-14175 DEFICITS IN HUMAN VISUAL SPATIAL ATTENTION FOLLOBNN 1/1

THALAHIC LESIONS(U) WASHINGTON UNIV ST LOUIS NO

I1 R D FAFAL ET AL 28 JUL 87 TR-87-4-ONR N88814-86-K-8289

UNCLASSIFIED FIG 5/8A N

E/Inh///////EI|iEEEEE

~ 140

MICROCOPY RESOLUTION TEST CHARTNArIUNAL BUREAu nFl STANDARPS 1963-A

-av

% %

00

0 DEFICITS IN HUMAN VISUAL SPATIAL ATTENTIONFOLLOWING THALAMIC LESIONS

Robert D. RafalRoger Williams General Hospital and

Brown University Program in Medicine, Providence

and

Michael I. PosnerMcDonnell Center for Studies of Higher Brain Function

Washington University, St. Louis

pl. ELECTL

ONR 87-4 UG25

Research sponsored by: Personnel and Training Research Program,Psychological Sciences Division,Office of Naval Research

Under Contract Number: N0014-86-K-0289Contract Authority Number: NR-442a554

Reproduction in whole or part is permitted for any

purpose of the United States Government.

F7 "2.' 01

8.25 0.9

SECURITY CLASS;F:CATION OF THIS PAGE

REPORT DOCUMENTATION PAGEla. REPORT SECURITY CLASSIFICATION 1b. RESTRICTIVE MARKINGS

Unclassified2a. SECURITY CLASSiFiCATION AUTHORITY 3. DISTRIBUTION/ AVAILABILITY OF REPORT

Approved for public release:2b. DECLASSIFICArION, DOWNGRADING SCHEDULE distribution unlimited.

4. PERFORMING ORGANIZATION REPORT NUMBER(S) S. MONITORING ORGANIZATION REPORT NUMBER(S)

Technical Report #87-4

6a. NAME OF PERFORMING ORGANIZAT!ON 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION

Washington University (if applicable) Personnel & Training Research ProgramsI Office of Naval Research (CODE 1142PT)

6c. ADDRESS 'City, State, and ZIPCode) 7b. ADDRESS (City, State, and ZIP Code)

660 S. Euclid, Box 8111 800 North Quincy StreetDepartment of Neurology Arlington, VA 22217St. Louis, MO 63110

Ba. NAME OF FUNDING /SPONSORING 8b. OFFICE SYMBOL 9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION I (it applicable) N00014-86-K0289

Sc. ADDRESS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS

PROGRAM PROJECT TASK 1W0RK UNITELEMENT NO. NO NOI ACCESSION NO

__61153N IRR04206 1RR04206-OA NR442a55411 TITLE (Include Securyt Clazfication)

Deficits in Human Visual Spatial Attention Following Thalamic Lesions

12 PERSONAL AUTHOR(S)Rafal, Robert D. and Posner, Michael I.

13a. TYPE OF REPORT 13b. TIME COVERED 14. DATE OF REPORT (Year. &ontfi. Day) 1S. PAG ouI4TTechnical FROM 01MAY87 TO 01MAY881 July 20, 1987

16. SUPPLEMENTARY NOTATION

,7 COSATI COOES 18 SUBLECT TERMS (Continue on reverse if necessary and identify by block nunber)

FIELD GROUP SU8-GROUP Selective attention, visual attention, thalamic lesions05 i0

19. ABSTRACT (Contnue on reverse ,f neceisary and denrtuy by block numbet'ihere has been recent speculation

concerning the role that thalamic nuclei play in directing attention to locations in visualI, space. We measured covert shifts of visual attention in three patients with unilateral

thalamic hemorrhages both shortly after the lesion and after a six month delay. The experi-ment measured reaction time to targets that occurred at locations to which attention hadbeen previously cued (valid trials) or at a currently unattended location (invalid trials).Although the patients showed no deficits in visual fields with perimetry and no neglectin the six month followup, we found slow reaction times for targets on the side contra-lateral to the lesion whether or not attention had been cued to that location. Deficitshave also been found in this task with cortical and midbrain lesions, but the patterns of

*. performance are quite different. The results with thalamic patients suggest they have aspecific deficit in the ability to use attention to improve the efficiency of processingvisual targets contralateral to the lesion (engage operation). This finding is in accordwith ideas of a thalamic link between cortical visual attention and pattern recognition syste..0 STR.3UT1ON,A4AiLAILT7Y OF ABSTRACT 21 ABSTRACT SEC'.RITY CLASS;FCA7iCN

JUNCLASSIFIEDJNLIMITED M SAME AS RPT rl DTC USERS Unclassified22& NAME O 4ESPCNSIBLE .NDIVIOUAL 22b. 'E.EPHONE (Include Area Code) I 21c FCE SrMBOL

FOMihg IPosner (314) 362-3317 ONR 1142 PT% 00 FORM 1473, 34 MAR 83 APR eition may :e wsea untiI eUtea SECURITY CLASSiFiCAT7ON OF 4HIS PAGE,V.' All other eoitions are obsolete.

