Citation
The effects of stimulus characteristics on the relationship between the visual evoked response and intelligence

Material Information

Title:
The effects of stimulus characteristics on the relationship between the visual evoked response and intelligence
Creator:
Ondercin, Patricia Ann, 1948-
Publication Date:
Copyright Date:
1974
Language:
English
Physical Description:
viii, 41 leaves. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Amplitude ( jstor )
Correlations ( jstor )
Culture fair tests ( jstor )
Evoked potentials ( jstor )
Hemispheres ( jstor )
Intelligence tests ( jstor )
Nonsense ( jstor )
Philosophical psychology ( jstor )
Syllables ( jstor )
Wechsler scales ( jstor )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Electroencephalography ( lcsh )
Mental work ( lcsh )
Psychology thesis Ph. D ( lcsh )
Visual evoked response ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis -- University of Florida.
Bibliography:
Bibliography: leaves 36-40.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022778497 ( AlephBibNum )
14083438 ( OCLC )
ADA8911 ( NOTIS )

Downloads

This item has the following downloads:


Full Text









THE EFFECTS OF STIMULUS CHARACTERISTICS ON THE RELATIONSHIP
BETWEEN THE VISUAL EVOKED RESPONSE AND INTELLIGENCE














By
PATRICIA ANN ONDERCIN














A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLOIIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE
OF DOCTOR OF PHILOSOPHY







UNIVERSITY OF FLORIDA


1974























To my parents who made it all possible.

To David who made it all worthwhile.








ACKNOWLDEGEMENTS

I would like to express my gratitude to Dr. Nathan

Perry, Chairman of my supervisory committee,for his counsel

and encouragement during this study and throughout my

graduate work. I also wish to thank the members of my

committee, Dr. Jacquelin Goldman, Dr. Wiley Rasbury, Dr.

Calvin Adams, and Dr. Arnold Nevis,for their thoughtful

comments and guidance.

Special thanks are due Judy McCoy, Janet Falgout, and

Alan Pope for their continued interest and assistance.

Finally, I would like to thank Dr. David Spray for his

support, encouragement, and understanding.








TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS... . . . . . . . ... .iii

LIST OF TABLES . . . . . . . . .. v

LIST OF FIGURES. . . . . . . . ... vi

ABSTRACT . . . . . . . .. . ... vii

INTRODUCTION . . . . . . . . . 1

METHOD . . . . . . . .. . . . 8

RESULTS . . . . . . . ... . . .14

DISCUSSION . . . . . . . . .. . ... 29

LIST OF REFERENCES . . . . . . . ... 36

BIOGRAPHICAL SKETCH . . . . . . ... 41









LIST OF TABLES

Page

Table 1. Significant correlations between VER
measures and intelligence tests under
four stimulus conditions . . . . .. 17

Table 2. Significant t-scores between the means
of VER measures of bright and dull groups. 24









LIST OF FIGURES


Page

Figure 1. Examples of the stimuli used for the Check
and Word conditions. Stimuli used for the
Flash and Nonsense conditions were compar-
able . . . . . . . . ... . 11

Figure 2. Representative VER with component designations
exemplified by Gastaut and Regis (1965),
illustrating the components used for data
analysis in this study . . . . ... 15

Figure 3. Typical VERs recorded from a bright and dull
subject. . . . . . . . . ... 16

Figure 4. Schematic VER, based on mean latencies and
amplitudes of each component for all subjects,
under each of the four stimulus conditions. 22

Figure 5. Mean latencies and amplitudes for 11 bright
subjects compared with mean latencies and
amplitudes for 11 dull subjects under each
of the four stimulus conditions. . . .. .27









Abstract of Dissertation Presented to the Graduate Council of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE EFFECTS OF STIMULUS CHARACTERISTICS ON THE RELATIONSHIP
BETWEEN THE VISUAL EVOKED RESPONSE AND INTELLIGENCE

By

Patricia Ann Ondercin

August, 1974

Chairman: Nathan W. Perry, Jr.
Major Department: Psychology

A number of investigators have studied the relationship

between the visual evoked response (VER) and intelligence.

However, these studies have been characterized by the use of

relatively simple stimuli to elicit the response. Electro-

physiological research with animals has indicated that neural

processing varies with different types of stimuli. This

study was designed to assess the effects of different types of

stimuli on the VER, and on the relationship between the VER

and intelligence. Four stimulus conditions were used: Flash,

Checkerboard, Word, and Nonsense syllable. VERs were recorded

from positions C3 and C4 of the 10-20 International electrode

system of 37 boys, ages ten and eleven, whose short form WISC

scores ranged from 88-138. The Culture Fair Test was also

administered to provide a measure which was re) ;oly free

from sociocultural bias. It was thought that t measure ;'uld

yield higher correlations with VER characteristic';.






It was found that the highest and largest number of

significant correlations with intelligence occurred under the

Flash condition, and the fewest under the Word. The short form

WISC yielded the largest number of significant correlations

with VER and the Verbal score almost as many; the Culture Fair

Test proved to correlate poorly with the VER. The results

confirmed the findings of other investigators that the latency

of the VER is negatively correlated with intelligence. Am-

plitude was found to correlate negatively with intelligence

in the first three components (IV, Va, Vb), and positively in

the last two components (Vc, VI).

The results suggest that ongoing cognitive processes as

well as underlying neural organization are reflected in the

VER. The relationship of attention and arousal to the

correlations between the VER measures and intelligence was

discussed in terms of the different stimulus conditions and

intelligence tests. It was pointed out that while the

latency and amplitude of the VER were found to be significant-

ly correlated with intelligence, variations between individuals

were great, and these measures have little practical value at

this time in the assessment of intelligence.


viii









INTRODUCTION

Since the times when phrenologists attempted to assess

brain power by the bulges of the forehead, science has searched

for a relationship between brain and intelligence. It is gen-

erally accepted today that neural structure and function under-

lie cognitive capabilities. However, as recently as 1965, a

review of the literature concluded that no broad principles

of the neurophysiological correlates of intelligence had yet

been established (Ferguson, 1965).

Early work seeking electrophysiological measures of

cognitive ability was concerned with EEG frequency, particu-

larly 10-14 Hz or alpha waves. Ellingson (1966) reviewed

the area and concluded that the bulk of the evidence suggested

no relationship between EEG and intelligence in adults. In

children, results were contradictory and confounded by the

effect.which organic brain dysfunction has on both EEG ac-

tivity and intelligence. These conclusions were refuted by

Vogel and Broverman (1966). Recent studies have found posi-

tive correlations between slow waves and general ability

(Vogel, Broverman, and Klaiber, 1968) and positive results

using factor analytic techniques (Ishihara and Yoshii, 1972).

It would seem that no unequivocal statements can be made at

present about the relationship between the EEG and intelli-

gence.

In 1965, Chalke and Ertl reported striking correlations,









as high as -.70, between IQ scores and the latency of the

visual evoked response. The visual evoked response, or VER,

is the computer-averaged sum of individual electrical respon-

ses elicited by a repetitive photic stimulus. Spontaneous,

ongoing cortical activity averages 50 uV, while the response

evoked by a single stimulus is less than 10 uV. Therefore,

specialized computer techniques are required to extract the

small signal, or response, from the larger "noise" of the EEG.

Evoked responses are time-locked to a repetitive stimulus,

and when summed, they provide a record of the response to

that stimulus. The averaged EEG, which is not time-locked,

appears as a relatively straight line. Auditory and somato-

sensory, as well as visual stimuli can be used to generate an

evoked response which is recorded on the scalp with surface

electrodes. A diffuse light flash is the most commonly used

visual stimulus, although patterned light may also be used.

The early components of the evoked response are postulated to

represent perceptual processing; the later components, infor-

mation processing (John, Ruchkin, and Villegas, 1964; Uttal

and Cook, 1964; Ertl, 1968).

Since Chalke and Ertl's report, a number of investigators

have studied the area. It is somewhat difficult to assess the

VER-IQ literature, as the studies are not strictly comparable,

because of methodological differences. Subject populations

studied vary widely, electrode recording sites differ, diverse

measures of intellectual performance are used, and different

VER characteristics are studied.







A number of studies have been able to replicate Chalke

arl Ertl's negative correlations of latency and intelligence

in bright and dull adults (Plum, 1968; Shucard and Home,

1972), in children (Ertl, 1968; Ertl and Schaefer, 1969), and

with retardates (Bigum, Dustman, and Beck, 1970; Galbraith,

Gliddon, and Busk, 1970; Marcus, 1970). In general, correla-

tions were lower than those found by Chalke and Ertl, but sta-

tistically significant. Highly significant negative corrola-

tions were found between latency of the neonatal VER and men-

tal and motor development at eight months (Butler and Engel,

1969), but there was no latency correlation with language at

three years, or with IQ at four years (Engel and Fay, 1972),

or at seven years (Henderson and Engel, 1974). Neither was a

relationship between VER latency and intelligence found by

Rhodes and his co-workers (Rhodes, Dustman, and Beck, 1969).

Other characteristics of the VER have also been found to

correlate with intellectual ability. Greater amplitudes of

response components have been found in brighter children and

adults than in those less bright (Rhodes et al., 1969; Bigum

et al., 1970; Galbraith et al., 1970). However, Marcus (1970)

reports larger response amplitudes in Mongoloid than in normal

infants. Hemispheric asymmetry of VER amplitude has often

been noted, but results have been highly inconsistent. Several

studies report amplitude asymmetry to be characteristic of

normals, but not of dull or retarded children (Rhodes et al.,

1969; Bigum at al., 1970, Galbraith et al., 1970). However,

another study (Richlin, Weisinger, Weinstein, Gianni, and









Morganstern, 1971) found amplitudes greater in the right

hemisphere than in the left in normals, and the reverse, left

greater than right, in retarded children. Plum (1968) found

no relation between asymmetry and intelligence.

A few investigators have used auditory evoked response

(AERs) in studies of intelligence. These studies use clicks,

white noise, or pure tones as stimulus, and record from cen-

tral and temporal areas of the scalp. The amplitude of certain

AER components were found to be larger in Mongoloid infants

than in normal infants (Barnet and Lodge, 1967; Barnet, 1971).

Response decrement, i.e., the progressive decrease in response

amplitudes with repetitive stimulation, was seen in normal six

to twelve month old infants, but was not seen in Mongoloid

infants of the same age (Barnet and Lodge, 1967; Barnet,

Ohlrich, and Shanks, 1971). Latency differences were generally

not found between normal and retarded, or normal and Mongoloid

subjects (Barnet and Lodge, 1967; Barnet, 1971; Barnet et al.,

1971; Richlin et al., 1971). The exception is Shimizu (1969)

who reports a trend, although not a statistically significant

one, toward larger response latencies in retarded subjects.

He also found that AER latency and wave shape were reliable in

normal adults, but inconsistent in mentally retarded adults.

Despite the diversity of methodologies used in studies of

correlations between the VER and intelligence, there is general

agreement that low, but statistically significant correlations

exist. It seems reasonable to expect a more complex visual

stimulus to require more complex neural processing. Just as









a more difficult behavioral task is a more precise indicator

of intelligence than a simpler one, so a more complex neuro-

logical task should yield a more accurate picture of the

cognitive efficiency of the organism. Simple diffuse flashes

of light have been used as stimuli in all VER-IQ studies with

two exceptions which have used checkerboard patterns (Galbraith

et al., 1970; Marcus, 1970). There is strong evidence that

diffuse light is processed differently in the cortex than

patterned light (Hubel and Wiesel, 1962; Perry and Childers,

1968). It seems possible then, that a VER evoked by patterned

stimuli might be more reflective of differences in cortical

processing related to intellectual ability, than would a VER

evoked by diffuse stimulation. This rationale is readily

subject to experimentation and testing.

