II

,pic

'yand

pp.

OR.

al-

Imsmal

'.;AND

zation

Meet-

1995h.

SELER.

1.;lon-

Re-

nonce

.;AND

t03lp'edi-

. Re-Is in a

.fER.

meas-

,using~S

'OKLO

MeasuringElectrical Activity

of the Brain

ERP Mapping inAlcohol Research

DAVID B. CHORLIAN, M.S., BERNICE PORJESZ, PH.D.,

AND HOWARD L. COHEN, PH.D.

The recording of brain electrical activity from scalp electrodes provides a noninvasive, sensitivemeasure of brain function. Event-related potentials (ERP's) are brain waves that are recordedwhile the subject is exposed to a specific sensory stimulus. Depending on experimentalconditions, ERP's are useful in studying many brain functions, such as sensory and informationprocessing (e.g., memory). The assessment of ERP's is useful in studying the effects of alcoholon brain function and in identifying people at risk for developing alcoholism. Computerizedmapping techniques produce graphs or color-coded images to summarize data about thegeneration of ERP's in time and space. KEYWORDS:brain function; brain wave; evoked potential;electrodiagnosis; diagnostic imaging; computer technology

Rcording brain electrical activity usingscalp electrodes provides noninva-sive, sensitive measures of brainfunction (e.g., cognition). These electricalrecordings consist of two phenomena: thecontinuous electroencephalogram (EEG)and timepoint-specific event-related brainpotentials (ERP's). Unlike positron emis-sion tomography (PET) or magnetic reso-nance imaging (MRI) techniques, brainmapping does not construct an image of ahidden anatomical structure; rather, it illus-

trates spatiotemporal brain activity (i.e.,brain activity as it occurs in both space andtime). This article presents a variety of brain

VOL. 19, No.4, 1995

mapping techniques, focusing on the ERPcomponent P300.

--

DAVIDB. CHORL/AN,M.S., is a seniorscientific programmer in the Neurody-namics Laboratory, State University ofNew York (SUNY)Health Science Centerat Brooklyn, Brooklyn, New York.

RECORDING BRAINELECTRICAL ACTIVITY

--

The EEG is not linked in time to any spe-cific event but instead reflects the activation

level of various brain regions. For example,a person in a relaxed state will manifest agreat deal of simultaneous brain wave ac-tivity occurring at a wave frequency between8 and 12 cycles per second. Conversely,ERP's represent brain electrical activity in

BERNICEPORJESZ,PH.D., is an assistantprofessor in the Department of Psychiatry,SUNY Health Science Center at Brooklyn,Brooklyn, New York.

HOWARDL. COHEN,PH.D., is a research

scientist in the Neurodynamics Laboratory,SUNY Health Science Center at Brooklyn,Brooklyn, New York.

315J

response to specific sensory or cognitiveevents occurring at a specific time. In ad-dition, ERP's consist of characteristic,highly reproducible wave forms that canbe measured to within a fraction of a sec-ond, providing an immediate record of thebrain activity associated with informationprocessing. Because ERP signals are smalland are embedded in the ongoing EEG,statistical techniques are used to extractthem from the background EEG.

Depending on experimental conditions,ERP's may represent overlapping activityof many brain circuits. Therefore, they are

F7

:;-~OJ"C

.~a.E~OJ

~

[o

I400

I600

I200

useful in studying complex brain functions,such as sensory and information processing(e.g., memory). ERP's offer a unique ap-proach for assessing brain function becausethey allow scientists to observe simultane-ously electrical activity (i.e., electrophysi-ology) and cognition.

How Are ERP's Obtained?

To record ERP's, subjects wear a cap em-bedded with from 20 to 128 noninvasivescalp electrodes. ERP's can be measuredin response to stimuli from any sensory

~PZ -0. ~~~---~:

FZ

CZ

PZ

1800

ro

I I I200 400 600

Time (ms)

modality (e.g., sight or sound) and alsomay be recorded while a behavioral taskis being conducted. For example, ERP'scan be obtained for visual stimuli using amethod in which shapes or letters arepresented on a computer screen while thesubject performs a simple cognitive task,such as recognizing a particular shape.The subject may be asked to make a be-havioral response (e.g., press a buttonwhenever a particular shape appears) orkeep count of how many specific stimulihave been presented. Scalp signals reflect-ing brain electrical activity are recorded

F8

Control subjects

Alcoholic subjects

-

simultansubject {:

What D

The peal<(i.e., neg;also kno'ponents,These cotheir poswell as tlrence of Imeasurelnegativefollowinthe Nl 0(with a Iresponseof the stithe laterpsychol(

What I!