'S.4 ,"% ,, "% . ,, ".. , -. - . ° . - .',-, -- . . . , . - . - . --. -, -. -. - . -. . . . ., -- . . . . % - -

-N. # - ." .. ' . . . ., ' -. '..-.'.. .' . '..,...'''., . '' . .• • ., .. €

Deficits in Human Visual Spatial Attention

Following Thalamic Lesions*

Robert D. Rafalt

Division of Neurology

Roger Williams General Hospital and

Brown University Program In Medicine

Providence, RI 02902

Michael I. Posner

McDonnell Center for Studies of Higher Brain Function and

Departments of Neurology, Neurological Surgery and Psychology

Washington University

St. Louis, MO 63110

(Selective attention / visual attention / thalamic lesions)

....................

I T

',

.5.. N

ABSTRACT

There has been recent speculation concerning the role thatthalamic nuclei play in directing attention to locations in visualspace (Crick, F., 1984 Proc. Nat. Acad. Sci. USA 81, 4586-4590). Wemeasured covert shifts of visual attention in three patients withunilateral thalamic hemorrhages both shortly after the lesion and aftera six month delay. The experiment measured reaction time to targetsthat occurred at locations to which attention had been previously cued(valid trials) or at a currently unattended location (invalid trials).Although the patients showed no deficits in visual fields withperimetry and no neglect in the six month followup, we found slowreaction times for targets on the side contralateral to the lesionwhether or not attention had been cued to that location. Deficits havealso been found in this task with cottical and midbrain lesions, butthe patterns of performance are quite different. The results withthalamic patients suggest they have a specific deficit in the abilityto use attention to improve the efficiency of processing visual targetscontralateral to the lesion (engage operation). This finding is inaccord with ideas of a thalamic link between cortical visual attentionand pattern recognition systems (Crick, 1984).

V-

.- " -.. %'

INTRODUCTION

In recent years a number of specific experimental methods havebeen used with alert monkeys (1-3) and humans (4-6) that time lockcovert shifts of attention to the presentation of cues. Inneurophysiological studies the orienting of attention is inferred fromselective enhancement in neuron firing rate in response to the cue.Cognitive studies measure the allocation of attention in terms ofimproved efficiency in responding to signals at the cued locations incomparison to other spatial locations. These approaches have begun toconverge to identify the neural mechanisms controlling visualattention. Cognitive studies with normal humans using visual cues todirect attention covertly to a location eccentric from fixation showmore efficient processing of signals at the cued location. Thisenhancement includes lowered manual (5) and saccadic (7) reactiontimes, reduced sensory thresholds (8), improvement in conjoiningfeatures (9) and modulation of evoked electrical potentials recordedfrom the scalp (10). These observations support the concept ofattention as a mechanism for relative enhancement of informationprocessing at a selected spatial location. There is also evidence thatthe area of enhancement becomes larger as cues are presented moreeccentrically in correspondence with the known characteristics of theneural magnification factor (11,12).

Areas of the monkey brain showing selective neuronal enhancementinclude the posterior parietal lobe (1,2), the superior colliculus (2)and substantia nigra (pr) (13) of the midbrain, and the lateralpulvinar (14). The same visual cueing method described above was usedto demonstrate that modulation of GABAergic transmission in thalamus(with iontophoretic injections of muscimol or bicuculline)systematically affect the orienting of attention contralaterally (15).Reaction time studies using cueing in neurologic patients haveconfirmed that lesions of the parietal lobe (16) and peritectal regionsof the midbrain (17) produce distinctly different deficits in orientingvisual attention.

Three computations have been suggested in orienting of visualattention. First attention must "disengage" from the current location;then "move" to a new location; then "engage" at the new location.Deficits in each of these three elementary operations can be identifiedin cueing studies. At the beginning of the trial the subject ismaintaining fixation at the center of the display without activelyattending to any spatial position (no targets occur at the center).When the cue is presented the subject must move attention to the cued

location and engage attention there in anticipation of the forthcoming

i.

%'3

target. The efficiency of moving attention can be inferred, then, fromthe rate of improvement of RT with cue-to-target delay on valid trials.A deficit in the move operation can be inferred by a deficiency (i.e. adelay or reduction) in this improvement.

A deficit in the move operation has been found in patients with'rogressive supranuclear palsy who have degenezation of the superior

lliculus and peritectal region (19,20). In these patients saccadiceye movements are relatively more impaired in the vertical dimensionthan are horizontal eye movements. We therefore compared vertical andhorizontal attention shifts. RT on valid trials improved more slowlywith time following the cue in the vertical dimension.

A different pattern of results was shown for patients withparietal lesions (16). Reaction times improved at the same rate inboth visual fields following a valid cue. This indicates that parietallesions do not slow the movement of attention toward the contralateralfield. Moreover, the asymptote of these functions differed very littlebetween fields showing that the ability to use attention to engage thetarget location did not differ greatly between visual fields. Incontrast to the midbrain patients there was a dramatic increase in RTsto targets in the contralateral field following invalid cues.According to our scheme, if attention is shifted to the cue but thetarget appears elsewhere, it is necessary to disengage attention fromthe cue before moving to the target. The selective slowing ofdetection RT in the invalid cue condition suggests, therefore, that theparietal lobe plays a special role in mediating the disengageoperation.

