Induced coding strategies and hemispheric differences in matching letter pairs

OOZS-3932/82/060669-4l6W3.C0/0 x-8 1982 Pergamon Press Ltd.

Nnropsycholoy~o. Vol. 20. No. 6. pp. 6694~74, 1982. Pnnted m Great Britam.

INDUCED CODING STRATEGIES AND HEMISPHERIC DIFFERENCES IN MATCHING LETTER PAIRS* JOHN EDWARDS-~ Department of Psychology,

University

of Reading,

Earley Gate, Whiteknights,

Reading

RG6 2AL, U.K.

and

PETER H. VENABLES Department

of Psychology,

University (Receir&

of York, Heslington,

York YOl

5DD, U.K.

16 Ma): 1982)

importance of task requirements in producing visual half-field asymmetries was investigated through manipulation of the criteria by which a central probe and a unilateral target letter were matched. Physical characteristics of the letters were emphasized (those made up of straight lines only vs those having at least one curved feature, e.g. F vs P). An interaction ofvisual field and type of judgment (match vs mismatch) was found for manual response time. Matching responses were faster to RVF than LVF probes and mismatches were faster to LVF than RVF probes. The significance of this interaction for models of hemispheric asymmetry is discussed. Abstract-The

INTRODUCTION RECENT studies

have demonstrated the influence of induced coding strategies on the processing of words and letters [ 1,2]. For example, both right and left hemifield superiorities have been generated with the same set of letters as a function of coding instructions (naming vs physical judgment) given to the subjects [3]. These studies move away from descriptions of functional specialization in terms of simple dichotomies towards a more process oriented approach, in which particular emphasis is placed on the mode of information processing induced by a particular task. As a consequence, it has become important to be aware of the interaction of the effects that the stimuli, the task, and the nature of the decision required have (e.g. same vs different) in producing a particular pattern of visual half-field asymmetries. For example, MARTIN [4] used the same set ofletters in two tasks: (1)judgment ofwhether pairs of letters contained a “curve” (a “visual” task); (2) judgment of whether pairs of letters rhymed with the sound of the letter “E” (a “verbal” task). While Martin did not obtain a Task by Visual Field interaction his results are of interest because they suggest that both his tasks induced verbal coding (a RVF advantage was found overall) and highlight the problem of ad hoc designation of tasks as “visual/verbal” or “right/left” hemisphere. The present study was designed to assess the role of task and decision processes by constructing two judgment criteria for matching pairs of letters that emphasized their “visual” or “physical” characteristics. Positive judgments were made to letter pairs consisting

*This study was carried out while the first author was in receipt of a Research Studentship Research Council of Great Britain. tAuthor to whom requests for reprints should be addressed. 669

awarded

by the Science

670

JOHN EDWARDSand PETERH. VWAHLES

of all straight lines or where both letters had a curved element; negative judgments were required when one letter was “curved” and the other “straight”. If it is assumed that both these tasks will be performed more efficiently as a result of visual coding ofthe letter pairs then a right hemisphere advantage should be obtained for both types of judgment. On the other hand, it has been proposed that two simultaneously occurring processes operate in a letter matching task: an “identity reporter” which emits a response if the stimuli are the same and a “serial process” which produces a response if they are different [S]. It is possible that the positive and negative judgments of this task correspond, respectively, to these processes. Taken together with evidence to suggest that when a task demands comparison on the basis of separate and single features (analytic processing) the left hemisphere will predominate for difference judgments whereas when the task favours gestalt matching, right hemisphere processes are more compatible for both same and different judgments [6-83, a Judgment by Hemisphere interaction is predicted with positive judgments favouring right hemisphere processes and negative judgments favouring the left hemisphere. METHOD Subjects were 12 male volunteers, all undergraduates at the University of York. Average age was 21.9 yr (range 19 24 yr). All were strongly right-handed as shown by the self-report inventory developed by WHrrti and ASH I ou 193. All were right-eye dominant as assessed by the A B C Vision Test of sighting dominance. and had normal or corrected to normal vision.