The major purpose of this study, then, is to determine

how the correlations between the VER and intellectual ability

are affected by stimuli of greater complexity. It is hypothe-

sized that VERs generated by patterned stimuli (checkerboard,

word, and nonsense syllable) will yield higher correlations

with IQ than the VER generated by diffuse stimuli.

It is, of course, impossible to quantify -visual complexity.

One can confidently say that a patterned stimulus is more com-

plex than a diffuse stimulus, but beyond that a rank ordering

of complexity is hypothetical. One could argue that a word

is the most complex because of its symbolic verbal content and

"meaningfulness." Yet, it could also be said that the non

sense syllable, as the most novel and unfamiliar stimulus,









might initiate more sustained cognitive activity and atten-

tion. A case could also be made for the checkerboard, as it

has the largest number of edges, has a verbal label, and may

stimulate a variety of associations. The various stimuli

cannot be ranked authoritatively for complexity then. How-

ever, it is postulated for the purposes of this discussion

that the checkerboard is the least complex of the patterned

stimuli since it is essentially non-verbal, is highly repeti-

tive in content, and has a somewhat limited association value.

The word is considered to be more complex, since it requires

cortical processing as a verbal and "meaningful" information.

The nonsense syllable will be considered most complex, as it

is a novel verbal stimulus which might have a large number of

associations attached to it because it lacks any well-defined

meaning.

Another variable which has not been considered in the

VER-IQ literature is the validity of the instrument used to

measure intelligence. It is well-documented that socio-

economic status (SES) is related to poor performance on IQ tests,

poor school achievement, lack of motivation, and slow develop-

ment of language skills (Terman and Merrill, 1927; McNemar,

1942; Jones, 1954; Bloom, 1964a,1964b;Kagan, 1970; Ginsburg,

1972). Conventional intelligence tests contain items which

are educationally and culturally biased to the advantage of

middle and upper SES groups, at the expense of the lower SES

groups (Cattell and Cattell, 1959). In order to control for

this bias, a test which is relatively free of contamil: ion





7


by the effects of school achievement will be administered in

addition to a conventional intelligence test. It is hypothe-

sized that VER-IQ correlations will be higher when intelligence

is measured by this test than when it is measured by the con-

ventional test.











METHOD


Subjects

Thirty-seven boys between the ages of ten and eleven,

who attended P. K. Yonge Laboratory School, were chosen

as subjects. This age group was chosen because the children

were old enough to sit quietly and attend to the stimulus,

yet young enough to avoid what unknown neurological effects

puberty might have. The visual acuity of each boy was

measured with a Snellen chart, and only those with an acuity

of 20/25 or better in each eye were accepted for the study.

In addition, Ss with a history of neurological dysfunction

or visual defect were excluded. The IQs of the Ss, as

measured by the short form of the Weschler Intelligence

Scale for Children (WISC), ranged from 88 (dull normal) to

138 (very superior), with a mean of 119.

Psychometric testing

Two intelligence tests were administered to each S. One,

a short form of the WISC, consisted of the following subtests:

Information, Arithmetic, Vocabulary, Picture Arrangement, and

Block Design. This is the pentad which correlates best with

the Full Scale WISC, r = .92, when corrected for subtest re-

liability (Silverstein, 1970). Verbal and Performance scores

for the WISC were calculated separately as well. In addition,

the Culture Fair Intelligence Test was administered in a group









to all Ss. This was chosen as a measure because it is con-

sidered to be relatively free from specific educational and

social biases (Cattell, 1940; Cattell, Feingold, & Sarason,

1941), as well as a valid indicator of general ability

(Cattell et al., 1941; Tilton, 1949; Geist, 1954; Marquant &

Bailey, 1955).

VER recording procedure

Silver-silver chloride electrodes (Beckman) were used to

record VERs monopolarly from the scalp, from positions C3 and

C4 of the International 10-20 electrode system, with the

reference electrode clipped to the ipsilateral ear. These

locations have been used by a number of investigators in both

monopolar and bipolar derivations (Ertl, 1968; Plum, 1968;

Rhodes et al., 1969; Weinberg, 1969; Richlin et al., 1971).

Microdot cable was used for leads from the electrodes

to the amplifiers, in order to minimize movement artifacts.

Electrical activity from the scalp was amplified by Grass

P-511 Amplifiers (Bandwidth 0.15-50 Hz) during stimulus pre-

sentation. The EEG signals were monitored visually on a

Tektronix Type 564 oscilloscope. After amplification, the

electrical activity was simultaneously routed to four channels

of a Computer of Average Transits (CAT 400B) for on-line

summation, and onto a seven channel FM magnetic tape recorder

(Sanborn 7000) for subsequent analysis.

The stimuli were prepared on 2" x 2" slides and were pro-

jected onto a diffuse light screen by two Viewlex V-27 pro-

jectors which were custom-mounted on a common base. Stimulus









duration was 500 msec. with an interstimulus interval of 1500

msec., determined by Gerbrandt electronic shutters controlled

by a Grass 5-8 Stimulator. The long stimulus duration was

used to avoid the "on" and "off" response mixtures obtained

with short pulse or strobe stimulation. The stimuli subtended

a visual angle of 6 on a side and were viewed under binocular

conditions. The S was seated in a padded chair and stimuli

appeared on the screen 6 ft. in front of the S. The luminance

level of the stimuli was 8 ft. cdl. on a dark surround for all

stimulus conditions, and was equated by a Variac variable

transformer. Luminance was measured by a photometer (UDT 40A

Opto-Meter), with the sensor at approximately the same distance

from the screen as the S.

Each VER recording was the summation of responses to 60

stimulus repetitions; two such recordings were obtained for

each of the four stimulus conditions. The four stimulus con-

ditions used were:

1. Diffuse light flash

2. Checkerboard pattern

3. Word ("FOR")

4. Nonsense syllable("RFO")

(see Figure 1).

The specific word was chosen because it is a high frequency

word (Kucera and Francis, 1967) which is classified at a pri-

mary reading level. The nonsense syllable is a recombination

of the same three letters. In order to minimize the effects

of tuation (response decrement occurring with repetitive














|I( ] .... ] 'A 1- 1* |






1 -1 i -A
LA L







stimulation), two stimuli were presented in random order

during a single trial. An incremental film strip reader was

used to program the random order of presentation. A binary

signal was recorded on a channel of the tape to enable relay

switching in the CAT to summate responses evoked by each stimu-

lus separately. The Flash and Check were in two trials, and

the Word and Nonsense in the other two. A trial consisted of

60 presentations of each stimulus, or a total of 120 presenta-

tions. Each trial took four minutes, and a brief (two minute)

rest period was allowed after each trial. In addition, two

control trials in which no light stimulation reached the eyes

were performed to test for the intrusion of artifacts. The

control trials consisted of 60 repetitions, and took two min-

utes. The order in which stimulus conditions and control trials

were presented was randomized.

Experimental procedure

Upon arrival in the laboratory, the S was tested for

visual acuity with a Snellen chart. The intelligence testing

had been completed previously at the school. The S's head

was measured for electrode placement, the sites cleaned with

alcohol, and the electrodes placed on positions C3 and C4.

The S was then seated in an electrically shielded, sound-

dampened and light-proof room. He was instructed to sit quiet-

ly without moving his head, and to watch the flashing lights.

The S was then fitted with earphones through which white noise

was transmitted in order to prevent the sound of the shutters

from evoking an auditory response. The lights inside tLu room








were turned off, and the trials began approximately 2 min.

later. The session lasted about 30 min.

Data analysis

Analog data tapes were played back following the experi-

mental procedure and data obtained in analog form by a Varian

F-50 Plotter. This yielded for each S two VERs for each con-

dition, which were then superimposed and averaged by visual

inspection. All subsequent analyses utilized this single

averaged VER. Latency was measured in milliseconds from the

beginning of the response to each peak. Amplitude was deter-

mined by measuring vertical distance in microvolts, with

reference to the preceding peak. Latencies and amplitudes of

the components were then correlated with the intellectual

measures using the Pearson product-moment procedure. Only

the later VER components (80-400 msec.) were correlated with

the intelligence measures, since most investigators have found

that correlations with intelligence occur within that range

(Rhodes et al., 1969; Galbraith et al., 1970).









RESULTS


Latency and Amplitude

The data appear to best fit the waveshape described by

Gastaut and Regis (1965), and the components were labelled

IV, Va, Vb, Vc, and VI (see Figure 2 ). Only these five waves

were analyzed for the purposes of this study, since these have

been found to be most related to measures of intelligence.

The components IV, Va, and VI were quite stable across Ss,

and Vb and Vc less so. A sample of the VER data can be seen

in Figure 3 .

There were a total of 66 significant correlations out

of a possible 320. Correlations ranged from +.55 to -.72,

with a mean of -.17. With a sample size of 37 Ss, correlations

of .33 and above are significant at the .05 level; however, due

to the absence of particular components in the VERs of some Ss,

the actual sample size for statistical purposes numbered as

low as 20, requiring a correlation of .42 for significance at

the .05 level (see Table 1 ).

Some of the correlations are strikingly high, up to -.72,

which are as high as those achieved by Chalke and Ertl (1965)

and by Galbraith and his co-workers (1970). They are con-

siderably higher than those achieved by several other studies

(Plum, 1968; Shucard and Horn, 1972, 1973). These high corre-

lations appear to be densely clustered around the three central

components of the response, Va, Vb, and Vc, especially under
14





















VI
IV






Vb
Vc I0 ./V

Va 100 msec



Figure 2. Representative VER with component disignations
exemplified by Gastaut and Regis (1965),
illustrating the components used for data
analysis in this study.













Left


/ -'


5 p.V

100 msec


Right


---Bright
-Dull


Figure. Typical VERs recorded from a bright and dull
subject.


J





TABLE 1

Significant correlations between VER measures and
intelligence tests under four stimulus conditions


WISC
VERB
PERF
CFT


Wechsler Intelligence Scale for Children
WISC Verbal Score
WISC Performance Score
Culture Fair Test


FLASH L I R L R L b R L c R L VI R
WISC -.49* -.57** -.59**-.72** -.72** -.57** -.64** -.33*
VERB -.43* -.54** -.54**-.71** -.59** -.56** -.56** -.35*
LATENCY PERF -.34* -.38* -.55**
CFT
WISC -.61** .43*
VERB -.61**
AMPLITUDE PERF .55* .37*
CFT -.51**
CHECK


-.37* -.62** -.37*
-.69** -.34*


-.53** -.62**
-.71** -.63**


WISC -.66** -.50** .40*
VERB -.42* -.67** -.40*
PERF -.38* .40*
CFT

WISC -.36* -.54** -.42*
VERB -.45*
PERF -.39*
CFT -.43* -.39*
WISC -.37*
VERB -.43*
PERF
CFT


WISC
VERB
PERF
CFT


WORD


LATENCY



AMPLITUDE



LATENCY




AMPLITUDE














TABLE 1 continued
TI7 V


L R R L R L L v R
NONSENSE
WISC -.35* -.58** -.37*
VERB -.43*
LATENCY PERF -.47** -.57**
CFT -.36* -.39*
WISC -.40* -.40* -.42*
VERB -.41* -.46**
AMPLITUDE PERF
CFT

Statistically significant at the .05 level
** statistically significant at the .01 level

L Left hemisphere
R Right hemisphere


t7b


T7T


i-









the Flash condition, but also with the Checkerboard. Corre-

lations tend to be lower and more scattered with the Word and

Nonsense conditions. The largest number of correlations and

those of the greatest magnitude were associated with response

latency, but some high correlations were also seen with response

amplitude. These high correlations occur primarily with the

short form WISC and the Verbal score.