Parietal lesions and midbrain lesions have distinctly differenteffects on orienting attention: midbrain lesions appear to produce aspecific deficit in the move operation; whereas parietal lesionsselectively appear to produce a specific deficit in the disengageoperation.

We now extend the use of cueing paradigms to measure attentionshifts in three neurological patients with thalamic hemorrhages. Thismethod permits us to compare the thalamic deficit with those found inmidbrain and parietal patients. Lesions of any of these areas canproduce clinical symptoms of neglect of contralateral stimuli (19).However, the computations performed by these areas may be quitedifferent. If the patterns of performance deficit due to lesions ofthese areas differ, it should be possible to further the analysis ofthe role of each area.

S 4

......... j.j ..

METHODS

Subjects

Three patients with hemorrhages in the thalamus were studied in anexperiment to measure covert shifts of visual attention on twooccasions: in the acute stage they were tested as soon as they wereable to perform the task; each was retested after 4-6 months ofrecovery (chronic stage). Patient VM, a 65 year old man, had a largehemorrhage centered in the left thalamus with rupture of the hemorrhageinto the ventricular system. He was initially comatose with righthemiplegia, hemianesthesisa and ocular deviation. Fig. 1 shows the CTscan findings at the time of initial testing seven weeks after theictus. At that time he still manifested some psychomotor retardationand mild visual neglect. At the time of retesting six months after theictus, he was alert, lucid and subtle visual neglect was evident onlyon a letter cancellation task. The other two patients had smallerlesions which did not impair alertness, and were first tested in thesecond week of their illness. Patient VL, a 67 year old woman, had ahematoma in the right thalamus (Fig. 2). Patient NA, a 54 year oldman, had a small hematoma in the right thalamus involving the nucleicentromedianum, ventrolateral and lateral posterior (Fig. 3A). Thehemorrhage extended into the posterior limb of the internal capsule,and ventral to the thalamus into the region of the zona incerta andperigeniculate region (Fig. 3B).1 Patients VA and NL had hemiparesisand hemisensory impairment contralateral to their lesions. Neither hadany signs of visual neglect on detailed clinical testing. At the timeof follow up testing 4-6 months after their strokes, perimetry testingconfirmed that the visual fields were intact in all three patients.

Procedure

Subjects sat facing a video display screen with one finger of thepreferred hand on a response key placed on a table between the subjectand the display. Light pressure on the key activated a microswitchwhich recorded RT. The display consisted of a + sign at the center,flanked five degrees to left and right by one-degree unfilled squares.Subjects were instructed to maintain gaze on a plus sign in the middleof the screen and not to move the eyes. Eye position was monitoredwith a closed circuit video camera to assure that the eyes remainedfixed at the center. Subjects practiced the task before data wascollected while the experimenter observed to ascertain that thedirections were understood and the subject was not moving the eyes.The intertrial interval was 2 seconds. At the start of each trial thefixation point was extinguished and .5 second later the cue was

5

1%

presented by brightening, randomly and with equal probability, one ofthe two peripheral boxes. The cue remained visible for 300 msec.After an interval (50, 150, 500 or 1,000 msec) following the onset ofthe cue, a target appeared either at the cued location or in theopposite visual field. Subjects were instructed to press the responsekey as quickly as possible any time the target (a bright asteriskfilling one of the peripheral boxes) appeared. The target remainedvisible until the subject responded (or for 5,000 msec). In thisexperiment, the target was on the cued side on 80% of trials (validtrials), while on 20% of trials, the target appeared in the boxcontralateral to the cue. The probabilities were designed to inducethe shift and maintenance of attention to the cued location. Since theeyes remained fixed at the center, and since the motor response (asimple key press) was always the same, any difference of RT betweenvalid and invalid cue conditions may be assumed to index a covertmovement of attention to the cued location.

RESULTS

We first excluded all RTs less than 100 or greater than 4,000millisec. Only a very few times were affected by this rule. Themedian RT for each patient in each condition was calculated.

A within factor analysis of variance was run with the followingfactors: stage of illness (acute vs. six month followup); target field(contralateral to lesion vs. ipsilateral to lesion); cue validity(target appeared at cued location (valid) versus at uncued location(invalid); and cue to target interval (50, 150, 550 or 1000 msec.).

When tested in the chronic stage (six months or more after thelesion) the patients were faster than in the acute stage but this didnot reach statistical significance (F[1,21 = 2.65). Thus we displaythe combined data for acute and chronic tests in Fig. 4.

Reaction times are faster in the ipsilateral field than in thecontralateral field for both validity conditions (F[1,21 = 36.8,p<.0 25 ). Validity has a significant effect with valid targets (solidlines) responded to faster than invalid targets (dashed lines) F[I,2J =23, p<.05) and validity interacts with interval such that its effectsare greater at short cue to target intervals (F[3,61=8, p<.025).Finally, this interaction of validity and interval is significantlygreater in the contralateral visual field than in the ipsilateral fieldresulting in a triple order interaction between validity, field andinterval (F[3,61 = 11, p<.Ol).