The stimulus set comprised the upper case letters V, A. E. F. U. B, R, P (Letraset: 4X-point Helvetica Light). The first four were chosen because they are made up of straight lines only and the remainder because they have at least one curved feature. The letters were also selected on the basis that the “straight” and “curved” set were similar in general shape and size. Two sets of tachistoscope cards were prepared: the target set consisting of one letter positloned at the centre of the card and the probe set with each card having a single letter centred 4’ to the left or right of the middle point. The closest edge of any letter was at least 3.5’ from the fixation point. An Electronic Developments three-field tachistoscope was used to present the stimuli. The subject Indicated a positive or negative judgment manually using a two-way switch which moved in a plane perpendicular to the visual field seen through the tachistoscope. The switch was held between thumb and forefinger (see below). Reaction time (RT) was measured to the nearest millisecond.

The subjects were instructed to match pairs of letters on the basis of particular physical characteristics unrelated to the name of the letter. Using a repeated measures design subjects were required. on the same block of trials. to make two types of judgment. “Positive” judgments required the subject to decide whether both letters were constructed from straight lines only or whether both had a curved element. A “negative”judgment was required when one letter was “curved” and the other “straight”. Subjects were instructed to perform the task as qutckly and as accurately as possible. Direction of lever movement and hand of responding were counterbalanced across subjects. Each subject received three blocks of 40 trials: each block consisted of 20 positive and 20 negative judgments. counterbalanced fol- visual field. All subjects received the three blocks in the same order. Visual half-field was randomized wzthin blocks.

On each trial the subject saw the following sequence ofevents, following a ready signal: a fixation field for 2 set: a centrally placed target letter for 20 msec; fixation field for 50 msec; probe letter for 20 msec. fixation field fat50 msec; dark field during an inter-trial interval of 18 lo 22 sec. A 2 min rest period was allowed between blocks and each block commenced with two non-recorded trials. Ten practice trials were given initially. No knowledge of results was given except during practice. The Importance of central fixation was repeatedly emphasized.

CODING

STRATEGIES

AND HEMISPHERIC

671

DIFFERENCES

RESULTS RTs were recorded for correct judgments only. Both mean and median RTs were computed for each subject (see Tables la and lb). This was because in previous studies either has been used rather arbitrarily. The aim here was to discover whether they provide the same results. A four-way ANOVA-based on mean RT-with one between (Responding Hand) and three within-subject factors (Visual Field; Judgment; Block) revealed only a significant main effect for Blocks, F (2, 20) = 24.724, P
Table

I(a). Means and standard

deviations:

RT (S-mean),

Judgment

LVF

RVF

Positive SD. Negative S.D.

1008 150 1000 1x5

962 153 1023 184

l(b). Means and standard

deviations:

RT (S-median),

Judgment

LVF

RVF

Positive S.D. Negative SD.

934 128 967 203

923 145 998 181

. .. .

msec

msec

.I

. . ..-. _ /. . ... . ...* / =_..

+

970

950

I

1

\

L

R Visual Field

FIG. 1. Manual

RT to letter probes

as a function

of visual field and type of judgment

(+ ve vs - ve).

612

Jorw

EDWAKI)S and PI n

K

H. Vt-\a~

IS

simple main effects [lo] used to unravel the two-way interaction revealed that positive judgments were made significantly faster than negative judgments to probes in the RVF, F (1, 44) =4.5, PO.l, although the pattern of results was the same as that shown by mean RT. An analysis of the error data revealed a main effect for Visual Field approaching significance: LVF probes produced more errors than RVF probe [4.3”/:, vs 2.6%); F (1, 1 l)= 10.6, P=O.O8). Finally, inspection of the data revealed no difference in the pattern of results obtained with the two components of the positive judgment [mean RTs of 997 msec (both straight). 1019 msec (both curved) vs 959 msec (both straight), 965 msec (both curved) for LVF and RVF probes, respectively].