With one exception, all the correlations between response

latency and the intellectual measures were in the negative di-

rection. Short latency was associated with higher intellectual

abilities and long latency with less ability. The picture is

more complex with regard to amplitude. With the early wave

components (IV, Va, Vb) all correlations were negative, indicat-

ing that larger amplitudes were related to less ability, and

smaller amplitudes with more ability. However, with the two

late components (Vc, VI) the correlations were in a positive

direction; larger amplitudes were associated with higher in-

telligence and smaller amplitudes with lower intelligence. The

number of significant correlations was different for latency

and amplitude measures. There were 43 correlations above the

level of significance with the latency measure, as compared

with 23 for the amplitude measure.

Measures of Intelligence

The number of significant correlations also varies with

the measure of intelligence used. The VER measures were

correlated with the short form WISC, its Verbal and Performance

Scores, and with the Culture Fair Test. It was the short form








WISC which correlated best with the VER, accounting for 29

of the 66 significant correlations. The Verbal score also

correlates well, with 23 correlations. Contrary to the hy-

pothesis that the Culture Fair Test would correlate especially

well with the VER, it accounted for only five of the signifi-

cant correlations. The Performance score was little better

with nine.

Stimulus Conditions

There were differences in the number of significant

correlations occurring under the four different stimulus con-

ditions. It was hypothesized that a more complex stimulus,

such as a verbal one, would correlate better with the VER.

Almost the opposite was seen to be true. It was the Flash

condition which accounted for both the highest and the most

correlations, 25, while the Word condition accounted for the

least, nine. The Check and Nonsense conditions fell about

midway between these with 17 and 15 correlations respectively.

Another way to look at differences between stimulus con-

ditions is to average the latencies and amplitudes of all Ss

to achieve one composite waveform for each hemisphere under

each condition (see Figure 4). Although in general the wave-

shapes appear highly similar, several differences are notable.

The composite VER elicited by the flash appears to be more like

a "W" in shape than do the other waves; the negative peaks Va

and Vc are more nearly on an even plane, while under all other

conditions, Vc is considerably higher than Va. Latency dif-

ferences between conditions are also apparent in the later

components. The latencies of peaks Vb, Vc, and VI show a clear








CHECK


200o-


0
~----L

LLJ
c -200
D


S 200'


(5 o
0 0
CT"


-200--


I ., I


I I


100 150 200 250


100 150 200 250 300


LATENCY (msec)


S. I I I I _


FLASH






NONSENSE


2001


-200


100 150 200 250 100 150 200 250 300
LATENCY (msec)
Figure 4. Schematic VER, based on mean latencies and amplitudes of each component
for all subjects, under each of the four stimulus conditions.


-200q

200]q


WORD








trend toward increasing across conditions. That is, the com-

ponent latencies are shortest under the Flash condition,

longer under the Check, increase little or not at all with

the Word, and are longest under the Nonsense condition.

Comparison of Bright and Dull Ss

It was thought that additional data could be gathered

by a comparison of the brightest and dullest of the Ss within

the sample. Twenty-two Ss were chosen by selecting the eleven

boys with the highest short form WISC IQs, and the eleven with

the lowest. The IQs of the bright group ranged from 125 to

138; those of the dull group from 88 to 106. These groups

were significantly different for intelligence at the .01

level, on all four measures of intelligence. Means and stan-

dard deviations were calculated for each group, and t-tests

performed between groups (see Table 2). As indicated by the

previously mentioned correlations between VER measures and

intelligence, the bright group has shorter latencies than the

dull group under all stimulus conditions. The bright children

tend to have smaller amplitudes in the earlier wave components

than did the duller children, but larger amplitudes in the

later components.

The difference between bright and dull groups is more

clearly illustrated in Figure 5. Variations among stimulus

conditions are apparent, as are variations between left and

right hemispheres. For both Flash and Check conditions, the

right hemisphere responses appear more flattened, with general-

ly smaller amplitudes, than those of the left hemisphere. The






TABLE 2


Significant t-scores between the means of
VER measures of bright and dull groups


L Va R L Vb R


L Vc R


Laten- YD 156 158
cy
t 2.76** 2.88**

Ampli- XB
tude
%D


3.50** 3.86** 2.82** 4.12**


379

2.91**


30 41

3.06** 2.24*


CHECK


Laten- -
cy D
t


2.19* 3.31** 2.11* 2.40* 2.35*


99 170


3.55** 1.99*


L IV R


FLASH


L IV R


2.18*


Ampli-
tude


1.91* 2.44*








TABLE 2 continued


IV R L Va
L R L


Vb Vc
R L R L R


2.55* 3.72** 3.44** 2.93** 2.80**

355
546

2.09*


2.01* 4.13*


2.49*


t 1.91*


B Mean of bright group
XD Mean of dull group
* Statistically significant at
** Statistically significant at
Note: Mean amplitude score x 10


2.25* 2.41* 3.06**


L Left hemisphere
R Right hemisphere
the .05 level
the .01 level


WORD


Laten-
cy



Ampli-
tude


141 184
163 234


L IV
L R


NONSENSE


Laten-
cy


Ampli- XB
tude -
XD


205
245

2.64*


390 177

624 469


_ _






400


200


* 0
0
LIJ
z-200
t-


-100
100


-100


-300


CHECK




Left


100 150 200 250


100 150 200 250


LATENCY (msec)


FLASH
--- Bright
-- Dull

Left


/
/


300





400- WORD NONSENSE
-- --- Bright
200- ---Dull /

17- o / /\ 1 J/
200 -
/ I



,/ -200 \
S 0
/ // /\ / ,
u -200-- /

/
S 200


S-/ _/ /
V V

-200-

,.. I I I I /, I I I I I
100 150 200 250 300 100 150 200 250 300
LATENCY (msec)
Figure 5. Mean latencies and amplitudes for 11 bright subjects compared with mean
latencies and amplitudes for 11 dull subjects under each of the four
stimulus conditions.








waveshapes of the responses from the right hemisphere are al-

so quite similar in bright and dull groups. This is not true

in the left hemisphere where the waveshapes for bright and

dull groups appear markedly different, primarily due to the

amplitude of the Vb component. With the Word and Nonsense

conditions, it is also in the left hemisphere that waveshape

differences between bright and dull groups are more apparent,

again, primarily due to the amplitude of Vb. These hemi-

spheric differences in waveshape cannot be accounted for by

hemispheric asymmetry in either group, since t-tests per-

formed between hemispheres for each group did not reach sta-

tistical significance. Rather they seem due to the differen-

tial amplitudes and to some extent, latencies, between groups.

However, the number of Ss in each group was small, and the

standard deviations, especially of the amplitude measures

large, so that hemispheric asymmetry cannot be completely dis-

counted as a contributing factor.












0
C\O













DISCUSSION

The major hypothesis of this study, that more complex

visual stimuli would correlate more highly with measures of

intelligence than simpler stimuli was not upheld. Although

there were clear differences between conditions, they were

in the opposite direction from that predicted: it was the light

flash which accounted for both the highest and the largest num-

ber of significant correlations with intelligence. Several

explanations might account for this. It may be that response

frequencies in the alpha range, 10-14 Hz, contribute heavily

to the relationship between the VER and intelligence, and the

correlations are best when this frequency is most in evidence,

i.e., with diffuse stimuli, as in the Flash condition. In

earlier studies, frequency analysis of the VERs elicited by

diffuse stimulation revealed a predominance of frequencies in

the alpha range (Ertl, 1971), while those frequencies were

rarely seen when visually complex stimuli were used (Perry,

Childers, and Falgout, 1972). Weinberg (1969) reports that

the highest correlations with intelligence are associated with

the frequencies of 12-14 Hz in VERs elicited by diffuse stimu-

lation.

Another explanation of these differences might be based

on the findings of several studies that VER differences re-

lated to intelligence tend to be obscured by increasing the Ss'









level of attention or arousal (Plum, 1968; Shucard and Horn,

1972). Response amplitudes increase during conditions of

high attention, while response latency decreases (Garcia-

Austt, 1963; Haider, Spong, and Lindsley, 1964; Gross,

Begleiter, Tobin, and Kissin, 1965). It will be remembered

that most studies have found both shorter latencies and larger

amplitudes were related to higher intelligence in children.

It seems likely that brighter children are generally in high-

er states of arousal and attention than duller children, but

the imposition of a simple task or arousal device stimulates

relatively more arousal in duller Ss, thus obscuring the

differences between them. Visually more complex stimuli, such

as the check, word, and nonsense syllable, may be more arousing

and attention-getting for the duller Ss than for the brighter,

and thus tend to obscure differences between groups. A flash

would have less arousing qualities, and therefore emphasize

the intrinsic differences in arousal level between Ss. If

component amplitudes, which are considered to reflect arousal,

are ranked for size across conditions, there is a suggestion

of a trend in this direction, although it is not of statistical

significance. For dull Ss, the Nonsense condition elicits the

highest amplitudes (suggesting higher arousal), and the Flash

condition the lowest amplitudes. With the bright Ss, amplitudes

are more nearly equal across conditions, indicating a more uni-

form level of arousal which seems less affected by extrinsic

characteristics of the stimuli. This would lend support to the

hypothesis that brighter children intrinsically maintain a







higher level of arousal, and that duller children are less

able to sustain attention when presented with simple stimuli,

but are relatively more aroused by complex or meaningful

stimuli.

It is interesting to note the differences in the com-

posite waveforms across stimulus conditions. The waveshape

evoked by the flash is distinct from those evoked by the

patterned stimuli, appearing more like a "W' This is consis-

tent with evidence that diffuse and patterned light are pro-

cessed differently in the cortex. The increasing latencies

of the composite VERs across conditions are also suggestive.

The most complex stimulus, the nonsense syllable, shows the

longest response latencies of the four stimulus conditions,

suggesting it requires a relatively longer processing time in

the cortex. The least complex stimulus, the flash,shows the

shortest latencies, and might indicate the relatively quicker

cortical processing of simple stimuli. The word and the check

show more nearly equal latencies, midway between those of the

flash and nonsense, and might indicate a similarity of cortical

processing. It is possible that the check is being given an

immediate verbal label by the S, and so is processed as a

word, as well as a configuration.

The differences between the Word and Nonsense conditions,

both in latencies and in the number of significant correla-

tions are not of statistical significance. They are intriguing,

however, because they are composed of identical letters, and

more similarity might be expected if cortical processing was









also identical. Although they vary in familiarity or novelty,

it is debatable how novel any stimulus can be after 60 repe-

titions, so it would seem that meaningfulness is the princi-

pal dimension along which they vary. The word is defined by

an assigned meaning, and in that sense is somewhat limited.