6

1

These results would be consistent with a primary visual defect inour patients. However, our thalamic patients had no clinical evidenceof visual impairment and, as mentioned, clinical neglect was notconspicuous (and was totally absent in two of the patients). All threepatients showed no contralateral visual field defect on formalperimetric examination, even with the smallest (3mm) target. Since thetarget in our experiment was a large (1 degree), bright signalpresented in the parafoveal (5 degree eccentricity) region, it seemsvery unlikely that a subtle visual field defect, beyond the sensitivityof perimeteric testing, could have accounted for the dramatic slowingof contralesional detection RT. A fourth patient with a posteriorcerebral artery stroke syndrome and CT evidence of infarction in theright thalamus and occipital lobe was also tested. He had a densehomonymous hemianopia, and could not respond to any signal presented inthis contralesional visual field. He was tested in an experiment whereall cues and targets were presented in his intact visual fieldipsilateral to the lesion (20). The target was presented at the samelocation on each trial, but was preceded by a cue which first summonedattention either to the left or right of the forthcoming target. Oneach trial, then, he had to disengage his attention to move it ineither an ipsilesional or contralesional direction. When he had toshift attention leftward (contralesionally) detection RTs weresystematically longer than when he had to shift attention rightward(ipsilesionally). This result, obtained entirely within the intactvisual field, could not have been due to differences in visualsensitivity since the target always occurred at the same location.

DISCUSSION

There are three salient features of the data depicted in Fig. 4.1) For the valid trials, the cue produces a similar improvement in RT,aos a function of cue-target interval, in both visual fields. 2) Forthe invalid trials, there are slow RTs in the contralesional field forthe short cue-target intervals. 3) There is a dramatic main effect ofvisual field, with mean RT to contralesional targets beingsubstantially slower. Consideration of these three findings incomparison to previous findings for patients with midbrain and parietallobe lesions, provides insights into the role of the thalamus in adistributed neural system for orienting visual attention.

IA Inspection of the data from the valid cue condition (solid lines)reveals a decrease in RT with interval. Although RT is slower for allcontralesional targets, the improvement in RT from valid cues over timeis equivalent in the two hemifields. This pattern for valid cue trialsdiffers from what we have found previously in patients with midbrain

7

'4|

- . - -- .. '~. ,'\ {

lesions in whom we have argued for a disorder in the "move" operation.In midbrain patients the improvement of RT on valid trials was slowerin the affected direction (vertical). Thus, midbrain patients wereslow in moving attention. In contrast, for the thalamic subjects RT tovalid trials improves following the cue with a similar time course inboth visual fields. In contrast to the midbrain lesion patients, theydo not appear to have a deficit in moving attention in response tocues.

The second feature of these results are the long RTs on theinvalid trials relative to valid trials in the contralesional field forthe short cue-target intervals. This pattern is similar to that foundin our parietal patients, and suggests that thalamic lesions do affectthe disengage operation in a qualitatively similar way. Indeed, themean reaction time to invalidly cued contralesional targets for theseearly cue-target intervals in the thalamic lesion patients is similarto that previously for patients with right parietal lesions.Nevertheless, the relative slowing on invalid trials when compared tovalidly cued targets in the same field is much less in the thalamicpatients. Moreover, the disengage deficit in the parietal lesionpatients persisted even through the longest (1000 msec) cue-targetinterval. In the thalamic lesion patients, the disengage deficit ismanifest only at the early cue-target intervals, while the cue is stillpresent. We conclude that, although intact thalamic function may benecessary for disengaging attention, the parietal lobe is chieflyresponsible for this operation. The thalamic lesion may have anindirect affect on parietal function to produce the "disengage"deficit.

In spite of their apparent ability to move their attention inresponse to the cue, the third and most striking aspect of the data isthe persisting main effect of visual field for both valid and invalidtargets. Even at the 1000 msec cue-target interval, when attention hashad time to reach the target location, RT to detect contralesionaltargets remains slower, and at no time is this difference less than 200msec. This difference between the two visual fields is about fourtimes as long as the mean difference which we found for parietal lesionpatients (16). Only one of those thirteen parietal patients showed aRT for validly cued contralateral targets at the 1000 msec cue-targetinterval which was as long as the mean for the three thalamic patients.The different pattern of results for the valid cue condition for thethalamic lesion patients, in comparison with that seen with midbrain orparietal lesions, is consistent with a deficit in the engage operation.

S 8

The different pattern of experimental results between parietal andthalamic lesion patients is especially interesting when one considersthat, even though the thalamic patients had much slower contralesionaldetection RTs than parietal patients in the valid cue condition, mostof the parietal lesion patients had more clinical neglect than did anyof the thalamic lesion subjects. The fact that these patients showless clinical neglect than do parietal lesion patients, whose deficitlies in the disengage operation, leads us to speculate that clinicalneglect, an important source of disability, can be linked most directlyto a disorder in the disengage operation.

It is not possible to make precise inferences about the specificneural structure responsible for the effects found in our patients. Intwo patients the hemorrhage involved large parts of the thalamus,including the pulvinar, as well as adjacent structures. The mostrestricted lesion was present in patient NA, who also had the leastsevere clinical impairment. Since this patient had the same pattern ofresults as the thalamic group as a whole, both at acute and chronictesting, the anatomic localization of his lesion on CT (Fig. 3)provides the best information bearing on this question. The lesioninvolves the nuclei lateral posterior, centromedianum andventrolateral. Unlike the other two patients, it does not clearlyinvolve the pulvinar (Fig. 3A). It extends ventral to the thalamus andinvolves the perigeniculate region (Fig. 3B).