DISCUSSION The most interesting aspect of the results concerns responses to RVF probes. Contrary to prediction, these were faster than to LVF probes for positive judgments and, within the RVF positive judgments were faster than negative decisions. In addition, RVF probes produced more accurate performance than LVF probes. Overall, this pattern suggests some form of left hemisphere involvement for positive judgments. It is possible that the positive judgment required a more analytic processing strategy of the subjects because they had to check every feature of the two letters in order to make a correct decision. Faster positive than negative judgments in the RVF may also reflect the operation of the proposed rapid “identity” reporter c5, 111. It is possible that the similar RTs for the two types ofdecision to LVF probes reflect similar coding strategies. However, the data cannot confirm this hypothesis because they do not provide information concerning the type of processing carried out. The results present problems for current models of hemisphere function. The simple structural model [12] predicts that hetnifield asymmetries are a function of the appropriateness of the stimulus for the hemisphere of entry and that verbal tasks should be performed better when the stimuli are presented to the RVF-left hemisphere as opposed to the LVF-right hemisphere pathway. This type of model ignores the fexibility of cortical processes and cannot effectively explain changes in hemispheric differences over time, for example. The interaction between type of judgment and visual field obtained in the present study is also outside the scope of such a model. That is, the general description of stimuli as being “visual” or “verbal” is shown to be of limited practical and theoretical utility since the pattern of asymmetries obtained could not be predicted from a model relying on simple dichotomies based on fixed pathways. While the structural model has been usefully extended with an emphasis on the analytic- wholistic processing distinction between the left and right hemispheres, respectively, the important point to emerge from the interaction found in this study is that the different demands made of the subject by a particular task can modify the processing required of the same physical stimuli within a single block of trials. It might also be argued that the interaction between visual field and judgment is not consistent with KINSBOURNE’S [ 131 attentional hypothesis that attributes lateral asymmetries to the relative activation of the hemisphere specialized for the type of load imposed by the task and the consequent inhibitory influence of this activated hemisphere on the performance of

CODING STRArEGIESANL)HEMISPHERICDIFFERENCES

613

other hemisphere. Since the stimuli are letters and the instructions were delivered orally this model might predict left hemisphere activation in the subjects leading to a RVF advantage for both types of judgments. This experiment was not designed to test Kinsbourne’s model but there is no indication in the data of prestimulus, set effects. The problem with these speculations is that they are necessarily post hoc since the interaction came out contrary to prediction. There is an additional problem concerning the different results obtained with mean and median RT analyses. This is an interesting, if slightly disconcerting, outcome since it could imply that use of one or the other metric of performance in similar studies may lead to increased chances of making Type 1 and Type 2 errors, respectively. This is clearly an issue for empirical study. In order to overcome the problems with interpretation it will be necessary to use tasks about which we have reliable information regarding the processing required for successful performance and, perhaps, integrate these with measurement of electro-cortical activity from appropriate regions of each hemisphere [ 143. While this study has succeeded in its primary aim of changing the pattern of lateral asymmetries normally associated with letter stimuli, by manipulation of the judgment criteria to be used by the subject it has not succeeded in identifying the underlying processes: whether serial vs parallel or analytic vs wholistic. Finally, there are two additional flaws associated with experiments such as the one described. First, current models of hemispheric specialization do not generate testable predictions except in the most simple experiments. Second, and very much related to the first, is the general lack of attention to and manipulation of the cognitive strategies adopted by the subjects. It has been pointed out [IS] that a subject’s performance is a joint function of the task and the way the subject chooses to approach it. If no knowledge of, or control over subjective strategies is available then problems of interpretation will always remain when group data are discussed. The influence of the subject is also of theoretical concern because of claims that certain individual differences in performance may themselves be based on hemisphere differences in “cognitive style” [16, 173. It is vital that the whole context of the experiment is carefully analysed if advances in the understanding of underlying mechanisms of cerebral asymmetry are to be made from the study of intact humans. the

REFERENCES SI:A~~ON..I. G. and G~zr,i~tc;n. M. S. Coding strategies and cerebral laterahty. Cog. Psyhol. 5, 249-256, 1973. J. G. Coding and retrieval processes and the hemispheres of the bram. In Hrtnisphcrc Funcrion in rhc, Hurwrl Brain. S. J. DlhlOM~ and J. G. BFAL~MON~(Edrtors). Elek Science, London, 1974. NI~IIXRHUHL.J. and SPRIWEK, S. P. Task requirements and hemispheric asymmetry for the processing of single letters. Nezrrops?‘c,hol~(lic 17, 689 692, 1919. MARTIN. C. M. Verbal and spatial encoding of visual stimuli: the effects of sex. hemisphere and yes no judgments. Car/c,.\- 14, 227 233, 197X. BAMREK,I>. Reaction times and error rates for “same” and “dtffcrent” judgments of multidimensional stmluh. Prrwp,. P.sn~hoph~x 6, 169 174. 1969. BEV~R, T. G. Cerebral asymmetries in humans are due to the drfferentiatron of two mcompatible processes: holistic and analytic. Ann. N. Y. Acud. Sci. 263, 251 262. 1975. BRAIISHAW.J. L.. GATFS. A. and PATTFRWN. K. Hemispheric differences in processing visual patterns. Q. J/c\~I. P.vrchol. 28, 667 682. 1976. PA~TFRSOY. K. and BRADSHAW. J. L. Differential hemtspheric medtatton of nonverbal visual stimuli. J. tsp. Ps~d~ol. Hwr. Prrwp~. Pcrfim?. I, 246 -1.52. 1975. WHITL, K. and ASH~OIL’.R. Handedness assessment inventory ,Yrlrr~,/,.?!.~,/lr,loclin14, 26 I 264. 1976. KEPPM.. G. /ksi~~n mcl Amr/~si.s. .4 Rc.vwwhrr’s Ifudbook-. Prentice Hall. Englewood Cliffs. NJ. 1973. SAMOU,

