Jeste (Eds.), Clinical handbook of schizophrenia (pp. 178-185). New
York: The Guilford Press.
Sakuta, A., & Fukushima, A. (1998). A study on abnormal brain findings pertaining to
the brain in criminals. International Medical Journal, 5(4), 283-292.
Saletu, B., Kufferle, B., & Anderer, P. (1986). On the orienting response in
schizophrenics: Quantitative EEG studies before and during therapy. In C.
Shagass, R. C. Josiassen, & R. A. Roemer (Eds.), Brain electrical potentials and
psychopathology (pp. 59-81). New York: Elsevier.
Salva, G. N., Moore, D. J., & Palmer, B. W. (2008). Cognitive functioning in
schizophrenia. In K. T. Mueser & D. V. Jeste (Eds.), Clinical handbook of
schizophrenia (pp. 91-99). New York: The Guilford Press.
Schanda, H., Knecht, G., Schreinzer, D., Stompe, T., Ortwein-Swoboda, G., &
Waldhoer, T. (2004). Homicide and major mental disorders: A 25-year study.
Acta Psychiatrica Scandinavica, 110, 98-107.
Schenkel, L. S., Spaulding, W. D., DiLillo, D., & Silverstein, S. M. (2005). Histories
of childhood maltreatment in schizophrenia: Relationships with premorbid
functioning, symptomatology, and cognitive deficits. Schizophrenia Research,
76, 273-286.
140
Schmitt, W. A., Brinkley, C. A., & Newman, J. P. (1999). Testing Damasio’s
somatic marker hypothesis with psychopathic individuals: Risk takers or risk
averse? Journal of Abnormal Psychology, 108(3), 538-543.
Schug, R. A., & Fung, M. T. (2007). Dimensions of childhood psychosocial deprivation
in psychopaths and nonpsychopathic criminals. Poster presented at the Second
Biannual Conference of the Society for Scientific Study of Psychopathy, April
26-28, 2007, in St. Petersburg, Florida.
Schug, R. A., & Raine, A. (2009). Comparative meta-analyses of
neuropsychological functioning in antisocial schizophrenic persons. Clinical
Psychology Review, 29I, 230-242.
Schug, R. A., Raine, A., & Wilcox, R. R. (2007). Psychophysiological and behavioural
characteristics of individuals with both antisocial personality disorder and
schizophrenia-spectrum personality disorder. British Journal of Psychiatry, 191,
408-414.
Schwartz, R. C., Petersen, S., & Skaggs, J. L. (2001). Predictors of homicidal ideation
and intent in schizophrenia: An empirical study. American Journal of
Orthopsychiatry, 71(3), 379-384.
Seltzer, J., Conrad, C., & Cassens, G. (1997). Neuropsychological profiles in
schizophrenia: Paranoid versus undifferentiated distinctions. Schizophrenia
Research, 23, 131-138.
Shagass, C. (1991). EEG studies of schizophrenia. In S. R. Steinhauer, J. H. Gruzelier,
& J. Zubin (Eds.), Handbook of schizophrenia, vol. 5: Neuropsychology,
psychophysiology, and information processing (pp. 39-69). New York: Elsevier.
Sham, P. C., Maclean, C. J., & Kendler, K. S. (1993). Risk of schizophrenia and age
difference with older siblings: Evidence for a maternal viral infection
hypothesis? British Journal of Psychiatry, 163, 627-633.
Silver, H., Goodman, C., Knoll, G., Isakov, V., & Modai, I. (2005). Schizophrenia
patients with a history of severe violence differ from nonviolent schizophrenia
patients in perception of emotions but not cognitive function. Journal of Clinical
Psychiatry, 66(3), 300-308.
Snook, B., Cullen, R. M., Mokros, A., & Harbort, S. (2005). Serial murderers’
spatial decisions: Factors that influence crime location choice. Journal of
Investigative Psychology and Offender Profiling, 2, 147-164.
141
Spencer, K. M., Nestor, P. G., Niznikiewicz, M. A., Salisbury, D. F., Shenton, M. E.,
& McCarley, R. W. (2003). Abnormal neural synchrony in schizophrenia.
Journal of Neuroscience, 23(19), 7407-7411.
Spencer, K. M., Nestor, P. G., Perlmutter, R., Niznikiewicz, M. A., Klump, M. C.,
Frumin, M., Shenton, M. E., & McCarley, R. W. (2004). Neural synchrony
indexes disordered perception and cognition in schizophrenia. Proceedings of the
National Academy of Sciences, 101(49), 17288-17293.
