|
An association between epileptiform disorders and psychopathology
has long been recognized (1). Flor-Henry reported left temporal
EEG abnormalities in schizophrenia patients and bi-temporal, but
mainly right temporal, dysrhythmia in mood disorders (2). Flor-Henry
and Koles postulated that, in the psychoses, there is disorganization
of the right hemisphere that is least severe in depression, intermediate
in mania, and maximal in schizophrenia (3). They considered left
hemisphere disorganization to also exist in both mania and schizophrenia.
Cook and others found that bipolar patients with abnormal EEGs
had a significantly negative family history of mood disorder, compared
with patients with normal EEGs (4). They confirmed previous findings
by Dalen (5), Hays (6), and Kadzmas and Winokur (7) in the postulation
of “acquired mania” that occurs independently of a genetic
loading. However, Dewan and others were unable to replicate earlier
reports suggesting lateralization of these abnormalities or the
negative correlation between EEG abnormalities and family history
of mood disorder (8). Levy and others (9) found a positive family
history of mood disorder in their 3 patients with rapid-cycling
Bipolar Mood Disorder (BMD) with EEG paroxysmal sharp waves, which
also seems to contradict the earlier reports. A major limitation
of these studies is the fact that the clinical application of the
conventional EEG is limited by the qualitative evaluation of the
results.
The introduction of microcomputers and the quantitative analysis
of EEG (QEEG) variables with appropriate statistical methods offer
more objective and reliable mechanisms for evaluating and extracting
diagnostic and discriminating EEG variables. The clinical sensitivity
of electroencephalography to psychiatric disorders has been significantly
enhanced with the advent of QEEG. Despite several methodological
limitations, QEEG variables have been used to investigate brain
activity in psychiatric disorders. Relations between psychiatric
diagnostic categories and some QEEG variables have been examined
in the attempt to characterize the QEEG abnormalities specific to
a particular diagnosis (10). QEEG coherence and power provide measures
of the correlation and intensity between 2 EEG signals for any given
frequency band. Coherence gives an indication of which particular
brain regions are correlated (or working together) in either the
resting or activated state. Power is the QEEG-derived parameter
of the EEG voltage amplitude, which indicates the intensity of the
wave between particular brain regions in either the resting or activated
state.
Using QEEG variables, Shagass and others reported that patients
with schizophrenia and mania were alike in their differences from
control subjects, although the differences were of larger magnitude
for patients with mania (11). Further, they found no important differences
between the patients with schizophrenia and those with mania. Kano
and others found that patients with mania displayed decreased alpha
power at F7 and increased beta at F8, compared with depression patients
(12). Koles and others conducted QEEG studies of hemispheric relations
between patients with schizophrenia and those with mood disorders
by examining spatial patterns including power, coherence, and phase
relations (13). Left hemispheric disorganization occurred in both
groups, but more so in the patients with schizophrenia. Small and
others’ study compared a small subsample of drug-free patients
with mania with normal control subjects and revealed lower QEEG
amplitudes in the left anterior and midtemporal regions in the patients
(14). Although some investigators indicated that some EEG measures
are heavily determined by trait factors, including heredity, others
concluded that, given the different activation levels among diagnostic
groups, the measures reflect state more than trait factors (11,15).
These results remain inconclusive.
Our study assessed the QEEG absolute power and coherence differences
between a group of patients with BMD I and a group with schizophrenia.
We also examined the correlation between QEEG measures and family
history of BMD, testing the hypothesis that there is a significant
positive association between QEEG measures and a positive family
history of BMDs.
|
|
Method
Subjects
We included in the study 18 inpatients meeting DSM-III-R (16) criteria
for BMD. None had a prior history of neonatal asphyxia, epilepsy,
head injury, schizophrenia, delirium, or dementia.
All the patients were assessed by a psychiatrist using a structured
clinical interview. Approximate length of illness was 13 years.
Symptom severity was rated using the National Institutes of Mental
Health (NIMH) Global Rating Scales (17). Subjects were on an average
daily dosage of 1380 mg lithium. All as-needed (PRN) medications
(for example, antipsychotics and benzodiazepines) were held for
at least 48 hours prior to EEG recording. The presence or absence
of family history of BMD in first-degree relatives was also documented
via patient reports and available clinical information. All patients
were in the active manic-depressive or mixed phase at the time of
study.
Eighteen inpatients meeting DSM-III-R criteria for schizophrenia
(chronic) served as a comparison group. All were diagnosed using
clinical information and an interview with a psychiatrist using
the Structured Clinical Interview for DSM-III-R (SCID) (18). Approximate
length of illness was 9 years. In this group, medication averaged
580.6 mg chlorpromazine equivalents daily. This sample was drawn
from a larger group of patients (n = 30) on which we have previously
published reports (19,20). In previous reports, no differences were
found between schizophrenia patients and matched normal control
subjects on measures of resting QEEG indices (that is, alpha bandwidth)
in either eyes-open or eyes-closed conditions. For this reason,
we did not include a normal control group in the present study.
The subjects with schizophrenia and those with BMD were matched
on the variables of age, sex, and handedness. All subjects provided
informed, written consent prior to participation in the study.
Quantitative Electroencephalograpy
(QEEG)
We obtained EEG data using a QSI 9000 computer-based acquisition
system with Grass gold-plated cup electrodes positioned according
to the International 10/20 system (21) using that employs a linked-ears
reference (Figure
1). All subjects underwent EEG recording in the morning hours.
Electrode impedences were not greater than 5000 ohms. Artefacts
caused by eye movement (monitored via electro-olfactogram [EOG]),
drowsiness, muscle tension, or technical difficulties were excluded
via an automatic-rejection computer program and by visual inspection
of the record. Because it was extremely difficult to maintain a
continuous state of immobility in the patients with BMD, the minimum
criteria for maintaining a sample of EEG for analysis was 20 seconds
of artefact-free EEG data. The samples were averaged at 2.5 second
epochs, digitized at a rate of 102.4 samples per second, and filtered
at 0.5 to 30 Hz. Each digitized epoch of eyes-closed EEG data was
quantified using a Fast Fourier Transformation (FFT) (22). FFT amplitude
and spectra were averaged and summed together, and individual bandwidths
were computed. EEG alpha (8 to 12 Hz) bandwidth power and coherence
were calculated for 18 pairs of electrodes in both left and right
hemispheres (F3F7, F3C3, F3T3, F3T5, F3P3, F3O1, F7C3, F7T3, F7T5,
F7P3, F7O1, T5C3, T5T3, T5P3, T5O1, P3C3, P3T3, P3O1, and homologous
locations in the right hemisphere).
Statistical Analysis
The clinical and demographic characteristics of both groups were
evaluated using 1-way analysis of variance (ANOVA) and SPSS software
(23). Between-group, resting eyes-closed measures of alpha power
and coherence were also assessed using 1-way ANOVA at the 18 pairs
of electrodes in both left and right hemispheres. A Pearson r
correlation analysis between ratings of family history of BMD and
EEG power and coherence measures within the BMD subjects was also
conducted.
|