The nonsense syllable has no particular meaning assigned to it,

and is therefore more open to interpretation and varied associ-

ations. This less restricted quality may be more stimulating

to the brighter Ss than the duller, and emphasize differences

between them. There is evidence to suggest that meaningful-

ness of the stimuli is associated with enhancement of the VER

(Symmes and Eisengart,1971), which might indicate increased

arousal or attention. It is interesting to speculate on the

possibility of a curvilinear relationship between the correla-

tions with intelligence and the arousal value of the stimulus.

The flash, as the simplest stimulus, is not very arousing for

either group, and their intrinsically different levels of

attention or arousal are made apparent. Word and check provide

extrinsic arousal, which is relatively more arousing for the

duller Ss, obscuring intersubject differences. The nonsense

syllable also provides extrinsic stimulation, more so than the

word or check, because of its lack of specificity, and makes

differences between bright and dull Ss more general.

Another major hypothesis of this study was that the

Culture Fair Test would prove ; better instrument for assessing

intelligence in relationship the VER than would the WISC.

However, just the reverse was shown to be the case. This is









surprising since the Culture Fair Test emphasizes both speed

and skill in the analysis and interpretation of visual in-

formation, abilities which one would expect to be important

in the processing of visual stimuli in the VER. Rather, it

is the short form WISC which is heavily loaded for verbal

abilities, and the Verbal subtest of the WISC, which yields

the majority of the significant correlations. These tests

primarily measure verbal comprehension and skills related

to school achievement.

There are generally thought to be three factors involved

in intelligence: the ability to encode information, the

ability to retain information over time, and the ability to

retrieve information. Retrieval of stored inforamtion has two

aspects: the recall of stored data in their original form, and

the manipulation of relevant data to form new combinations.

The WISC would seem to rely heavily on the more passive recall

of learned information. The Culture Fair Test, in contrast,

presents unfamiliar stimuli and demands a more active process

of retrieval and recombination of relevant data in a new

situation. This is consistent with the previous hypothesis

that the VER correlates better with intelligence under less

arousing conditions than under more arousing ones. The type

of intelligence reflected in the VER then, would seem to be

more in a passive, receptive mode than a more active, manipula-

tive mode.

It is apparent from the results of this and other studies

that the latency and amplitude of the VER have a significant








relationship to intelligence, brighter Ss tended to have short-

er response latencies, duller Ss tended to have longer laten-

cies. It is thought that the shorter latencies reflect faster

and more efficient neural processing of incoming stimuli,

while longer latencies reflect slower and less efficient pro-

cessing. However, a variation of one of Spitz's (1963) postu-

lated of neural functioning in the mentally retarded might

also apply. Spitz states that a relatively longer time is

required to induce a temporary change in stimulated cortical

cells in retardates than is required in normals. This im-

plies that longer response latencies would be seen in the

VERs of retardates as compared to normals. In a comparison

of the VERs of normal and retarded Ss, Galbraith and his co-

workers (1970) have lent support to this hypothesis. It seems

likely that this postulate would also apply more generally to

the range of intellectual functioning in a normal population,

i.e., in brighter individuals, cortical cells are more rapidly

activated by stimulation than in dull individuals. The impli-

cation of shorter response latencies for brighter individuals

was clearly borne out by this stuC

The correlations between intelligence and response ampli-

tude are confusing, however, and agree only in part with other

work. The ea ier wave components are negatively correlated

with intelligence. Higher amplitude is associated with the

duller Ss, lower amplitudes with brighter Ss.

Dustman and Beck (1969) reported that between the ages of

5 and 13, the amplitude of the BER response shows a marked

decrease. The components which are reduced fall in the range
















of 70-225 msec.,which roughly corresponds to the latencies

of components IV, Va, Vb. It seems possible that the brighter

children are developmentally more mature than less bright

children, and this difference is reflected in smaller compo-

nent amplitudes characteristic of more advanced development

of the central nervous system.

In summary, it seems clear that correlations can be found

between the VER and measures of intelligence. These correla-

tions appear to reflect both ongoing cognitive processes and

underlying neural organization. However, it must be stressed

that while VER-IQ correlations were significant, the possibility

of assessing the intelligence of an individual is very limited

due to the large variability of the response. Callaway (1973)

makes the point that even perfect VER-IQ correlations would

give a measure that was no better and very likely more expensive

than a conventional test. Yet VER measures are less affected

by specific learning and school performance, and may well cast

new light on individual differences in cognitive functioning.

It might prove especially useful in the early detection of

subtle learning dysfunctions, such as dyslexia.





-ar"'
'C












REFERENCES


Barnet, A. B. Evoked potentials in handicapped children.
Developm. Med. Child Neurol., 1971 13, 313-320.

Barnet, A. B. and Lodge, H. Click evoked EEG responses in
normal and developmentally retarded infants. Nature,
1967, 214, 252-255.

Barnet, A. B., Ohlrich, E. S., and Shanks, B. L. EEG evoked
responses to repetitive auditory stimulation in normal
and Down's syndrome infants. Developm. Med. Child
Neurol., 1971, 13, 321-329.

Bigum, H. B., Dustman, R. E. and Beck, E. C. Visual and
somato-sensory evoked responses from mongoloid and normal
children. Electroenceph. Clin. Neurophysiol., 1970,
28, 576-585.

Butler, B. V. and Engel, R. Mental and motor scores at eight
months in relation to neonatal photic responses. Developm.
Med. Child Neurol., 1969, 11, 77-82.

Bloom, B. S. Stability and change in human characteristics.
New York: Wiley, 1964a.

Bloom, B. S. Compensatory education for cultural deprivation.
Chicago: University of Chicago, 1964b.

Callaway, E. Correlations between averaged evoked potentials
and measures of intelligence. An overview. Arch. Gen.
Psychiatry, 1973, 29, 553-558.

Cattell, R. B. A culture free intelligence test. I. J. Educ.
Psychol., 1940, 31, 161-180.

Cattell, R. B. and Cattell, A. K. S. Handbook for the culture
fair intelligence test. Scale 2. Champaign, Ill: Insti-
tute for Personality and Ability Testing, 1959.

Cattell, R. B., Feingold, S. N., and Sarason, S. B. A culture
fair intelligence test. II Evaluation of cultural in-
fluences on test performance. J. Educ. Psychol., 1941,
32, 81-100.

Chalke, F. C. R. and Ertl, J. Evoked potentials and intelli-
gence. Life Sciences, 1965, 4, 1319-1322.









Dustman, R. E. and Beck, E. C. The effects of maturation and
aging on the waveform of visually evoked potentials.
Electroenceph. Clin. Neurophysiol., 1969, 26, 2-11.

Ellingson, R. J. Relationship between EEG and test intelli-
gence: A commentary. Psychol. Bull., 1966, 65, 91-98.

Engel, R. and Fay, W. Visual evoked responses at birth, ver-
bal scores at three years, and IQ at four years.
Developm. Med. Child Neurol., 1972, 14, 283-289.

Ertl, J. P. Evoked potentials, neural efficiency and IQ.
Paper presented at the International Symposium for
Biocybernetics, Washington, D. C., 1968.

Ertl, J. P. Fourier analysis of evoked potentials and human
intelligence. Nature, 1971, 230, 525-526.

Ertl, J. P. and Schaefer, E. W. P. Brain response correlates
of psychometric intelligence. Nature, 1969, 223, 421-422.

Ferguson, G. A. Human abilities. Ann. Rev. Psychol., 1965,
16, 39-62.

Galbraith, G. C., Gliddon, J. B., and Busk, J. Visually evoked
responses in mentally retarded and nonretarded subjects.
Amer. J. Ment. Deficiency, 1970, 75, 341-348.

Garcia-Austt, E. Influence of the states of awareness upon
sensory evoked potentials. Electroenceph. Clin.
Neurophysiol., 1963, Supp. 24, 76-89.

Gastaut, H. and Regis, H. Visually-evoked potentials recorded
transcranially in man. In L. D. Proctor and W. R. Adey
(Eds.) The analysis of central nervous system and cardio-
vascular data using computer methods. (1964 SymposiumT
1965. Washington, D. C., NASA, 7-34.

Geist, H. Evaluation of culture free intelligence. Calif. J.
Educ. Res. Dig., 1954, 5, 209-214.

Ginsburg, H. The myth of the deprived child. Poor children's
intellect and education. Englewood Cliffs, N. J.:
Prentice-Hall,1972.

Gross, M. M., Begleiter, A., Tobin, M., and Kssin, B. Audi-
tory evoked response comparison during cc :.ting clicks
and reading. Electroenceph. Clin. Neuror iiol., 1965,
18, 451-454.

Haider, M., Spong, P., and Lindsley, D. B. Cc cal evoked
potentials during visual vigilance task i ormance.
Electroenceph. Clin. Neurophysiol., 1964, 710-715.









Henderson, N. B. and Engel, R. Neonatal visual evoked poten-
tials as predictors of psychoeducational tests at age
seven. Developm. Psychol., 1974, 10, 269-276.

Hubel, D. H. and Wiesel, T. N. Receptive fields, binocular
interaction and functional architecture in the cat's
visual cortex. J. Physiol., 1962, 160, 106-154.

Ishihara, T. and Yoshii, N. Multivariate analytic study of
the EEG and mental activity in juvenile delinquency.
Electroenceph. Clin. Neurophysiol., 1972, 33, 71-80.

John, E. R., Ruchkin, D. S., and Villegas, J. Experimental
background: signal analysis and behavioral correlates
of evoked potential configurations in cats. Ann. N. Y.
Acad. Sci., 1964, 112, 362-420.

Jones, H. E. The environment and mental development. In
L. Carmichael (Ed.) Manual of child psychology, 2nd
edition. New York: Wiley, 1954, 631-696.

Kagan, J. On class differences and early development. In
V. H. Denenberg (Ed.) Education of the infant and young
child. New York: Academic Press, 1970.

Kucera, H. and Francis, W. N. Computational analysis of
present-day American English. Providence: Brown
University Press, 1967.

Marcus, M. M. The evoked cortical response:a technique for
assessing development. Calif. Ment. Health Res. Dig.,
1970, 8, 59-72.

Marquant, D. I. and Bailey, L. L. An evaluation of the culture
free test of intelligence. J. Genet. Psychol., 1955, 86,
353-358.

McNemar, Q. The revision of the Stanford-Binet. Boston:
Houghton Mifflin, 1942.

Perry, N. W., Jr. and Childers, D. G. Cortical potentials
in normal and amblyopic binocular vision. In E.
Schmoeger (Ed.) Advances in Electrophysiology and
Pathology of the Visual System. Leipzig, Thieme, 1968,
151-161.

Perry, N. W. Jr., Childers, D. G., and Falgout, J. C. Chroma-
tic specificity of the visual evoked responses. Science,
1972, 177, 813-815.

Plum, A. Visual evoked responses: Their relationship to
intelligence. Doctoral dissertation, University of Florida,
1968.









Rhodes, L. E., Dustman, R. E., and Beck, E. C. The visual
evoked response: A comparison of bright and dull children.
Electroenceph. Clin. Neurophysiol., 1969, 27, 364-372.

Richlin, M., Weisinger, M., Weinstein, S., Gianni, M., and
Morganstern, M. Interhemispheric asymmetries of evoked
cortical responses in retarded and normal children.
Cortex, 1971, 7, 98-105.

Shimizu, H. AER in the severely retarded. Excerpa Medica
International Congress Series, 1969, 206, 530-534.