614

JOHN EL)WARDSand PETFR H. VFNAIXFS

I I. DAVIS, R. and SCHMIT, V. Visual Psychol.

37, 229-240.

and verbal

coding

in the interhemispheric

transfer

of information.

Acta

1913.

12. 13.

KIMURA, D. Dual functional asymmetry of the brain in visual perception. Neurops&ologicr 4, 275-285, 1966. KINSBOURNE, M. The cerebral basis of lateral asymmetries in attention. Acta Psychol. 33, 193-201, 1970. 14. DONCHIN. E., KUTAS, M. and MCCARTHY. G. Electrocortical indices of hemispheric utllisation. In Luteralisarion in rlze Ner~,ou.s Sy~rem, S. HARNAD, R. W. DOTY, L. GOLDSTEIN. J. JAYNESand G. KRAU~FIAMFR (Editors), Academic Press, New York, 1977. IS. BRYDEN, M. P. Strategy effects in the assessment of hemispheric asymmetry. In Srrutryie.~ of fnfi,rmcrriorl Processing, G. UNDERW~~II (Editor). Academic Press, London. 1978. 16. GALIN, D. and ORNSTEIN. R. Individual differences in cognitive style~~l. Reflective eye movements. Neuropsycholoyia 17.

DoYLF.J.C.,

12, 367-~376,

1974.

ORNSTEIU,R.~~~GAI.IN.

Psychophysiology

567-578,

D.

Lateralspecialisationofcognitive

mode~~~Il.EEGfrequencyanalysis.

1974.

la consigne dans la production L'importance de d'asym@tries entre les hemichamos visuels a Bt6 inventori6e en manibulant les critkes qui permettent H un sujet d'apparier un stimulus central et une lettre orkent6e dans WI hgmichamo. Les caractkistiques physiques des lekres (faites soit uniqueknt de segments de' droite;, F, ou comportant au moins un segment courbe, P) ont 6t6 arises en comote. IJne interaction du cBt.6 du chamo stimulk et du tvoe he iuaement (aooari6 ou non-apparii) a visuel gte trouvee en testant-'le tekp; de r6poke manuel. Les'reponses d’appariement ont 6t6 p111s rapides pour le champ visuel droit que ooor le chamo visuel aauche. L'inverse a 6t6 observ6 pour les r6ponses de nbn-appariement. La signification de cette in&action pour la comprPhension de l'asyktrie entre les h6misphGres est discut6e.

Zusammenfassune: Die

Bedeutung

hllften ein

wurde

zentrales

wurden. z.

B.

der

Aufgabenstellung

durch

Manipulation

Objekt

und em fmse~tiger

Die geometrischen gerade

Linien

Wechselwrkung fiir

eine

Objekten schneller dieser diskutiert.

Kurven,

auf Objekte Wechselwirkung

als

Zielbuchstabe

Gesichtsfcld Da

im linken

im linken fiir

als

Modelle

in beiden

untersucht,

der

rinander

I’.

C,rsichtsfrld, urn rechten der

gelang

welchen zugeordnet

wurden

Es fnnd

und ZuverEisslgkcit Zuordnen

Gesichtsfeld-

nach

Buchstaben

wle m F gegen

Heaktmnszeit.

im rechten

Asymmetrien

Kriterien

Charakterlstika

oder

zwischen

manuelle

fiir der

des schneller

IJrtells bfl

und Fehlzuordnung Gesichtsfeld.

hemisphSrischen

betont,

such eme

erfolgte

Die Iiedeutung Asymmetric

wrd