495
ISSN 1758-2008
10.2217/NPY.11.45 © 2011 Future Medicine Ltd
Neuropsychiatry
(2011) 1(5), 495–514
Summary
The human brainweighsapproximately 3 lbsandconsumes 40–60%
ofbloodglucose. This disproportionateamountofenergyisusedtocreateelectricity in
approximately 100 billioninterconnectedneurons. Quantitative EEG is a real-time movieof
theelectricalactivityofthepreconsciousandconsciousmind at frequenciesofapproximately
1–300 Hz. Numerousstudieshavecross-validatedelectricalneuroimagingbystructural MRI,
functional MRI anddiffusionspectralimagingandtherebydemonstratedhow quantitative
EEG canaid in linking a patient’ssymptomsandcomplaintstofunctionalspecialization in the
brain. Electricalneuroimagingprovides an inexpensivemillisecondmeasureoffunctional
modules, includingtheanimationofstructuresthroughphaseshiftandphase lock. Today,
neuropsychiatristsusethesemethodsto link a patient’ssymptomsandcomplaintsto
functionalspecialization in thebrainandusethisinformationtoimplementtreatment via
brain–computer Interfaces andneurofeedbacktechnology.
†
Authorforcorrespondence: NeuroImaging Laboratory, Applied Neuroscience Research Institute, St Petersburg, FL 33722, USA;
Tel.: +1 727 244 0240;
Review
Robert W Thatcher
†
Neuropsychiatryand quantitative
EEG in the 21st Century
Practice points
Useconventionalclinicalevaluationtoderive a diagnosisandidentifypatientsymptoms.
Measureeyes open andeyesclosedartifact-free quantitative EEG.
Calculate auto- andcross-spectratoidentifyscalplocationsandnetworkdeviationsfrom normal.
Use EEG tomographyto link thepatient’ssymptomsandcomplaintstofunctionalsystems in thebrain.
Identifyand separate the ‘weak’ systemsfromcompensatorysystems.
Use Z-score biofeedbacktotargetthederegulatedbrainsubsystemstoreinforce optimal andhomeostatic
statesoffunctionwhiletheclinicianmonitorsthepatient’ssymptomreduction.
Use quantitative EEG toevaluatepre- versus post-treatment and follow-upevaluationstodetermine
treatmentefficacy (e.g., medications, repetitive transcranialmagneticstimulation, electroconvulsivetherapy,
brain–computerinterfacesandbiofeedback, amongothers).
Neuropsychiatry
(2011) 1(5)
futuresciencegroup
496
r
eview
Thatcher
EEG isthemeasurementofthebrain-gener
-
atedelectrical potential betweenlocations on
thescalpand/orwithrespectto a reference.
Quantitative EEG (qEEG) invovlestheuse
ofcomputerstopreciselyquantifyelectrical
potentialsofapproximately 1–300
Hz, repre
-
sentingsubsecondmeasuresofsummatedlocal
fieldpotentialsgenerated in groupsofcortical
pyramidal neurons
[1]
. In thelast 40
years, over
90,000 qEEGstudieshavebeenlisted in the
National Library ofMedicine’sdatabase
[201]
.
Toreviewthisvastliterature, itisbesttousethe
searchterms ‘EEG and x’ where ‘x’ is a topic such
asschizophrenia, dyslexia, attentiondeficit, reli
-
ability, validity, obsessive–compulsivedis
orders,
evidenced-basedmedicine, anxietyorphobia,
amongothers. A readingofthestudiesand
abstractsshowsthatthevastmajorityofthese
studiesareqEEGstudiesinvolvingcomputer
analyses (e.g., spectralanalyses, ratiosof power,
coherenceorphase, amongothers). The search
term ‘EEG’ and not ‘qEEG’ isnecessarybecause
the National Library ofMedicinesearchesarti
-
cletitles/abstracts, andtheserarelyifeveruse
theterm ‘qEEG’ in the title (e.g., thisauthor
haspublishedsixbooksandover 200 total
publicationsandneverusedtheterm ‘qEEGor
QEEG’ in the title orabstract). This iswhy a
small ‘q’ isused in thispapertoemphasizethat
thesummationofelectricalpotentialsgenerated
by pyramidal neuronsynapsesarethesourcesof
the EEG andthe ‘q’ designatesquantification
asopposedto ‘eye-ball’ orvisualexamination
ofthe EEG tracesorsquiggleswithoutquanti
-
ficationasused in clinicalroutine. This article
iswrittenwith a specialemphasis on theuseof
qEEG after visualexaminationbypsychiatrists,
neuropsychiatrists, clinicalpsychologists, psy
-
chologists, neuro
psychologistsandneuroscien-
tistswhoaretheprimaryusersandpublishersof
psychiatric-relatedarticlesusingqEEG.