Shucard, D. W. and Horn, J. L. Evoked cortical potentials
and measurement of human ability. J. Comp. Physiol.
Psychol., 1972, 78, 59-68.

Shucard, D. W. and HOrn, J. L. Evoked potential amplitude
change related to intelligence and arousal.
Psychophysiology, 1973, 10, 445-452.

Silverstein, A. B. Reappraisal of the validity of WAIS, WISC
and WPPSI. J. Consult. and Clin. Psychiat., 1970, 34,
12-14.

Spitz, H. H. Field theory in mental deficiency. In N. R.
Ellis (Ed.) Handbook of mental deficiency. New York:
McGraw-Hill, 1963.

Symmes, D and Eisengart, M. A. Evoked response correlates
of meaningful visual stimuli in children. Psychophysiology,
1971, 8, 769-778.

Terman, L. M. and Merrill, M. A. Measuring intelligence: A
guide to the administration of the new revised Stanford-
Binet tests. Boston: Houghton Mifflin Co., 1937.

Tilton, J. R. A survey of the reliability, validity and use-
fulness of the Cattell Culture Fair Test. Persona., 1949,
1, 17-19.

Uttal, W. R. and Cook, L. Systematics of the evoked somato-
sensory cortical potential: a psychophysical-electro-
physiological comparison. Ann. N. Y. Acad. Sci., 1964,
112, 60-79.

Vogel, W. and Broverman, D. M. A reply to "Relationship be-
tween EEG and test intelligence: A commentary."
Psychol. Bull., 1966, 65, 99-109.

Vogel, W., Broverman, D. M., and Klaiber, E. L. EEG and mental
abilities. El -troenceph. Clin. Neurophysiol., 1968, 24,
166-175.





40



Weinberg, H. Correlation of frequency spectra of averaged
visual evoked potentials with verbal intelligence.
Nature, 1969, 224, 813-815.









BIOGRAPHICAL SKETCH

Patricia Ann Ondercin was born in Racine, Wisconsin

in 1948. She attended St. Catherine's High School there,

and received her diploma in 1966. Her undergraduate studies

were done at Marquette University, in Milwaukee, Wisconsin.

She was graduated cum laude in 1970 with a Bachelor of Arts

in Psychology and English. She began her graduate studies at

the University of Florida the following September, and was

awarded the degree of Master of Arts in December, 1971. In

July, 1973, she was appointed to an Internship in Clinical

Psychology at the New York Hospital-Cornell Medical Center,

in White Plains, New York.









I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.




Nathan W. Perry, Jr.,ChairyAn
Professor of Psychology



I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.




S(jcqulin R. Goldmant
Associate Professor of Psychology



I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.




Wiley C. rasbury
Assistant Professor of Psychology



I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.




Calvin K. Adams
Assistant Professor of Psychology







I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.




Arnold it. Nevis
Professor of Electrical Engineering



This dissertation was submitted to the Department of
Psychology in the College of Arts and Sciences and to the
Graduate Council, and was accepted as partial fulfillment
of the requirements for the degree of Doctor of Philosophy.

August, 1974



Dean, Graduate School




Full Text

PAGE 1

THE EFFECTS OF STIMULUS CHARACTERISTICS ON THE RELATIONSHIP BETVffiEN THE VISUAL EVOKED RESPONSE AND INTELLIGENCE By PATRICIA ANN ONDERCIN A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1974

PAGE 2

,^N;VERSITy OF FLORIDA 3 1262 08552 8494

PAGE 3

To my parents v7ho made it all possible. To David who made it all v/orthwhile.

PAGE 4

ACKNO WLDEGEMENT S I v;ould like to express my gratitude to Dr. Nathan Perry, Chairman of ray supervisory conanittee, for his counsel and encouragement during this study and throughout iny graduate work. I also wish to thank the members of my committee, Dr. Jacquelin Goldman, Dr. Wiley Rasbury , Dr. Calvin Adams, and Dr. Arnold Nevis, for their thoughtful comments and guidance. Special thanks are due Judy McCoy, Janet Falgout, and Alan Pope for their continued interest and assistance. Finally, I would like to thank Dr. David Spray for his support, encouragement, and understanding.

PAGE 5

TABLE OF CONTENTS Page iii V vi vii 1 ACKN0V7LEDGEMENTS LIST OF TABLES LIST OF FIGURES ABSTRACT INTRODUCTION METHOD g RESULTS ^4 DISCUSSION 29 LIST OF REFERENCES 3g BIOGRAPHICAL SKETCH ^j^

PAGE 6

LIST OF TABLES Table 1 Significant correlations between VER measures and intelligence tests under four stimulus conditions , Page 17 Table 2. Significant t-scores between the means of VER measures of bright and dull groups 24

PAGE 7

LIST OF FIGURES Page Figure 1. Examples of the stimuli used for the Chock and Word conditions. Stimuli used for the Flash and Nonsense conditions were comparable 11 Figure 2. Representative VER with component designations exemplified by Gastaut and Regis (1965) , illustrating the components used for data analysis in this study 15 Figure 3. Typical VERs recorded from a bright and dull subject Ig Figure 4. Schematic VER, based on mean latencies and amplitudes of each component for all subjects, under each of the four stimulus conditions. . 22 Figure 5. Mean latencies and amplitudes for 11 bright subjects compared with mean latencies and amplitudes for 11 dull subjects under each of the four stimulus conditions 27 VI

PAGE 8

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE EFFECTS OF STIMULUS CHARACTERISTICS ON THE RELATIONSHIP BETWEEN THE VISUAL EVOKED RESPONSE AND INTELLIGENCE By Patricia Ann Ondercin August, 1974 Chairman: Nathan W. Perry, Jr. Major Department: Psychology A number of investigators have studied the relationship between the visual evoked response (VER) and intelligence. However, these studies have been characterized by the use of relatively simple stimuli to elicit the response. Electrophysiological research with animals has indicated that neural processing varies with different types of stimuli. This study was designed to ac-sess the effects of different types of stimuli on the VER, and on the relationship between the VER and intelligence. Four stimulus conditions v/ere used: Flash, Checkerboard, V?ord , and Nonsense syllable. VERs v/ere recorded from positions C^ and C^ of the 10-20 International electrode system of 37 boys, ages ten and eleven, whose short form V.'ISC scores ranged from 88-138. The Culture Fair Test v;as also administered to provide a measure which v;as re.l /ely free from sociocultural bias. It was thought that t measure v/ovld yield higher correlations with VER characteristics.

PAGE 9

It was found that the highest and largest number of significant correlations with intelligence occurred under the Flash condition, and the fewest under the Word. The short form Wise yielded the largest number of significant correlations v/ith VER and the Verbal score almost as many; the Culture Fair Test proved to correlate poorly with the VER. The results confirmed the findings of other investigators that the latency of the VER is negatively correlated with intelligence. Amplitude was found to correlate negatively with intelligence in the first three components {IV, Va , Vb) , and positively in the last two components (Vc , VI). The results suggest that ongoing cognitive processes as v;ell as underlying neural organization are reflected in the VER. The relationship of attention and arousal to the correlations betv/een the VER measures and intelligence was discussed in terms of the different stimulus conditions and intelligence tests. It was pointed out that while the latency and amplitude of the VER were found to be significantly correlated with intelligence, variations between individuals were great, and these measures have little practical value at this time in the assessment of intelligence.

PAGE 10

INTRODUCTION Since the times when phrenologists attempted to assess brain power by the bulges of the forehead, science has searched for a relationship between brain and intelligence. It is generally accepted today that neural structure and function underlie cognitive capabilities. However, as recently as 1965, a reviev; of the literature concluded that no broad principles of the neurophysiological correlates of intelligence had yet been established (Ferguson, 1965). Early v/ork. seeking electrophysiological measures of cognitive ability was concerned with EEG frequency, particularly 10-14 Hz or alpha waves. Ellingson (1966) reviev;ed the area and concluded that the bulk of the evidence suggested no relationship between EEG and intelligence in adults. In children, results vzere contradictory and confounded by the effect which organic brain dysfunction has on both EEG ac• tivity and intelligence. These conclusions were refuted by Vogel and Broverman (1966) . Recent studies have found positive correlations between slow v.'aves and general ability (Vogel, Broverman, and Klaiber , 1968) and positive results using factor analytic techniques (Ishihara and Yoshii , 1972). It would seem, that no unequivocal statem.ents can be made at present about the relationship between the EEG and intelli-^ gence. In 1965, Cha].ke and Ertl reported striking correlations, 1

PAGE 11

as high as -.70, between IQ scores and the latency of the visual evoked response. The visual evoked response, or VER, is the computer-averaged sum of individual electrical responses elicited by a repetitive photic stimulus. Spontaneous, ongoing cortical activity averages 50 uV, while the response evoked by a single stimulus is less than 10 uV. Therefore, specialized computer techniques are required to extract the small signal, or response, from the larger "noise" of the EEC. Evoked responses are time-locked to a repetitive stimulus, and v/hen summed, they provide a record of the response to that stimulus. The averaged EEG, which is not time-locked, appears as a relatively straight line. Auditory and somatosensory, as well as visual stimuli can be used to generate an evoked response which is recorded on the scalp with surface electrodes. A diffuse light flash is the most commonly used visual stimulus, although patterned light may also be used. The early components of the evoked response are postulated to represent perceptual processing; the later components, information processing (John, Ruchkin, and Villegas, 1964; Uttal and Cook, 1964; Ertl , 196S). Since Chalke and Ertl's report, a nujiiber of investigators have studied the area. It is somev/hat difficult to assess the VER-IQ literature, as the studies are not strictly comparable, because of methodological differences. Subject populations studied vary widely, electrode recording sites differ, diverse measures of intellectual performance are used, and different VER characteristics are studied.