Much at!componctive com500 ms ;of P300a rear (i.

1 r800' 0

I200

I600

1800

I400

Figure 1 Electroencephalographic (Le., brain wave) tracings of the waveform P300 obtained from nonalcoholic (Le., control) andalcoholic subjects in response to a visual stimulus. Each tracing represents a different location on the scalp. The hori-zontal scales represent time in milliseconds (ms). The vertical scales represent wave amplitude measured in microvolts(!lV) of electrical potential. The data are statistically derived from 197 subjects and 61 scalp electrodes.

316 ALCOHOL HEALTH & RESEARCH WORLD

Figur,

VOL.19,

II

:;-2:

r:"EOJ(5a.(ij.guOJ

!lVjjJ.S;(/)OJC>CC1I

.

II

,0 simultaneously from all electrodes as thesubject processes the stimuli.lsk

)'S

What Do ERP's Indicate?

thelsk,

The peaks (i.e., positive waves) and troughs

(i.e., negative waves) of the ERP wave form,also known as positive and negative com-

ponents, are measured in microvolts (~V).These components are named according to

their positive (P) or negative (N) polarity aswell as their latency (i.e., the time of occur-rence of the peak wave after the stimulus,measured in milliseconds [ms]). Thus, a

negative ERP wave occurring 100 ms

following the stimulus would be calledthe NI or N100 wave. Early components

(with a latency of less than 100 ms) areresponses to the physical characteristics

of the stimulus (e.g., intensity), whereasthe later components are influenced bypsychological factors (e.g., cognition).

)e-Iormliflect -led

What Is a P300?

Much attention has focused on the P300

component of the ERP, a prominent posi-tive component peaking between 300 and500 ms after a stimulus. Scalp recordingsof P300 are strongest at the parietal area,a rear (i.e., posterior) brain region. The

ERP task most commonly used to elicitthe P300 is the so-called "oddball" task,in which subjects are asked to attend toand/orrespondto a rare stimuluspresent-ed in a series of other stimuli. For exam-ple, in an auditory paradigm, the subjectlistens to frequent "boops" and rare "beeps"(targets) in a random stream of tone burstspresented about 1.5 seconds apart. The sub-ject may be askedto press a button or counteach time a target occurs (for a review, seeRegan 1989). P300 results from auditoryparadigms, however, are not as consistentas those from visual paradigms. Conse-quently, the P300 data discussed here arefrom a study using a visual paradigm inwhich alcoholics, high-risk subjects, andcontrol subjects pressed a button in responseto the rare occurrence of the target letter"X" embedded in a sequence of other vis-ual shapes.

Advantages of ERP Measurementin Alcoholism Research

ERP's provide sensitive measures of brainfunctions. Unlike other imaging techniques,ERP's reflect subtle, dynamic, real-time,millisecond-to-millisecond transactionsthat are elicited while the brain is chal-lenged and are therefore highly sensitive

Measuring the Brain's Electrical Activity

to specific brain processes. Most sensoryactivity occurs within 100 ms after a stim-ulus is presented, and most cognitive ac-tivity takes place within 500 ms. Althoughimages obtained using PET also offer afunctional measure of brain processes, eIec-trophysiological measures link images ofbrain activity to specific timepoints (Le.,provide a temporal resolution) with a pre-cision far greater than that of PET's. Otherimaging methods, such as MRI and com-puted tomography (CT) scans,portray grossbrain structure but do not reflect ongoingelectrical activity. Thus, they cannot pro-vide direct measures of brain function. ERPabnormalities, however, can be observedeven when MRI and CT images reveal noanatomical changes. In addition,many EEGand ERP characteristicsare extremely sen-sitive to acute and chronic effects of alcoholon the brain and are responsive to intoxi-cation, tolerance, withdrawal, and the ef-fects of prolonged abstinence.