Historically, visuallyrecognized EEG pat
-
ternsandotherelectrophysiologicalmeasures
(evokedpotentialsand event-related potential)
wereusedtodiscernetiologicalaspectsofbrain
dysfunctionrelatedtopsychiatricdisorders
withreasonablesuccess, but not at thelevelthat
qEEGcanbeusedas a standalonediagnostic
methodforpsychiatricdisorders
[2]
. Instead,
qEEG was usedas an indicatoroforganicityor a
physiologicaletiologyofunknownoriginsimilar
tohow a clinicalbloodtestisusedaswellas an
objectiveevaluationoftreatmentefficacy upon
follow-up. In the 1960s and 1970s, priortothe
adventof MRI or PET scansor modern knowl
-
edgeofbrainfunction, it was speculatedthatthe
developmentof large qEEGdatabasesofpatients
with different clinicaldisorders will result in the
developmentofqEEGdiagnosticmeasuresthat
provideindicationsofpsychiatricdisorders
[3]
.
However, it was quicklyshownthatonly a sta
-
tisticalapproachisfeasible due tothenumberof
measuresandthefactthatthe EEG changeswith
age. As a consequence, ageregressionandstrati
-
fiedreference normative databasesweredevel
-
opedbyMatousekand Petersen in 1973
[4,5]
and
laterby John
[3,6–8]
, Duffy
[9]
, Thatcher
[10]
and
CongedoandLubar
[11]
, amongothers
[12–17]
.
The Stockholm, Sweden, normsofMatousek
and Petersen wereindependentlyreplicatedby
John andcollaborators in New York, USA
[3,6]
.
Subsequent replicationsof different qEEGnor
-
mativedatabasesdemonstratedthestatistical
stabilityandvalueofusingreference normative
databasestoaid in identifyingdeviant EEG fea
-
turesand in linkingthelocationof deviant fea
-
turestosymptomsandcomplaints
[2–8,12,16,18]
.
The referencedatabaseprovides a statistical
matchtoreliable quantitative featuresavailable
in the 1970s and 1980s. However, thespectral
methods in the 1970s relied upon the Fourier
transformthatdid not havesufficient temporal
resolutiontomeasurehigh-speeddynamics such
as rapid shifts in phasedifferencesandphase
lock. This problem was solved in thelate 1980s
withtheapplicationofjoint time–frequency
analysis (JTFA), where a time seriesof real-
time measuresofphasedifferencesisproduced.
JTFA providedprecisemeasuresofphaseshift
and lock durationsacrossthe human lifespan
for all combinationsoftheten- or 20-electrode
systemsand normative JTFA databasesthatwere
soondeveloped
[12,19]
.
Effortsare still beingundertaken in a fewlabo
-
ratoriestorecordandclassifyqEEGfromthou
-
sandsofpatientswiththe belief that a standalone
diagnosiscanbedevelopedfor different psychiat
-
ricdisorders. However, asexplainedby John
[2,3]
and Duffy
[9]
, itisunlikelythatqEEGcanserve
as a standalonediagnosticmeasureno matter how
large thedatabases. Forexample, meta-analyses
ofevidenced-basedmedicinecriteriaonlyshow
moderate to strong effectsizesforparticular EEG
features in schizophrenia
[4]
and obsessive–com
-
pulsivedisorder, post-traumatic stress disorder,
panicdisorder, generalizedanxietydisorderand
phobias
[2,20–22]
. This scientificliteratureshows
thatthereare a widevarietyof different changes