This area may correspond to the region of the perigeniculatenucleus considered by Crick as possibly mediating the "searchlight" ofvisual attention. This structure, related to the thalamic reticularnuclei, sends GABAergic projections to the dorsal thalamic nuclei whichmay gate their processing of sensory information. Petersen et al (15),using the experimental task described here in monkeys, have shown thatmanipulation of GABAergic transmission to pulvinar, with iontophoreticinjections of muscimol or bicuculline, systematically affect theorienting of attention contralaterally. It would be of interest tocompare, in experimental animals, the effects of discrete lesions ofpulvinar and of the thalamic reticular region, in this task.

Recent PET scan studies in patients with thalamic lesions showthat these lesions produce diffuse hypometabolism throughout theipsilateral hemisphere (21). These results suggest that the thalamusis involved in cortical "activation" in some way. Whether suchactivation can be interpreted in terms of a defect in attention, in thesense applied in this communication, remains conjectural. Thehypometabolism (21) was most pronounced in the acute phase, and haddiminished substantially within 4-6 months.

9

According to current neurobiological views, the visual cortexinvolves somewhat separate areas for signal localization and directingof visual attention (parietal) than for pattern recognition (occipito-temporal) (22). We have shown that patients with parietal lesions havedefects in pattern recognition on the side contralateral to the lesion(23). This suggests that the ability to recognize patterns rests inpart upon an intact visual attention system. The route by which theparietal system interacts with the pattern recognition system is notknown. The current results agree with the ideas of others thatthalamic nuclei may play a role in this interaction (3,24). Moreover,it suggests that the thalamic effects on attention are not due toremote effects on cortical or midbrain areas alone. Our evidence isthat thalamic lesions produce a different pattern of deficit than foundfor midbrain or cortical lesions. Thus the computations performed bythalamic structures are distinct and do not appear to be an indirectreflection of damage elsewhere. Even closer contact between humanstudies and alert monkey studies should be useful in developing a morecomplete model of how these neural systems interact in orchestrating ashift of visual attention.

10

;

FIGURE LEGENDS

FIG 1: CT scan from patient VM at the time of acute phase testing seven weeks

after his stroke. There is a resolving large hematoma (arrow) centered in the

left pulvinar.

FIG 2: CT scan from patient VL at the time of her stroke showing a large

hematoma in the posterior right thalamus.

FIG. 3: CT scan from patient NA at the time of his stroke. A. There is a

small hematoma in the right thalamus centered in the ventrolateral nucleus andinvolving nuclei lateral posterior and centromedianum. B. The hematomaextends into the posterior limb of the interanal capsule, and ventral to thethalamus into the area of the zona incerta and into the perigeniculate region(arrow).

FIG. 4: Mean RT for three thalamic patients as a function of cue-target

interval. Stimulus onset asynchrony between cue and target in millisec.

AN

N. L

11

FOOTNOTES

* This research was supported in part by ONR Contract N-0014-86-0289.

Reprint requests should be addressed to:

Robert D. Rafal

Roger Williams HospitalDepartment of Neurology

825 Chalkstone Avenue

Providence, RI 02902

± Localization of the lesions was determined by relating the CT findings

to De Armond SJ, Fusco MM and Dewey MM: Structure of the Human Brain:A Photographic Atlas (Second Edition) New York: Oxford UniversityPress, 1976.

PIZ2

REFERENCES

1. Mountcastle, V.B. (1978) J. R. Soc. Med. 71, 14-28.

2. Wurtz, R.H., Goldberg, M.E., & Robinson, D.L. (1980) Prog. in Physiol.Psych. & Psychobiol. 9, 43-83

3. Moran, J. & Desimone, R. (1985) Science, 229, 782-784.

4. Treisman, A.M. & Gelade, G. (1980) Cog. Psych. 12, 97-136.

5. Posner, M.I. (1980) Quart. J. Exp. Psych. 32, 3-25.

6. Eriksen, C.W. & Hoffman, J.E. (1973) Percept. & Psycho. Phys. 14,155-160.

7. Fischer, B. & Breitmeyer, B. (1987) Neuropsychologia 25, 73-83.

8. Bashinski, H.S. & Bachrach, V.R. (1980) Percept. & Psycho. Phys. 28,241-248.

9. Prinzmetal, W., Presti, D. & Posner, M.I. (1986) J. Exp. Psychol. 12,I. 361-369.

10. Mangun, G.R., Hansen, J.C. & Hillyard, S.A. (1986) ONR TechnicalReport SDEPL 001. Proceedings of the Eighth InternationalConference on Event Related Potentials, Stanford, CA. June.

