EEG Signals in Psychiatry: Biomarkers for Depression Management - Quantitative EEG Analysis Methods and Clinical Applications - page 301

asymmetry measure to predict any TESI with 65% accuracy ( p = 0.14). Accuracy,

0.04) when considering only those subjects who
exhibited an increase of 2 or more points in suicidal ideation on item 3 of the
HamD 17 . These findings are promising but await replication including replication in
a placebo-controlled trial.

however, was higher (81%, p =

11.2.2.6 qEEG and the Placebo Response
Placebo response is of interest in psychiatry from a clinical perspective, as well as
from research and drug development perspectives. Placebo response rates are higher
for MDD than those for some other medical conditions [130], ranging from 20% to
80% [131]. On one hand, high placebo response rates are evidence of the potential
impact of nonpharmacological “placebo-related” factors such as patient expecta-
tions, classical conditioning effects, and the patient-physician relationship. On the
other hand, high placebo response rates may obscure the efficacy of specific inter-
ventions making it difficult to evaluate treatment effects. qEEG imaging may have
the potential to address both of these issues.
In a landmark study, serial EEGs were examined over 8 weeks of treatment in
51 MDD subjects assigned randomly to receive antidepressant medication
(fluoxetine or venlafaxine) or placebo [132]. Subjects were categorized according to
one of four outcome groups: medication responders, medication nonresponders,
placebo responders, or placebo nonresponders. Analysis of regional changes in
qEEG theta-band cordance revealed significant differences in the prefrontal region
in placebo responders as compared to all other groups; placebo responders uniquely
showed early increases in prefrontal cordance prior to achieving response. This
finding documents neurophysiological changes associated with placebo response in
depression, and suggests that drug and placebo response have at least some distinct
underlying mechanisms.
Results of this study can be considered alongside a fluorodeoxyglucose (FDG)
PET study of depressed males treated with fluoxetine or placebo [133]. PET scans
obtained at pretreatment baseline, and again after 1 and 6 weeks of treatment,
revealed some regional changes that were common to both medication and placebo
responders (i.e., metabolic increases in prefrontal, parietal, and posterior cingulate
regions, and decreases in subgenual cingulate), and other changes that were
uniquely seen in fluoxetine responders (metabolic changes in subcortical and limbic
regions including increases in pons and decreases in striatum, hippocampus, and
anterior insula). Considering these data, it is possible that qEEG measures are espe-
cially well suited to capturing functional changes that are unique to placebo
response.
In another report, pretreatment baseline features of the EEG, as well as other
pretreatment neurophysiological and clinical characteristics, were examined for
their ability to predict who will respond to placebo [134]. At baseline, those subjects
who would later be classified as placebo responders exhibited lower theta-band
frontocentral qEEG cordance as compared to all other subjects ( p 0.006). An

exploratory multiple-variable model including this frontocentral qEEG marker, in
addition to measures of cognitive processing time and insomnia, accurately identi-
fied 97.6% of eventual placebo responders. The ability to prospectively identify pla-

<