Sponheim, S. R., Clementz, B. A., Iacono, W. G., & Beiser, M. (1994). Resting EEG in
first-episode and chronic schizophrenia. Psychophysiology, 31(1), 37-43.
Stafford-Clarke, D., & Taylor, F. H. (1949). Clinical and electro-encephalographic
studies of prisoners charged with murder. Journal of Neurology, Neurosurgery,
and Psychiatry, 12, 325-330.
Stanford, M. S., Conklin, S. M., Helfritz, L. E., & Kockler, T. R. (2007). P3 amplitude
reduction and executive function deficits in men convicted of spousal/partner
abuse. Personality and Individual Differences, 43, 365-375.
Stern, R. M., Ray, W. J., & Quigley, K. S. (2001). Psychophysiological recording.
New York: Oxford University Press.
Stevens, J. R. (1995). Clozapine: The yin and yang of seizures and psychosis.
Biological Psychiatry, 37, 425-426.
Swanson, J. W., Swartz, M. S., Van Dorn, R. A., Elbogen, E. B., Wagner, H. R.,
Rosenheck, R. A., Stroup, T. S., McEvoy, J. P., & Lieberman, J. A. (2006). A
national study of violent behavior in persons with schizophrenia. Archives of
General Psychiatry, 63, 490-499.
Teichner, G., Golden, C. J., Van Hasselt, V. B., & Peterson, A. (2001). Assessment of
cognitive functioning in men who batter. International Journal of Neuroscience,
111, 241-253.
Ticehurst, S. B., Ryan, M. G., & Hughes, F. (1992). Homicidal behaviour in elderly
patients admitted to a psychiatric hospital. Dementia, 3, 86-90.
Tiihonen, J., Isohanni, M., Räsänen, P., Koiranen, M., & Moring, J. (1997). Specific
major mental disorders and criminality: A 26-year prospective study of the 1966
Northern Finland birth cohort. American Journal of Psychiatry, 154(6), 840-845.
Trevaskes, S. The death penalty in China today: Kill fewer, kill cautiously. Asian
Survey, 48(3), 393-413.
142
Tseng, W. S., Griffith, E., Ruiz, P., & Buchanan, A. (2007). Culture and
psychopathy in the forensic context, In A. Felthous & H. Saβ (Eds.),
International handbook on psychopathic disorders and the law (pp. 473-488).
West Sussex, England: John Wiley & Sons Ltd.
Tsuchiya, K. J., Matsumoto, K., Miyachi, T., Tsujii, M., Nakamura, K., Takagai, S., et
al. (2008). Paternal age at birth and high-functioning autistic-spectrum disorder
in offspring. British Journal of Psychiatry, 193, 316-321.
Tsuchiya, K. J., Takagai, S., Kawai, M., Matsumoto, H., Nakamura, K., Minabe, Y.,
Mori, N., & Takei, N. (2005). Advanced paternal age associated with an elevated
risk for schizophrenia in offspring in a Japanese population. Schizophrenia
Research, 76, 337-342.
United States Census Bureau (n.d.). Annual estimates of the population by sex, race, and
Hispanic origin for the United States: April 1, 2000, to July 1, 2007 (NC-
EST2007-03). Retrieved May 1, 2009 from
http://www.census.gov/popest/national/asrh/NC-EST2007-srh.html
Valliant, P. M., Asu, M. E., Cooper, D., & Mammola, D. (1984). Profile of dangerous
and non-dangerous offenders for pre-trial psychiatric assessment. Psychological
Reports, 54, 411-418.
Varma, L. P., & Jha, B. K. (1966). Characteristics of murder in mental disorder.
American Journal of Psychiatry, 122(11), 1296-1298.
Venables, P. H., Raine, A., Dalais, C., Liu, J., & Mednick, S. A. (2006).
Malnutrition, cognitive ability and schizotypy. In A. Raine (Ed.), Crime and
schizophrenia: Causes and cures (pp. 131-150). New York: Nova Science
Publishers, Inc.
Vogel, F., & Schalt, E. (1979). The electroencephalogram (EEG) as a research tool in
human behavior genetics: psychological examinations in healthy males with
various inherited EEG variants. III. Interpretation of the results. Human
Genetics, 47, 81–111.
Volavka, J. (2006). Treatment approaches to aggressive behavior in schizophrenia.
In A. Raine (Ed.), Crime and schizophrenia: Causes and cures (pp. 301-314).
New York: Nova Science Publishers, Inc.
Wagner, K. A. (2007). Thuggee and social banditry reconsidered. The Historical
Journal, 50(2), 353-376.