11. Downing, C.J. & Pinker, S. (1985) in Attention and Performance XI ed.Posner, M.I. & Marin, O.S.M. Erlbaum: Hillsdale, N.J. 171-187

12. Sagi, D. & Julesz, B. (1986) Nature, 321, 693-694.

13. Hikosaka, 0., & Wurtz, R.H. (1985) J. Neurophys. 53, No. 1, 292-308.

14. Petersen, S.E., Robinson, D.L. & Keys. (1985) J. Neurophys. 54,867-886.

15. Petersen, S.E., Lee D., Robinson, D.L. & Morris, J.D. (1987)Neuropsychologia 25, 97-105.

16. Posner, M.I., Walker, J.A., Friedrich, F.J. & Rafal, R.D. (1984) J.Neurosci. 1984, 4:1863-1874.

13

aiz

i

17. Posner, M.I., Choate, L., Rafal, R.D., & Vaughan J. (1985) Cog.Neuropsych. 2:211-228.

18. Rafal, R.D. & Inhoff, A.W. (1985) Proceedings of the Eighth Meeting ofthe Cognitive Science Society, Amherst, Mass. Hillsdale, NJ:LawrenceErlbaum Assoc., p. 260-271. (abstr.)

19. Mesulam, M.M. (1981) Ann. of Neurol. 10, 309-325.

1 20. Posner, M.I., Walker, J.A., Friedrich, F.J. & Rafal, R.D.Neuropsychologia, 1987, 25, 135-146.

21. Baron, J.C., D'Antona, R., Pantano, P., Serdaru, M., Samson, Y., &Bousser, M.G. (1986) Brain, 109, 1243-1259.

22. Mishkin, M., Ungerleider, L.G. & Macko, K.A. (1983) TINS 6, 414-417.

23. Friedrich, F.J., Walker, J.A., & Posner, M.I. (1985) Cog. Neuropsych.2, 250-264.

24. Mountcastle, V.B., Motter, B.C., Steinmetz, M.A. & Sestokas (in press)J. Neurosci.

14

U ' '/ '. , .- ",,'" " o . ", .'"% . .v ". '" .- '- -. '

FIG 1

0of

04C.

4-4-.w- 2 4.A ... - - .A.~A.'

* PIGZ

6

d~.

S

~ *.*.v*.. -- ...~ ..-. -. *

F G3

FIG 36

Figure 4

1100

1000 -

900\

U.W 800

CONTRA

~- 700-600 -0',-"

600

5-0 -4 Valid

S D'0 Invalid

200 400 600 800 1,000Cue to Target Interval

C

a C.NC

X . a N cr - 0 -

a- A c E-

~ U"C Is - .

C 0. .. -, - 0-9* f. C. C. r 6

*o * o 6 cc "C w 3. J .

o0 of 0

0 cc

-30 a W6a 2C -

c c

C 0~ 0 0S. a m C C. C

-- 0 LC -6 0

cS Lc a w . 6 .1.

IS 0 0 S9 -~ 0 .aa

Cc~ O S S

- 16 a6 U x

OWN a~ aLS axz a US *U~..Z 4.S.. - cSC' ~~0 IC C B a. UE OB

0-

0.00 a-

1 -0 C n 2 -A

-a~~~ 0 S

CS 00 0J IVa0 c c 01 a,, C

-- ~ Js - I ~ S A. SL~~~~~9~~~ 0 4S 9C, US2cC4 @~ N S U0 U

II~ S C

I40 C U ao~ 0 V ~ CO - *.

-0 -0 0 0cv. * 05 0 S

O U .a C .5

t 00 a -4

-~~~~10 0. V .C Q =9o ~ ~~~ 2*V 2s ~*S

a. 1- 1U 0 1

LqO mc mc A . 0

C. ' n~ c c cc o ~ - U a. m .c~~~ at 4. c. * . u

0.4 16 acmo I~N c- . W o- 0L

MAou .U.a 8..n a a .00 . A

0

10 c

C4

u 0 0c 0

L 4.0 - 44 . C

40~ 0 r4 - 44. c. a r ml. n

4. c C4 c0. ceM

aO c a 0,1 0 4AO OV 0, - 1 &. "I2 c~ L, 0

4J*'D 4 0 z.. v 0.~m0 4.5

000 a~ C.- 0 e-o om 4.0 UO

V4.

"".0 C mq -n aE C EL L104

71-I

00.0 0 ID

6-cc C 0C

Ap a4a.

*N. Et.

~~O4.4. 4.NL 0L NN 4

2~~~~-- "Icep I U a " .k '.

L

c I ~ c - ~ 6 ~ ~ " 6

C UN

41 NN4 N.: ",

.c~ ~~~ c --. 0 4 - 4

.4~ ~ 0 CN. 04 C C N .N

2

-. - -00 -r-' 0 Zi

6: CA "I 0-i c.

fl-uc 6 0nfl a.0n 0n a- n. 11C,- .

-C O u 3~ ' . 01 0

i ..0 =IC 0 rO

Oll

6 Cc

00&

t =0 C4 0 046 c a WOU.C

", ' I M L. C40Z- -- C Z af C.30,-C

0 6 00 coa 0 a 0 0 0 360 . 0 - 6. 1, w a nL

a 22 1- 11 *1 0CCc c .3 fi ; 0

00.N * WD~ A 0 ; 0 aC 0 w - - . -,C h0 -C .flO CC0o0 0 1 c 6. .3 A~ -- -0 2- V0 1 c 0 fi L, u

CAZS *lIlo .2 ; C a tf .ZOL cc u- 5 a m2

AAX A ' SZC wC 20 k @0 r A l~ k00 w60- C09 hs000I.:

0W U C fl h U £ 6.* C.