PAGE 12

A number of stud.i.e:s have been able to replicate Chalke ajxl Ertl's negative correlations of latency and intelligence in bright and dull adults (Plum, 196 8; Shucard and Home, 1972), in children (Ertl, 1968; Ertl and Schaefer, 1969), and with retardates (Bigum, Dustman, and Beck, 1970; Galbraith, Gliddon, and Busk, 197.0; Marcus, 1970). In general, correlations v;ere lower than those found by Chalke and Ertl, but statistically significant. Highly significant negative corKolations were found between latency of the neonatal VER and mental and motor development at eight months (Butler and Engel, 1969) , but there was no latency correlation with language at three years, or with IQ at four years (Engel and Fay, 1972), or at seven years (Henderson and Engel, 1974). Neither was a relationship between VER latency and intelligence found by Rhodes and his co-workers (Rhodes, Dustman, and Beck, 1969). Other characteristics of the VER have also been found to correlate with intellectual ability. Greater amplitudes of response components have been found in brighter children and adults than in those less bright (Rhodes et al . , 1969; Bigum et al. , 1970; Galbraith et al., 1970). Hov.'ever, Marcus (1970) reports larger response amplitudes in Mongoloid than in normal infants. Hemispheric asyimaetry of VER amplitude has often been noted, but results have been highly inconsistent. Several studies report c^iplitude asyr.unetry to be characteristic of normals, but not of dull or retarded children (Rhodes et al., 1969; Bigum at al., 1970, Ga.lbraith et al . , 1970). However, anotlier study (Richlin, Weisinger, V7einstein, Gianni, and

PAGE 13

Morganstern, 1971) found amplitudes greater in the right hemisphere than in the left in normals, and the reverse, left greater than right, in retarded children. Plum (196 8) found no relation between asynunetry and intelligence. A few investigators have used auditory evoked response (AERs) in studies of intelligence. These studies use clicks, white noise, or pure tones as stimulus, and record from central and temporal areas of the scalp. The amplitude of certain AER components were found to be larger in Mongoloid infants than in normal infants (Barnet and Lodge, 1967; Barnet , 1971). Response decrement, i.e., the progressive decrease in response amplitudes with repetitive stimulation, was seen in normal six to twelve month old infants, but was not seen in Mongoloid infants of the same age (Barnet and Lodge, 196 7; Barnet, Ohlrich, and Shanks, 1971). Latency differences were generally not found between normal and retarded, or normal and Mongoloid subjects (Barnet and Lodge, 1967; Barnet, 1971; Barnet et al., 1971; Richlin et al., 1971). The exception is Shimizu (1969) who reports a trend, although not a statistically significant one, toward larger response latencies in retarded subjects. He also found that AER latency and wave shape were reliable in normal adults, but inconsistent in mentally retarded adults. Despite the diversity of methodologies used in studies of correlations between the VER and intelligence, there is general agreement that low, but statistically significant correlations exist. It seems reasonable to expect a more complex visual stimulus to require more complex neural processing. Just as

PAGE 14

a more difficult behavioral task is a more precise indicator of intelligence than a simpler one, so a more complex neurological task should yield a more accurate picture of the cognitive efficiency of the organism. Simple diffuse flashes of light have been used as stimuli in all VER-IQ studies v/ith two exceptions v/hich have used checkerboard patterns (Galbraith et al. , 1970; Marcus, 1970). There is strong evidence that diffuse light is processed differently in the cortex than patterned light (Hubel and Wiesel, 1962; Perry and Childers, 1968). It seems possible then, that a VER evoked by patterned stimuli might be more reflective of differences in cortical processing related to intellectual ability, than would a VER evoked by diffuse stimulation. This rationale is readily subject to experimentation and testing. The major purpose of this study, then, is to determine how the correlations betv/een the VER and intellectual ability are affected by stimuli of greater complexity. It is hypothesized that VERS generated by patterned stimuli (checkerboard, word, and nonsense syllable) will yield higher correlations with IQ than the VER generated by diffuse stimuli. It is, of course, impossible to quantify 'visual complexity. One can confidently say that a patterned stimulus is more complex than a diffuse stimulus, but beyond that a rank ordering of complexity is hypothetical. One could argue that a word is the most complex because of its symbolic verbal content and "meaningfulness. " Yet, it could also be said that the nonsense syllable, as the most novel and unfamiliar stimulus,

PAGE 15

might initiate more sustained cognitive activity and attention. A case could also be made for the checkerboard, as it has the largest number of edges, has a verbal label, and may stimulate a variety of associations. The various stimuli cannot be ranked authoritatively for complexity then. However, it is postulated for the purposes of this discussion that the checkerboard is the least complex of the patterned stimuli since it is essentially non-verbal, is highly repetitive in content, and has a somev/hat limited association value. The word is considered to be more complex, since it requires cortical processing as a verbal and "meaningful" information. The nonsense syllable will be considered most complex, as it is a novel verbal stimulus which might have a large number of associations attached to it because it lacks any well-defined meaning. Another variable which has not been considered in the VER-IQ literature is the validity of the instrument used to measure intelligence. It is well-documented that socioeconomic status (SES) is related to poor performance on IQ tests, poor school achievement, lack of motivation, and slow development of language skills (Terman and Merrill, 1927; McNemar , 1942; Jones, 1954; Bloom, 1964a , 1964b;Kagan , 1970; Ginsburg, 1972) . Conventional intelligence tests contain items v/hich are educationally and culturally biased to the advantage of middle and upper SES groups, at the expense of the lower SES groups (Cattell and Cattell, 1959). In order to control for this bias, a test which is relatively free of contamii ton

PAGE 16

by the effects of school achievement will be administered in addition to a conventional intelligence test. It is hypothesized that VER-IQ correlations will be higher when intelligence is measured by this test than when it is measured by the conventional test.

PAGE 17

METHOD Subjects Thirty-seven boys between the ages of ten and eleven, v/ho attended P. K. Yonge Laboratory School, were chosen as subjects. This age group was chosen because the children were old enough to sit quietly and attend to the stimulus, yet young enough to avoid what unknown neurological effects puberty might have. The visual acuity of each boy was measured with a Snellen chart, and only those with an acuity of 2 0/25 or better in each eye were accepted for the study. In addition, Ss with a history of neurological dysfunction or visual defect were excluded. The IQs of the Ss, as measured by the short form of the Weschler Intelligence Scale for Children (WISC) , ranged from 8 8 (dull normal) to 138 (very superior), with a mean of 119. Psychometric testin g T\^7o intelligence tests were administered to each S. One, a short form of the WISC, consisted of the following subtests: Information, Tvrithmetic, Vocabulary, Picture Arrangement, and Block Design. This is the pentad which correlates best with the Full Scale WISC, r .92, when corrected for subtest reliability (Silverstein, 1970). Verbal and Performance scores for the V;fISC were calculated separately as well. In addition, the Culture Fair Intelligence Test was administered in a group

PAGE 18

to all Ss. This was chosen as a measure because it is considered to be relatively free from specific educational and social biases (Cattell, 1940; Cattell, Feingold, & Sarason, 1941) , as well as a valid indicator of general ability (Cattell et al. , 1941; Tilton, 1949; Geist, 1954; Marquant & Bailey, 1955) . VER recording procedure Silver-silver chloride electrodes (Beckman) were used to record VERs monopolar ly from tlie scalp, from positions C^ and C^ of the International 10-20 electrode system, with the reference electrode clipped to the ipsilateral ear. These locations have been used by a number of investigators in both monopolar and bipolar derivations (Ertl, 1968; Plum, 1968; Rhodes et al. , 1969; K-einberg, 1969; Richlin et al., 1971). Microdot cable vjas used for leads from the electrodes to the amplifiers, in order to minimize movement artifacts. Electrical activity from the scalp was amplified by Grass P-511 Amplifiers (Bandwidth 0.15-50 Hz) during stimulus presentation. The EEG signals were monitored visually on a Tektronix Type 564 oscilloscope. After amplification, the electrical activity was simultaneously routed to four channels of a Computer of Average Transits (CAT 4 OB) for on-line summation, and onto a seven channel FM magnetic tape recorder (Sanborn 7000) for subsequent analysis. The stim.uli were prepared on 2" x 2" slides and were projected onto a diffuse light screen by two Viewlex V-27 projectors which v;ere custom-mounted on a conmion base. Stimulus

PAGE 19

10 duration was 500 msec, with an interstimulus interval of 1500 msec, determined by Gerbrandt electronic shutters controlled by a Grass 5-8 Stimulator. The long stim.ulus duration was used to avoid the "on" and "off" response mixtures obtained with short pulse or strobe stimulation. The stimuli subtended a visual angle of 6 " on a side and were viewed under binocular conditions. The S was seated in a padded chair and stimuli appeared on the screen 6 ft. in front of the S_^ The luminance level of the stimuli was 8 ft. cdl . on a dark surround for all stimulus conditions, and was equated by a Variac variable transformer. Luminance was measured by a photometer (UDT 40A Opto-IJeter) , with the sensor at approximately the same distance from the screen as the S. Each VER recording was the summation of responses to 6 stimulus repetitions; two such recordings were obtained for each of the four stimulus conditions. The four stim.ulus conditions used v/ere: 1. Diffuse light flash 2. Checkerboard pattern 3. V7ord ("FOR") 4. Nonsense syllable ("rfo" ) (see Figure 1) . The specific v/ord was chosen because it is a high frequency word (Kucera and Francis, 1967) which is classified at a primary reading level. The nonsense syllable is a recombination of the same three letters. In order to minimize the effects of . tuation (response decrement occuring with repetitive

PAGE 20

11

PAGE 21

12 stimulation) , two stimuli were presented in random order during a single trial. An incremental film strip reader was used to program the random order of presentation. A binary signal was recorded on a channel of the tape to enable relay switching in the CAT to summate responses evoked by each stimulus separately. The Flash and Check were in two trials, and the V7ord and Nonsense in the other two. A trial consisted of 60 presentations of each stimulus, or a total of 120 presentations. Each trial took four minutes, and a brief (two minute) rest period was allowed after each trial. In addition, tv;o control trials in which no light stimulation reached the eyes were performed to test for the intrusion of artifacts. The control trials consisted of 60 repetitions, and took two minutes. The order in which stimulus conditions and control trials were presented was randomized. Experimental procedure Upon arrival in the laboratory, the S was tested for visual acuity vi/ith a Snellen chart. The intelligence testing had been completed previously at the school. The S ' s head was measured for electrode placement, the sites cleaned with alcohol, and the electrodes placed on positions C-, and C.. The S was then seated in an electrically shielded, sounddampened and light-proof room. He was instructed to sit quietly without moving his head, and to watch the flashing lights. The S was then fitted with earphones through which white noise was transmitted in order to prevent the sound of the shutters from evoking an auditory response. The lights inside th^ room

PAGE 22

13 were turned off, and the trials began approximately 2 min. later. The session lasted about 30 min. Data analysis Analog data tapes were played back following the experimental procedure and data obtained in analog form by a Varian F-50 Plotter. This yielded for each S two VERs for each condition, which were then superimposed and averaged by visual inspection. All subsequent analyses utilized this single averaged VER. Latency was measured in milliseconds from the beginning of the response to each peak. Amplitude was determined by measuring vertical distance in microvolts, with reference to the preceding peak. Latencies and amplitudes of the components were then correlated with the intellectual measures using the Pearson product-moment procedure. Only the later VER components (80-400 msec.) v^ere correlated with the intelligence measures, since most investigators have found that correlations with intelligence occur within that range (Rhodes et al. , 1969; Galbraith et al., 1970).

PAGE 23

RESULTS Latency and Amplitude The data appear to best fit the waveshape described by Gastaut and Regis (1965) , and the components were labelled IV, Va, Vb, Vc, and VI (see Figure 2 ). Only these five waves were analyzed for the purposes of this study, since these have been found to be most related to measures of intelligence. The components IV, Va, and VI were quite stable across S s , and Vb and Vc less so. A sample of the VER data can be seen in Figure 3 . There were a total of 66 significant correlations out of a possible 320. Correlations ranged from +.55 to -.72, with a mean of -.17. With a sample size of 37 S s , correlations of .33 and above are significant at the .05 level; however, due to the absence of particular components in the VERs of some Ss, the actual sample size for statistical purposes numbered as low as 20, requiring a correlation of .42 for significance at the .05 level (see Table 1 ). Some of the correlations are strikingly high, up to -.72, which are as high as those achieved by Chalke and Ertl (196 5) and by Galbraith and his co-workers (1970) . They are considerably higher than those achieved by several other studies (Plum, 1968; Shucard and Horn, 1972, 1973). These high correlations appear to be densely clustered around the three central components of the response, Va, Vb, and Vc, especially under 14

PAGE 24

15 Va 100 msec Figure 2. Representative VER with component disignations exemplified by Gastaut and Regis (1965) , illustrating the components used for data analysis in this study.