Electrophysiological features also maybe hereditary markers of risk for alcoholism.Evidence indicates that the characteristicsof both EEG's and ERP's are geneticallydetermined and that P300 attributes gen-erally differ between alcoholics and non-alcoholics (Porjesz and Begleiter 1995). Arecent twin study (O'Connor et aI. 1994)

B. Potential Map for Alcoholic Subjects

o

18.016.014.012.010.08.06.04.02.00.0

-2.0-4.0-6.0-8.0

-10.0

250 500 750

Time (ms)

1,000

Figure 2 Spatiotemporal (Le., space-time) potential maps displaying electroencephalographic (Le., brain wave) tracings of thewave form P300 obtained from (A) nonalcoholic (Le., control) and (B) alcoholic subjects in response to a visual stimulus.This same data are shown in figure 1. The horizontal scales represent time in milliseconds (ms). The color-coded verticalscales represent changes in electrical potential measured in microvolts (IJV) in a spatial continuum along a scalp line fromfront to rear. Several such maps are needed to demonstrate the spatiotemporal changes at all electrodes. The maps showthat although large amplitude differences exist over time in P300 between the nonalcoholic and alcoholic subjects, spatialdistributions between the two groups are similar.

VORLD VOL. 19, No.4, 1995 317

A. Potential Map for Control Subjects

18.0

:;- 16.0 :;-2: 14.0 2:ro 12.0 roE FZ E FZ

10 I I OJ10.0 OJ

'0 '0Q.. 8.0 Q..

_"If j ',rj' f\ro 6.0 rou u. CZ 4.0 CZ0 I I OJ 2.0 OJiiJ 0.0

iiJ. .(/) -2.0 (/)OJ OJ

PZ -4.0C>c:

Figure 3 Topographic maps of (A) nonalcoholic (Le., control) and (B) alcoholic subjects representing electrical potentials(microvolts) recorded 430 milliseconds after exposure to a target stimulus (when P300 is at its maximum). Because themaps indicate fewer electrodes than points, mathematical models must be used to estimate values for the intermediatepoints. These two maps show little variation in potential.

reported that P300 amplitude is highly heri-table; similar findings on the heritability ofP300 are being obtained by researchers inthe Collaborative Study on the Genetics ofAlcoholism (COGA) through a large fam-ily study (Porjesz et al. in press).

METHODS OF ERP ANALYSISAND MAPPING

Amplitude and Latency Measures

Researchers have used various auditory andvisual paradigms to study the P300 compo-nent and have found the P300 amplitude tobe smaller in alcoholics than in control sub-

jects (figure I) (for a review, see Porjeszand Begleiter 1993). Researchers previouslythought that these low P300 voltages re-sulted from neurotoxic effects of alcohol onthe brain. Low P300's, however, do not re-cover with prolonged abstinence and canoccur in subjects at risk for alcoholism priorto alcohol exposure. Thus, low P300 ampli-tudes may characterize populations at risk fordeveloping alcoholism (Polich et al. 1994).

Topographic Representations

Although amplitude and latency measuresprovide important information about brainfunction, they do not address the spatial dis-tribution ofERP's across the scalp. Spatial

318

distributions provide information aboutbrain areas possibly involved in generatingERP activity.Displaying ERP's from multi-ple electrodes does not show values acrossthe entire scalp surface but only at the sitesof the electrodes. Mathematical techniquesenable researchers to assign values to thepoints that have not been directly meas-ured; maps can then be created by calcu-lating the values between electrode points.Figures 2 and 3 show two types of maps-spatiotemporal and topographic-basedon this approach (Fein et al. 1992).

Topographic maps are frozen at a sin-gle point in time, thereby losing the dy-namic aspect of the recorded data. Withthe advent of computer technology, thesespatial maps can be presented as a seriesof rapidly changing, computer-generatedanimations that describe how the topo-graphic distribution of the recorded activitychanges over time.

the radially oriented current density at apoint on the scalp. The CSD values de-note current sources and "sinks" (i.e.,locations where the current becomes un-detectable). Researchers create CSD mapsby taking simultaneous data from all theelectrodes and converting them into spatialdata. A CSD topographic map or spatio-temporal map (figures 4 and 5) may dis-play results in a more useful fashion thana topographic map based on ERP's (figures2 and 3).