0

C. I

C, 0 iC

acC 0. -- 0 --3- _0 00- z- 0" ';a:

"' CON Ccg

a - -C, 20 a ,- 0 U' -

S 3C 9-1 AC om 3c. *N : 3. 0~

C6 C? R '

aC u6 a-.I

6 1 6 0 ~~~ c 0 - -0l "0ff ~ C 3 U~O i 2~

*CP. wO - 1" xU - 0 &

-Cc -- 0* Z 0 C Ce*n

CC

w02 0 A -

a As' a C - 05

fi U.- C . 0

a 0% 140 0 0 0 00- 00Im I u -. 3 1 S. 0o A. -0 a ~ ** wt

2 ~ N 6 9! I-iN90o- - 1

6o 30C M O1:- ~ -00= O all~8 3 2 -W8 U 0

. 0 00 =0 00 -.b C- &-CC- ~ a W 0 0 AL 0 C f-

I-ow 0L a - S.L0 0%W . 1 E 6 -

omc0 61 WC * in

CC

00

0~ 00 .c

W- *0 Cv 6aC -c

-C N0 OX N. w. 0. 32.C-~ i 00. aU* 0C c4 CC~ --C 0.C CC U -12 a -; 00 0..u 0C -o01 wI c1 C 0CIn C. LI CC0C uC cI -- OX0mC 0 2 0 C o

A.4 0* *e 0 60 0 LI0 %Ja ~* 5-4 01.~~ 1.0 0,0 0. 0L ca C fo c Oa* 0s

Gn I~ 1^ C6 Ci. '.n "' 0 ;Z2C. i1CO 0 ~ 0 E ~

C)c

0 0

0D C0 MO4. CC LI -C 0-4

-0 0 a .j c

*a oc Co - . - .01. III -o 01 1".0

0A 0.C 0 C0 c a C 1.0 mc000

S. A WO C 10N C iX6~0. A .5-

c "C0. O0. OC c c . 0.C U C-- S ~ AsX NSo 2 I 0 JS am CO. Is. . L .4 is L, t1.0 . 0.LCC 0. 06

-' .- ; ~ o. 0£ ~ 0. 00

4, -C

13 IC I

0. 00 a c. 0. Ou c. c, 0 " 0 zJ C

~0 06 cL , C OCC0i aO x., 0' :

3F CC C c.~ N c C . zCC C-. -- 0 -. , ,- C *Z C 1Do.C .JC C. .. 5-.

-~ ~~~ _0 C ac. 0 4ft- .r. C cO IL ca .C Ca. 00. .0 .U314.. c . C C OL a -Ca '. ma0SC ..

.s-..boc VL0 0 5 ~ . C . ) O ~ C . C V4 .. U

P I O N. 5 4 0 6 ~ O C , ~ * V O L 4 ~ . 4 4 0 1l1ib.C 0 * ~ *' C 1f - ~ ~ l O C .. t 0 ~ 40

0. a0.-- C c C6 ,

0 0 50 0. 1

O 0 a 0~. o s

. .

LU N ciUr I I.1 1 i -. ~a .c 0 C =C.f'

C 0. 0% 0 0C C -450- 0 .. ~ 0

ft w - C . 00C.

C E OCN CW0 . c so a CC 0C1 . CC. .

1. ~ ~ ~ ~ - o0 1..3)1 *t411 . - CN C S a 10 fO a.

.00 00.Ow . N0i O0C~o 0 a.1 C. 0 0 NO a~~N x -

C C

22

a 0

S. C .Z t

-ff OC a 0c . C C. 0 0 0 .

0~, 5- 0- 50 -Sai CC 00 A 0 0C0L

MO I, A 1 2 .0

.A 41 So* a * 0 -A-4C '

-~G C000 0 W"' N O0 0 CO0 0 - 0 .cC

CULC. .iJ. I N 0 Uotn 0

0.~- aSI -0 -U0 a4 c..A . 0 a .

0a 0.is c b Ca * q 0 a .a CCo0 a O

00~~ a CO . 0a . .C C C a O. 4CO-C40 c a",

.6 0 aeL r. t - . CC. b o 0w 10 =C"

to1. -1 cN . 0C 1N K- 1, 0 Vn.0 c r20 0 C .

fi; .a ::1 :c 0 - - m. . 41.. 0 a

Ocz ~ ~ ~ ~ ~ 1c 001C .IU o

0 In

InI

C C 0. c-

0 NC 0 0 a c

1. 0. 0 0

a0 c : CCa Ca

00 a gI *e . - 0

2 - -cN 0 - *M c;;1 . -.-. O S.t . 1 0-.1 0 - 0CIVCo 0. IaN * N pNa C . 0 N .