PAGE 25

16 Left 100 msec Right Figures. Typical VERs recorded from a bright and dull subject.

PAGE 26

ro LO 17 T5 W C C (0 O -rH W -P Qj -H d a W O ns u 0) g w W 3 > g OJ W > M 4J rJ 0) O CO ^ C (1) O TJ •H C •4J 3 rH W GJ 4J >H W 5-1 Q) O 4J n3 (U O Cn Cn-P rH c G S-l H o ^^ o o u CO O Q) o c CU W x; u u 4J U 0-) C/3 rH 0) H M 3 IS :s 3: U I I I I O a Cm E-i r/l d:; o; [i, H w w u s > p^ U3 in r~ to LT) IT) «

PAGE 27

ca m r^ tn •^ in 18 U CQ fri Eh cn Di Oi [ii H W W U S > D^ O KD O rH o m t| B CO cr; 2 pm M w w u > a. rH M > > rH rH in iH 0) -p -p -P -P c a (13 (T3 O O •H -H •rH -H G c: •H -H to w >1 >1 rH iH to m o o •H -H P -P W W H -H P +J •P 4J en -H •H g e d) 0) Xi p -p ^ Q) -H h^l Pi r I

PAGE 28

19 the Flash condition, but also with the Checkerboard. Correlations tend to be lower and more scattered with the Word and Nonsense conditions. The largest nuniber of correlations and those of the greatest magnitude were associated with response latency, but some high correlations were also seen with response amplitude. These high correlations occur primarily with the short form WISC and the Verbal score. With one exception, all the correlations between response latency and the intellectual measures were in the negative direction. Short latency was associated with higher intellectual abilities and long latency with less ability. The picture is more complex with regard to amplitude. With the early wave components (IV, Va, Vb) all correlations were negative, indicating that larger amplitudes were related to less ability, and smaller amplitudes with more ability. However, with the two late components (Vc, VI) the correlations were in a positive direction; larger amplitudes were associated with higher intelligence and smaller amplitudes with lower intelligence. The number of significant correlations was different for latency and amplitude measures. There were 43 correlations above the level of significance with the latency measure, as compared with 23 for the amplitude measure. Measures of Intelligence The number of significant correlations also varies with the measure of intelligence used. The VER measures were correlated with the short form WISC, its Verbal and Performance Scores, and with the Culture Fair Test. It was the short form

PAGE 29

20

PAGE 30

21 ld31 (A^) 3ani 1H9 HdlAIV

PAGE 31

22 id3n iHom 0) u 3 •H t4

PAGE 32

23 trend toward increasing across conditions. That is, the component latencies are shortest under the Flash condition, longer under the Check, increase little or not at all with the Word, and are longest under the Nonsense condition. Comparison of Bright and Dull Ss It was thought that additional data could be gathered by a comparison of the brightest and dullest of the Ss within the sample. Twenty-two Ss were chosen by selecting the eleven boys with the highest short form WISC IQs , and the eleven with the lowest. The IQs of the bright group ranged from 125 to 138; those of the dull group from 88 to 106. These groups were significantly different for intelligence at the .01 level, on all four measures of intelligence. Means and standard deviations were calculated for each group, and t-tests performed between groups (see Table 2). As indicated by the previously mentioned correlations between VER measures and intelligence, the bright group has shorter latencies than the dull group under all stimulus conditions. The bright children tend to have smaller amplitudes in the earlier wave components than did the duller children, but larger amplitudes in the later components . The difference between bright and dull groups is more clearly illustrated in Figure 5. Variations among stimulus conditions are apparent, as are variations between left and right hemispheres. For both Flash and Check conditions, the right hemisphere responses appear more flattened, with generally smaller amplitudes, than those of the left hemisphere. The

PAGE 33

oi « 24 o CO (0 o Pi Pi ci IX IX 4J ix ix c

PAGE 34

25 rj* (N rH \0
PAGE 35

26 I I ' ' O o o o CM -L— ±.A^5 — L_J — ^L — L L^ o o o o o o o o C4 ~ "T »^ • » I (A^) 3anindt^'v

PAGE 36

27 ld3"1 IHOId (A^) 3anind{Aiv P4

PAGE 37

28 waveshapes of the responses from the right hemisphere are also quite similar in bright and dull groups. This is not true m the left hemisphere where the waveshapes for bright and dull groups appear markedly different, primarily due to the amplitude of the Vb component. With the Word and Nonsense conditions, it is also in the left hemisphere that waveshape differences between bright and dull groups are more apparent, again, primarily due to the amplitude of Vb. These hemispheric differences in waveshape cannot be accounted for by hemispheric asymn\etry in either group, since t-tests performed between hemispheres for each group did not reach statistical significance. Rather they seem due to the differential amplitudes and to some extent, latencies, between groups. However, the number of Ss in each group was small, and the standard deviations, especially of the amplitude measures large, so that hemispheric asymmetry cannot be completely discounted as a contributing factor. O C

PAGE 38

DISCUSSION The major hypothesis of this study, that more complex visual stimuli would correlate more highly with measures of intelligence than simpler stimuli was not upheld. Although there were clear differences between conditions, they were in the opposite direction from that predicted: it was the light flash which accounted for both the highest and the largest number of significant correlations with intelligence. Several explanations might account for this. it may be that response frequencies in the alpha range, 10-14 Hz, contribute heavily to the relationship between the VER and intelligence, and the correlations are best when this frequency is most in evidence, i.e., with diffuse stimuli, as in the Flash condition. in earlier studies, frequency analysis of the VERs elicited by • diffuse stimulation revealed a predominance of frequencies in the al£ha range (Ertl, 1971), while those frequencies were rarely seen when visually complex stimuli were used (Perry, Childers, and Falgout, 1972). Weinberg (1969) reports that the highest correlations with intelligence are associated with the frequencies of 12-14 Hz in VERs elicited by diffuse stimulation. Another explanation of these differences might be based on the findings of several studies that VER differences related to intelligence tend to be obscured by increasing the Ss29

PAGE 39

30 level of attention or arousal (Plum, 1968; Shucard and Horn, 1972) . Response amplitudes increase during conditions of high attention, while response latency decreases (GarciaAustt, 1963; Haider, Spong, and Lindsley, 196 4; Gross, Begleiter, Tobin, and Kissin, 1965). It will be remembered that most studies have found both shorter latencies and larger amplitudes were related to higher intelligence in children. It seems likely that brighter children are generally in higher states of arousal and attention than duller children, but the imposition of a simple task or arousal device stimulates relatively more arousal in duller S s , thus obscuring the differences between them. Visually more complex stimuli, such as the check, word, and nonsense syllable, may be more arousing and attention-getting for the duller Ss than for the brighter, and thus tend to obscure differences between groups. A flash would have less arousing qualities, and therefore emphasize the intrinsic differences in arousal level between S s . If component amplitudes, which are considered to reflect arousal, are ranked for size across conditions, there is a suggestion of a trend in this direction, although it is not of statistical significance. For dull S s , the Nonsense condition elicits the highest amplitudes (suggesting higher arousal) , and the Flash condition the lowest amplitudes. With the bright S s , amplitudes are more nearly equal across conditions, indicating a more uniform level of arousal which seems less affected by extrinsic characteristics of the stimuli. This would lend support to the hypothesis that brighter children intrinsically maintain a

PAGE 40

31 higher level of arousal, and that duller children are less able to sustain attention v/hen presented with simple stimuli, but are relatively more aroused by complex or meaningful stimuli. It is interesting to note the differences in the composite waveforms across stimulus conditions. The waveshape evoked by the flash is distinct from those evoked by the patterned stimuli, appearing more like a "WV This is consistent with evidence that diffuse and patterned light are processed differently in the cortex. The increasing latencies of the composite VERs across conditions are also suggestive. The most complex stimulus, the nonsense syllable, shows the longest response latencies of the four stimulus conditions, suggesting it requires a relatively longer processing time in the cortex. The least complex stimulus, the flash, shows the shortest latencies, and might indicate the relatively quicker cortical processing of simple stimuli. The word and the check show more nearly equal latencies, midway between those of the flash and nonsense, and might indicate a similarity of cortical processing. It is possible that the check is being given an immediate verbal label by the S, and so is processed as a word, as well as a configuration. The differences between the Word and Nonsense conditions, both in latencies and in the number of significant correlations are not of statistical significance. They are intriguing, however, because they are composed of identical letters, and more similarity might be expected if cortical processing was

PAGE 41

32 also identical. Although they vary in familiarity or novelty, it is debatable how novel any stimulus can be after 60 repetitions, so it would seem that meaningfulness is the principal dimension along which they vary. The word is defined by an assigned meaning, and in that sense is somewhat limited. The nonsense syllable has no particular meaning assigned to it, and is therefore more open to interpretation and varied associations. This less restricted quality may be more stimulating to the brighter Ss than the duller, and emphasize differences between them. There is evidence to suggest that meaningfulness of the stimuli is associated with enhancement of the VER (Symmes and Eisengart ,1971) , which might indicate increased arousal or attention. It is interesting to speculate on the possibility of a curvilinear relationship between the correlations with intelligence and the arousal value of the stimulus. The flash, as the simplest stimulus, is not very arousing for either group, and their intrinsically different levels of attention or arousal are made apparent. Word and check provide extrinsic arousal, which is relatively more arousing for the duller Ss, obscuring intersubject differences. The nonsense syllable also provides extrinsic stimulation, more so than the word or check, because of its lack of specificity, and makes differences between bright and dull Ss more general. Another major hypothesis of this study was that the Culture Fair Test would prove better instrument for assessing intelligence in relationship the VER than would the WISC. However, just the reverse was shown to be the case. This is

PAGE 42

33 surprising since the Culture Fair Test emphasizes both speed and skill in the analysis and interpretation of visual information, abilities which one would expect to be important in the processing of visual stimuli in the VER. Rather, it is the short form WISC which is heavily loaded for verbal abilities, and the Verbal subtest of the WISC, which yields the majority of the significant correlations. These tests primarily measure verbal comprehension and skills related to school achievement. There are generally thought to be three factors involved in intelligence: the ability to encode information, the ability to retain information over time, and the ability to retrieve information. Retrieval of stored inforamtion has two aspects: the recall of stored data in their original form, and the manipulation of relevant data to form new combinations. The WISC would seem to rely heavily on the more passive recall of learned information. The Culture Fair Test, in contrast, presents unfamiliar stimuli and demands a more active process of retrieval and recombination of relevant data in a new situation. This is consistent with the previous hypothesis that the VER correlates better with intelligence under less arousing conditions than under more arousing ones. The type of intelligence reflected in the VER then, would seem to be more in a passive, receptive mode than a more active, manipulative mode. It is apparent from the results of this and other studies that the latency and amplitude of the VER have a significant

PAGE 43

34 relationship to intelligence, brighter Ss tended to have shorter response latencies, duller Ss tended to have longer latencies. It is thought that the shorter latencies reflect faster and more efficient neural processing of incoming stimuli, while longer latencies reflect slower and less efficient processing. However, a variation of one of Spitz's (1963) postulated of neural functioning in the mentally retarded might also apply. Spitz states that a relatively longer time is required to induce a temporary change in stimulated cortical cells in retardates than is required in normals. This implies that longer response latencies would be seen in the VERS of retardates as compared to normals. In a comparison of the VERS of normal and retarded S s , Galbraith and his coworkers (1970) have lent support to this hypothesis. It seems likely that this postulate would also apply more generally to the range of intellectual functioning in a normal population, i.e., in brighter individuals, cortical cells are more rapidly activated by stimulation than in dull individuals. The implication of shorter response latencies for brighter individuals was clearly borne out by this stuc , The correlations between intelligence and response amplitude are conf-.-ing, however, and agree only in part with other work. The ea ler wave components are negatively correlated with intellig. r.ce. Higher amplitude is associated with the duller Ss, lower amplitudes with brighter Ss. Dustman and Beck (1969) reported that between the ages of 5 and 13, the amplitude of the BER response shows a marked decrease. The components which are reduced fall in the range

PAGE 44

35 of 70-225 msec, which roughly corresponds to the latencies of components IV, Va, Vb. It seems possible that the brighter children are developmentally more mature than less bright children, and this difference is reflected in smaller component amplitudes characteristic of more advanced development of the central nervous, system. In summary, it seems clear that correlations can be found between the VER and measures of intelligence. These correlations appear to reflect both ongoing cognitive processes and underlying neural organization. However, it must be stressed that while VER-IQ correlations were significant, the possibility of assessing the intelligence of an individual is very limited due to the large variability of the response. Callaway (1973) makes the point that even perfect VER-IQ correlations would give a measure that was no better and very likely more expensive than a conventional test. Yet VER measures are less affected by specific learning and school performance, and may well cast new light on individual differences in cognitive functioning. It might prove especially useful in the early detection of subtle learning dysfunctions, such_ as dyslexia.