In the P300 experiment mentionedearlier, for example, the peak of the P300component (occurring at 430 ms) from 61scalp electrodes is transformed to CSD. Inhealthy control subjects, these maps indi-cate both a large posterior focus of activityas well as front (i.e., anterior) foci (figure4). These brain-mapping techniques aid inpinpointing possible sites of brain dysfunc-tion in alcoholics and in subjects at riskfor developing alcoholism.

A comparison of CSD maps obtainedfrom alcoholics and nonalcoholics showsweaker activity in the alcoholics' brains.Figure 5 displays the distribution of P300activity across the alcoholics' scalps. Com-pared with control subjects, alcoholics canappear to have a weaker posterior focusof activity and no clear frontal focus. There-fore, control subjects and alcoholics ex-hibit not only differences in amplitudebut differences in spatial distributions,particularly in the brain's frontal regions.

MAPPING OF CURRENTSOURCE DENSITY

Because voltages measured at a point onthe scalp result from both local (i.e., di-rectly under the electrode) and remotesources by mathematically transformingthe ERP data, researchers can emphasizelocal sources. The data transformationproduces a quantity known as the currentsource density (CSD), which represents

ALCOHOL HEALTH & RESEARCH WORLD

heite

vat a; de-,e.,~sun-

D mapsill the

) spatialpatio-y dis-In than

(figures

1ede P300'rom61:':SD.In,s indi-

activity(figure:said inIysfunc-t risk

,tainedshows

brains.)f P300s.Com-,lics canfocus. There-cs ex-tudeIons,egions.

WORLD

Measuring the Brain's Electrical Activity

A. Current Density Map for Control Subjects

FZ(f)c:.21iiu.3 CZa.

performs a task. The results are interpret-ed as reflecting the activity of activenetworks in the cerebral cortex and can be

applied to various ERP paradigms.

Structural Findings

To determine the relationship betweenbrain structure and function, researchershave investigated the neuroanatomicalorigins of P300 using various techniques.Evidence from recordings obtained fromelectrodes implanted in the human brainimplicates both the medial temporal lobe'as well as source(s) within the frontal lobeas contributing to P300 generation. Thesefindings, coupled with the rather small ef-fect that temporal lobe removal has on scalpP300 during auditory oddball tasks, sug-gest that multiple brain sites contribute tothe P300 (for a review, see Regan 1989).

Researchers have used noninvasiveimaging techniques to assess brain struc-ture. In a recent study using P300 ampli-tude and MRI, Ford and colleagues (1994)reported that reduced P300's recorded dur-ing both automatic and effortful attentiontasks (i.e., tasks requiring subjects to focusintently) correlated with frontal and pari-etal gray-matter volumes. These findingsare consistent with the CSD maps of P300described earlier in which both parietal andfrontal foci of activity were found. The re-duced P300 amplitudes in alcoholics andtheir CSD maps may be manifestations offrontal lobe damage (see Oscar-Bermanand Hutner 1993).

'For a definition of this and other technical terms

used in this anicle. see central glossary, pp. 293-295.

CONCLUSIONS

Electrophysiological recordings of ERP'sreflect the activation of neural circuits in-volved in the mediation of sensory and cog-nitive processes. The assessment ofERP'shas proven extremely useful in studyingthe effects of alcohol(ism) on brain func-tion and in identifying people at risk fordeveloping alcoholism. In contrast, con-ventional imaging techniques have helpedidentify brain structural changes resultingfrom alcoholism.Electrophysiologicalbrainmapping is the representation of brain func-tioning in its spatiotemporal dimensions.These electrophysiological techniques pro-vide exquisite temporal resolution of brainprocesses and have now been enhanced toprovide spatial information about possiblebrain generators of this activity. Thesenovel techniques not only potentially pro-vide a window into the brain's function butalso may contribute simple clinical toolsfor identifying both the adverse effects ofalcohol consumption on the brain and peo-ple who may be at risk for alcoholism.