J-N CO JN Ji 0. A5 00 C a. C 0

OOW

C S C

-C a,-it,6 v c

*~~~d a. a ft aa a .... my

.c 10S.2 cl. 04 C aU aI- .0 .U

0. WO SS -C c U 0

S~ US

.0N =1 CO.. S ICc !j bC Wf~ it u,: ICS0.nC

C

00

* 1 07- -

"0 c0- t 0 10C

5 -0 1s. 1. 9~ 2 . . -_ l; ;-

ON C i.- tZ0 GO 0 " 0 C I &~ I fiaO 'M 0 EE

1ia 0 1g s1 -C m 0U 1 O 0 V CC.5IC~~Z 0..C v

1 clt- IS C O . a U ' O o Ei

0..n c it LO0- CL- ffOI -- I0 0 L C .

I5 CL 0 0ii ,OO *UI *S. 0c ..0 x~ LU CID 00 C. S O La *6 1I Cm h -aU0cC

axo Oa O~ Z 12 t z 11~ cy A t O Ict

I-iC

a1 C

.a , - CL. 0 a. 0 L ow a. 0a w 0 -0

V)- L Si w U V) r a a 0 X

0 U0 00~ 1 ~ £, .

C.n r. a;0a l lVI)- W- aC.- C U

tfl~ ~ ait it I -0 (n- C i - 6. * 5 S . -i

0 U 0 a-t I 1ia .1cA a z3~ E a~ 1111> N ... u. 2 10 L N a a £ ) I , ~

m C o.C C ~ 53 Ii as. UC , -X s C0- 0- O aCIVI v~M re i Iac a . a - a a .&06 U -ca a a il~I *CI Ca .a0.,- Co I OIc i ' 5n a s q C 52Va-0O

C' o

Z: w C N 0

C a ik

--------------- M

.0 0 - 0

- - 0

Sn~C o CO *0 C nC 1

v C 1a 0" an S00 N N Uc N 0-O

CuC 04 C QM= .a3C~ ~~ ~~~~~ ~~~~ x f- ) n 3 c C s o t 3 4 - 0 s , 4 .

a4 0 5 0C L L S I. , . C 0 -ey- C 4 '0 C. . n

C C .C" a c..~ Uc .. J. 1 .4 U .0CU~~~~~ C t . . C l ~ S C .a - .' 05.

IV 04 5 0 04- ft.. CU .. 'O a. C. 3 t

CC

* 30. 1' C

c U0 c

0 0 - a C. '4S SO 0- 1a 3

*0 a5 - -* C4C nO~~~ ~~ C aC~ to 4. 4 - o O C J-~~ 100 - wo Ut. r n U - .Ca ., 2 - C6 g * C n 0 C

(- wl C 'm 2U Ns a4 S 3 9sS16 5 - - * - .

C0US 0 3 0 1 C 3 O U S 4 I

L. .4 5

CC. S . C R 4 3 4 .. 'L. V5 .'a, a.XS~~~ ~ ~ ~ ~~cZ 0. c * a - Ca . * .C U . C O QO .

ms a un . . c .0 .c C a o N 01 *

C r C 0-0 0- v CI5n ~ C .1 4 . 4

1 A 5.0C A- A a. I.N CSn. X6 =1~ 13 154 .. 0 2s WU = I -2' 0 1

Distribution List (Washington University/Posner) 1987/11/10

AFOSR, Life Sciences Division Dr. Jaime CarbonellUniversity of Minnesota Carnegie-Mellon UniversityDepartment of Psychology Department of PsychologyMinneapolis, MN 55455 Pittsburgh, PA 15213

Chief of Naval Education Dr. John J. Collinsand Training Director, Field Research Office,

Liaison Office OrlandAir Force Human Resource Lab. NPRDC Liaison OfficerOperations Training Division NTSC Orlando, FL 32813Williams AFB, AZ 85224

Captain P. Michael Curran Dr. Joel DavisOffice of Naval Research Office of Naval Research800 N. Quincy Street 800 North Quincy StreetCode 125 Code 1141NPArlington, VA 22217-5000 22217-5000

Defense Technical Information Ctr. ERIC Facility AcquisitionsCameron Statiopn, Bldg. 5 4833 Rugby AvenueAlexandria, VA 22314 Bethesda, MD 20014Attn: TC

Dr. Marshall J. Farr J. D. Fletcher2520 North Vernon Street 9931 Corsica StreetArlington, VA 22207 Vienna, VA 22180

Dr. John R. Frederiksen Dr. David NavonBolt Beranek & Newman Institute for Cognitive Science50 Moulton Street University of California

SCambridge, MA 02138 La Jolla, CA 92093

Dr. Donald A. Norman Dr. Robert SasmorInstitute for Cognitive Science Army Research InstituteUniversity of California 5001 Eisenhower AvenueLa Jolla, CA 92093 Alexandria, VA 22333

- Dr. Steven Zornetzer Dr. Michael J. ZydaOffice of Naval Research Naval Postgraduate SchoolCode 1140 Code 52CK800 N. Quincy St. Monterey, CA 93943-5100Arlington, VA 22217-5000

pp.

___________.-- Wflflwm~~,~

I.

'4

* p 0-p

'p.'4

'p.

4,M

4: -

4...

'44.

~R- 'if V U S S U S U if V U 0 V U'

-. .r~ ~ ~ ~ ~ .P~ . . *. . -- - -. - -, - .' -

- ~ ./.4. ~ *** * -.* -