PAGE 45

REFERENCES Barnet, A. B. Evoked potentials in handicapped children. Developm . Med. Child Neurol. , 1971 13_, 313-320. Barnet, A. B. and Lodge, H. Click evoked EEC responses in normal and developmentally retarded infants. Nature, 1967, 21_4, 252-255. Barnet, A. B., Ohlrich, E. S., and Shanks, B. L. EEC evoked responses to repetitive auditory stimulation in normal and Down's syndrome infants. Developm . Med. Child Neurol . , 1971, 13_, 321-329. Bigum, H. B. , Dustman, R. E. and Beck, E. C. Visual and somato-sensory evoked responses from mongoloid and normal children. Electroence ph. Clin. Neurophysiol . , 1970. 28^, 576-585': Butler, B. V. and Engel, R. Mental and motor scores at eight months in relation to neonatal photic responses. Developm. Med . Child Neurol . , 1969, 1]^, 77-82. Bloom, B. S. Stability and change in human character istics. New York: Wiley, 1964a. Bloom, B. S. Compensatory education for cultural deprivation . Chicago: University of Chicago, 1964b. ~ Callaway, E. Correlations between averaged evoked potentials and measures of intelligence. An overview. Arch. Gen Psychiatry , 1973, 29_, 553-558. ' Cattell, R. E. A culture free intelligence test. I. J. Educ Psychol . , 1940, 31^, 161-180. ~ ' Cattell, R. B. and Cattell, A. K. S. Handbook for the culture f^^^ intelligence test . Scale 2. Cham|)aign, lITT Institute for Personality and Ability Testing, 1959. Cattell, R. B. , Feingold, S. N. , and Sarason, S. B. A culture fair intelligence test. II Evaluation of cultural influences on test performance. J. Educ. Psychol. , 1941 32 , 81-100. ~ — Chalke, F. C. R. and Ertl, J. Evoked potentials and intelligence. Life Sciences, 1965, 4_, 1319-1322. 36

PAGE 46

37 Dustman, R. E. and Beck, E. C. The effects of maturation and aging on the waveform of visually evoked potentials. Electroenceph . Clin . Neurophysiol . , 1969, 26_, 2-11. Ellingson, R. J. Relationship between EEG and test intelligence: A commentary. Psychol . Bull . , 1966, 6_5, 91-98. Engel, R. and Fay, V7. Visual evoked responses at birth, verbal scores at three years, and IQ at four years. Developm . Med. Child Neurol . , 1972, 14_, 283-289. Ertl, J. P. Evoked potentials, neural efficiency and IQ. Paper presented at the International Symposium for Biocybernetics, Washington, D. C, 196 8. Ertl, J, P. Fourier analysis of evoked potentials and human intelligence. Nature, 1971, 230 , 525-526. Ertl, J. P. and Schaefer, E. W. P. Brain response correlates of psychometric intelligence. Nature , 1969, 223, 421-422. Ferguson, G. A. Human abilities. Ann. Rev. Psychol., 1965 16, 39-62. — ^ ' Galbraith, G. C. , Gliddon, J. B. , and Busk, J. Visually evoked responses in mentally retarded and nonretarded subjects. Amer . J. Ment . Deficiency , 1970, 75_, 341-348. Garcia-Austt, E. Influence of the states of awareness upon sensory evoked potentials. Electroenceph . Clin . Neurophysiol . , 1963, Supp. 24, 76-89. Gastaut, H. and Regis, H. Visually-evoked potentials recorded transcranially in man. In L. D. Proctor and W. R. Adey (Eds.) The analysis of central nervous system and cardio vascular data using computer methods. (1964 Symposiuin) 1965. Washington, D. C, NASA, 7-34. Geist, H. Evaluation of culture free intelligence. Calif J Educ. Res. Di£. , 1954, 5, 209-214. ~ Ginsburg, H. The myth of the deprived child . Poor children's intellect and education . Englewood Cliffs, N J 'Prentice-Hall ,1972 . Gross, M. M. , Begleiter, A., Tobin, M. , and K: 3sin, B. Auditory evoked response comparison during cc nting clicks and reading. Electroencep h. Clin. Neuror -^iol. , 1965 18, 451-454. Haider, M.,_Spong, P., and Lindsley, D. B. Cc cal evoked potentials during visual vigilance task f ormance. Electroenceph . Clin . Neurophysiol . , 1964, , 710-715.

PAGE 47

38 Henderson, N, B. and Engel, R. Neonatal visual evoked potentials as predictors of psychoeducational tests at age seven. Developm . Psychol . , 1974, 10_, 269-276. Hubel, D. H. and Wiesel, T. N. Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. J. Physiol . , 1962, 160, 106-154. Ishihara, T. and Yoshii , N. Multivariate analytic study of the EEG and mental activity in juvenile delinquency. Electroenceph . Clin . Neurophysiol . , 1972, 33, 71-80. John, E. R. , Ruchkin, D. S., and Villegas, J. Experimental background: signal analysis and behavioral correlates of evoked potential configurations in cats. Ann N Y Acad. S£i . , 1964, 112, 362 420. -* "' Jones, H. E. The environment and mental development. In L. Carmichael (Ed.) Manual of child psychology , 2nd edition . New York: VViley, 1954, 631-696" Kagan, J. On class differences and early development. In V. H. Denenberg (Ed.) Education of the infant and young child. New York: Academic PressT l9ToT Kucera, H. and Francis, W. N. Computational anal ysis of present-da^ American English . Providence! Brown — University Press, 1967. Marcus, M. M. The evoked cortical response: a technique for fnf^^^i"^^ development. Calif. Ment . Heal th Res. Dig., 1970, 8, 59-72. — ^ Marquant, D. I. and Bailey, L. L. An evaluation of the culture 353-35r intelligence. J. Genet . Psychol ., 1955, 86, McNemar, Q. The revision of the StanfordBine t. BostonHoughton Mifflin, 1942. Perry, N. W., Jr. and Childers, D. G. Cortical potentials in normal and amblyopic binocular vision In E Schmoeger (Ed.) Advances in Electrophysiology and Pathology of tho Visual System. Leipzig, Thiemrri96 8, 1-31-161. Perry N. W. Jr., Childers, D. G., and Falgout , J. C. Chroma^J^/P?5^fi?ity of the visual evoked responses. Science , 1 -^ ' ^ , 1 / / , oXj— 815. ~~~" Plum, A. Vis ual evoked responses : Their relationship to 1968 ^^^''''^* °°''^°''^^ dissertation. University of Florida,

PAGE 48

39 Rhodes, L. E. , Dustman, R. E., and Beck, E. C. The visual evoked response: A comparison of bright and dull children. Electroenceph . Clin . Neurophysiol . , 1969, 27, 364-372. Richlin, M. , Weisinger, M. , Weinstein, S., Gianni, M. , and Morganstern, M. Interhemispheric asymmetries of evoked cortical responses in retarded and normal children. Cortex , 1971, 7, 98-105. Shimizu, H. AER in the severely retarded. Excerpa J>ledica International Congress Series , 1969, 206 , 530-534T Shucard, D. W. and Horn, J. L. Evoked cortical potentials and measurement of human ability. J. Comp. Physiol Psychol . , 1972, 78^, 59-68. ~ — * Shucard, D. W. and HOrn , J. L. Evoked potential amplitude change related to intelligence and arousal. Psychophysioloqy , 1973, 10, 445-452. Silverstein, A. B. Reappraisal of the validity of WAIS, WISC and WPPSI. J. Consult , and Clin. Psychiat. , 1970, 34. 12-14. — — Spitz, H. H. Field theory in mental deficiency. In N. R. Ellis (Ed.) Handbook of mental deficiency. New York • McGraw-Hill, 19G3. Symmes, D and Eisengart, M. A. Evoked response correlates of meaningful visual stimuli in children. Psychophysioloqy 1971, 8, 769-778. — ^^-^ ^^ Terman,_L. M. and Merrill, M. A. Measuring intelligence ; A g^^^Q to the administration of the new revised Stan fordBmet tests. Boston: Houghton Mifflin Co. , 193T; Tilton, J. R. A survey of the reliability, validity and usefulness of the Cattell Culture Fair Test. Persona . , 1949, l , 17 — 19. Uttal, W. R. and Cook, L. Systematics of the evoked somatosensory cortical potential: a psychophysical-electrophysiological comparison. Ann. N. Y. Acad. Sci 1964 112 , 60-79. ' Vogel, W. and Broverman, D. M. A reply to "Relationship between EEG and test intelligence: A commentary " Psychol. Bull . , 1966, 6_5, 99-109. Vogel, W., Broverman, D. M., and Klaiber, E. L. EEG and mental abilities. EI otroenceph . Clin . Neurophysiol . , 1968, 24,

PAGE 49

40 . Correlation of frequency spectra of averaqed Weinberg, H. Correlati visual Nature

PAGE 50

BIOGRAPHICAL SKETCH Patricia Ann Ondercin was born in Racine, Wisconsin in 1948. She attended St. Catherine's High School there, and received her diploma in 1966. Her undergraduate studies were done at Marquette University, in Milwaukee, Wisconsin. She was graduated cum laude in 197 with a Bachelor of Arts in Psychology and English. She began her graduate studies at the University of Florida the following September, and was awarded the degree of Master of Arts in December, 1971. In July, 1973, she was appointed to an Internship in Clinical Psychology at the New York Hospital-Cornell Medical Center, in White Plains, New York. 41

PAGE 51

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Nathan W. Perry, Jr., Chairman Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. (kj^ cqu^lin R. Goldman Associate Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Wiley C. l^asbury Assistant* Professor of Psychology _ I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Calvin K. Adams Assistant Professor of Psychology

PAGE 52

_ I certify that I have read this study and that in my opinion It conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. L^MuT^^/M 9/jc^ Arnold ^. Nevis Professor of Electrical Engineering This dissertation was submitted to the Department of Psychology in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1974 Dean, Graduate School