143
Wagner, E. E., & Klein, I. (1977). WAIS differences between murderers and
attackers referred for evaluation. Perceptual and Motor Skills, 44, 125-126.
Warnick, E. K. (2007). Cognitive heterogeneity in murderers. Unpublished doctoral
dissertation, University of Nevada, Las Vegas.
Wechsler, D. (1981). The Wechsler Adult Intelligence scale –R. New York: The
Psychological Corporation.
Wichniak, A., Szafrański, T., Wierzbicka, A., Waliniowska, E., & Jernajczyk, W.
(2006). Electroencephalogram slowing, sleepiness and treatment response in
patients with schizophrenia during olanzapine treatment. Journal of
Psychopharmacology, 20(1), 80-85.
Wilcox, R. R. (2003). Applying contemporary statistical techniques. New York:
Academic Press.
Wilcox, R. R. (2005). New methods for comparing groups: Strategies for increasing the
probability of detecting true differences. Current Directions in Psychological
Science, 14, 272-275.
Wilk, C. M., Gold, J. M., McMahon, R. P., Humber, K., Iannone, V. N., & Buchanan, R.
W. (2005). No, it is not possible to be schizophrenic yet neuropsychologically
normal. Neuropsychology, 19(6), 778-786.
Wilson, J. Q., & Herrnstein, R. (1985). Crime and human nature. New York: Simon and
Schuster.
Winkler, G. E., & Kove, S. S. (1961). The implications of encephalographic
abnormalities in homicide cases. Journal of Neuropsychiatry, 3, 322-330.
Wittels, F. (1937). The criminal psychopath in the psychoanalytic system.
Psychoanalytic Review, 24(1), 276-291.
Wong, M. T. H., Fenwick, P. B. C., Lumsden, J., Fenton, G. W., Maisey, M. N.,
Lewis, P., & Badawi, R. (1997). Positron emission tomography in male violent
offenders with schizophrenia. Psychiatry Research: Neuroimaging Section, 68,
111-123.
Wuebben, Y., & Winterer, G. (2001). Hypofrontality—a risk marker related to
schizophrenia? Schizophrenia Research, 48, 207-217.
Yang, J. (2007). China’s one-child policy and overweight children in the 1990s. Social
Science and Medicine, 64, 2043-257.
144
Yarvis, R. M. (1990). Axis I and Axis II diagnostic parameters of homicide. Bulletin
of the American Academy of Psychiatry and the Law, 18(3), 249-269.
Zhang, Y., & Goza, F. W. (2006). Who will care for the elderly in China? A review
of the problems caused by China’s one-child policy and their potential solutions.
Journal of Aging Studies, 20, 151-164.
Zilboorg, G. (1935). Some sidelights on psychology of murder. Journal of Nervous and
Mental Disorders, 81, 442.
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Psychophysiological and behavioral characteristics of individuals with both antisocial personality disorder and schizophrenia-spectrum personality disorder
PDF
The development of electrodermal fear conditioning from ages 3 to 8 years and relationships with antisocial behavior at age 8 years
PDF
Psychometric and biometric evaluation of psychopathic personality traits in 9 and 10 year old twins
PDF
Early temperamental and psychophysiological precursors of adult psychopathic personality
PDF
Effects of nicotine abstinence on orienting, executive function, arousal and vigilance
PDF
Neural correlates of learning and reversal of approach versus withdraw responses to in- and out-group individuals
PDF
Age differences in diffusivity in the locus coeruleus and its ascending noradrenergic tract
Asset Metadata
Creator
Schug, Robert August
(author)
Core Title
Biopsychosocial and forensic clinical correlates of schizophrenia and homicide
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Psychology
Publication Date
06/14/2009
Defense Date
04/27/2009
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
EEG,forensic psychology,Homicide,neuropsychology,OAI-PMH Harvest,schizophrenia
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Raine, Adrian (
committee chair
), Baker, Laura A. (
committee member
), Trickett, Penelope K. (
committee member
), Dawson, Michael E. (
committee member
)
Creator Email
robertschug@hotmail.com,schug@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m2296
Unique identifier
UC1496811
Identifier
etd-Schug-2973 (filename),usctheses-m40 (legacy collection record id),usctheses-m2296 (legacy record id),usctheses-c127-241331 (legacy record id)
Legacy Identifier
etd-Schug-2973.pdf
Dmrecord
241331
Document Type
Dissertation
Rights
Schug, Robert August
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
EEG
forensic psychology
neuropsychology
schizophrenia