ACKNOWLEDGMENT

The authors would like to thank DouglasHarsh for developing computer graphicssoftware to create brain images.

REFERENCES

FEIN, G.; RAZ, J.; ANDTuRETSKY, B. Brain electrical

activity: The promise of new technologies. In: Zakhari,

S., and Witt, E.. eds. Imaging in Alcohol Research.National Institute on Alcohol Abuse and Alcoholism

Research Monograph No. 21. DHHS Publication No.(ADM)92-1890. Bethesda, MD: the Institute, 1992.

pp. 49-78.

FORD, J.M.; SULLIVAN,L.M.; WHITE, P.M.; LIM, K.O.;

ANDPFEFFERBAUM,A. The relationship between P300

amplitude and regional gray maIler volumes depends

upon the attentional system engaged. Electroencephalo-

graphy and Clinical Neurophysiology 90:214-228,1994.

GEVINS, A.S.; BRESSLER.S.L.; CUTILLO,B.A.; LILES,

J.; MILLER, J.C.; STERN,J.; ANDJEX, H.R. Effects of

prolonged mental work on functional brain topography.

Electroencephalography and Clinical Neurophysio.

logy 76:339-350, 1990.

O'CONNOR, S.J.; MORZORATI,S.; CHRISTIAN,J.C.; AND

LI, T.-K. Heritable fealures of the auditory oddball

event-related potential: Peaks, latencies, morphology

and topography. Electroencephalography and Clinical

Neurophysiology 92:115-125,1994.

OSCAR-BERMAN,M., ANDHUTNER,N. Frontal lobe

changes after chronic alcohol ingestion. In: Hunt, W.A.,

and Nixon, S.J., eds. Alcohol-Induced Brain Damage.National Institute on Alcohol Abuse and Alcoholism

Research Monograph No. 22. NIH Publication No. 93-

3549. Bethesda, MD: the Institute, 1993. pp. 121-167.

POLICH,J.; POLLOCK,V.E.; ANDBLOOM, F.E. Meta-

analysis of P300 amplitude from males at risk for al-

coholism. Psychological Bulletin 115:55-73, 1994.

PORJESZ,B., ANDBEGLEITER,H. Neurophysiological

factors associated with alcoholism. Hunt, W.A., and

Nixon, S.1., eds. Alcohol-Induced Brain Damage. Na-tional Institute on Alcohol Abuse and Alcoholism Re-

search Monograph No. 22. NIH Publication No. 93-3549.

Bethesda, MD: the Institute, 1993. pp. 89-120.

PORJESZ,B., ANDBEGLEITER,H. Event-related poten-

tials and cognitive function in alcoholism. Alcohol

Health & Research World 19(3):108-112,1995.

PORJESZ,B.; BEGLEITER,H.; LITKE, A.; BAUER, L.O.;

KUPERMAN,S.; O'CONNOR, S.1.; ANDROHRBAUGH,J.

Visual P3 as a potential phenotypic marker for alco-holism: Evidence from the COGA national project. In:Ogura, C.; Koga, Y.; and Shimokochi, M., eds. RecentAdvances in Event-Related Brain Potential Research.

Amsterdam: Elsevier Science, in press.

REGAN,D. Human Brain Electrophysiology: EvokedPotentials and Evoked Magnetic Fields in Science andMedicine. New York: Elsevier, 1989. pp. 236-245.

320 ALCOHOL HEALTH & RESEARCH WORLD

Sm4Ani

MARGl

JAMES,CYNTHAND Rc

KEYWOliresonanciency;nence; ,drug eft

~cottlaing the brto diagno,diseases, !tural and :organs, Ftinvolvinginvasive 1under invmonly usbrain or cresonanc<exposes tpresence cmeasuresbody's ti!ergy. Thepending c(e.g" theihelp of c(convertec

This anology capatients blmals thatmodels faare findin

MARGARE.

the Deparcology, ScSciences,Buffalo, f.

JAMESAufellow in tcal Colle!

CYNTHIA1

sistant proPh.D., is cAnatomy 1cine and 1

versity of.

VOL.19,N

II