Motion after-effects in cat striate cortex elicited by moving texture

V&on Res. Vol. 26. No. 7. pp. 1055-1060. Printed in Great Britam

0042-6989;86 Pergamon

1986

RESEARCH

53.00 + 0.00 Journals Ltd

NOTE

MOTION AFTER-EFFECTS IN CAT STRIATE CORTEX ELICITED BY MOVING TEXTURE P. HAMMOND,’ ‘Department England

of Communication and ?Department (Receired

G. S.

V. MOUAT’ and A. T. SMITHS

and Neuroscience, University of Keele, Keeie. Staffordshire ST5 5BG, of Psychology, University College Cardiff. Cardiff CFI IXL, Wales 18 Nocember

1985; in revised form

13 January

1986)

Abstract-Responses of striate cortical neurones to randomly-textured test patterns (static visual noise), or to bars of optimal orientation and width, moving back-and-forth with fixed velocity. were recorded in the lightly anaesthetized cat. Effects of prior adaptation with textured patterns drifting continuously in each cell’s preferred or null directions. were assessed. Alterations of directional bias and responsiveness to the test stimulus were assessed in relation to the degree and time-course of texture adaptation. The effects of adaptation to moving texture were qualitatively similar to our previously published data on adaptation to drifting gratings in the same or similar cells.Responses to the test stimulus were transiently depressed in the direction of adaptation and enhanced following adaptation in the opposite direction. compared with responses following exposure to stationary texture. However. even in cells that were driven strongly by the adapting texture, i.e. particularly the special complex cells in cortical layers 111 and V, the after-effects were always weak in strength compared with those elicited by moving gratings. We conclude that, as a group, cortical cells strongly sensitive to texture motion are relatively unsusceptible to adaptation. Vision

Texture

after-effects

Cat

Visual cortex

We have previously described potential neural correlates of the motion after-effect which can be elicited from cells of the cat’s striate cortex by prolonged exposure to drifting square-wave gratings (Hammond et af., 1985). The onset magnitude of these neural “after-effects” is typically maximal following adaptation periods of 30-60 set, and they decline rapidly in strength, in more-or-less exponential fashion, over some 30 sec. In common with the few previous neurophysiological reports of such phenomena (Barlow and Hill, 1963; Maffei et al., 1973; Vautin and Berkley, 1977; von der Heydt et al., 1978), all the measured after-effects are directionspecific. The strength of response to the test pattern is reduced in the direction of motion to which a cell has been adapted, and is either unaffected or enhanced in the opposite direction. The recovery pattern is quite different from the habituation of response which characteristically follows a quiescent period of exposure to an unstructured field of similar average luminance. In this report we present the preliminary results of adaptation to drifting fields of 1055

random-texture, subsequently testing the ensuing after-effects with similarly textured fields, swept back-and-forth, or with sweeping bars of optimal orientation. As previously, we have investigated changes of directional bias and temporal variation in response magnitude, following drift in preferred or null directions, or following a comparable period of exposure to a textured, but stationary, field. These results are discussed in relation to parallel human psychophysical studies (Smith et al., 1984; Smith and Hammond, 1985), and our previous report of after-effects evoked by drifting square-wave gratings (Hammond et al.9 1985). Adult cats were lightly anaesthetized with 72.5% 127.5% nitrous oxide : oxygen and 0.2-0.5% halothane, with recovery between recording sessions. Antibiotics were administered prophylactically. Eyes were immobilized with intravenous gallamine triethiodide, protected with unpowered contact lenses and focussed with supplementary lenses on a CRT display at 57 cm; pupils were dilated with atropine sulphate; eyelids and nictitating membranes were retracted with phenylephrine hydrochloride;

10%

Research

5 mm diameter CirCuhr artificial pupils Were applied. EEC, ECG and pulse rate. neuronal firing rate. expired carbon dioxide and body temperature were monitored continuously. For K-137-19:

Note d&IS,

MacKay (1984).

Hammond

(1983)

(1980).

and

Complex (Speclolj

‘K-192-23:

Hammond

Hammond

Single ceils were recorded

Texture dr,ft: 6d.g/s.c

zr

See

and

Texture teat: 190dzg/l0dcg, 6drg/wc

Ttxturc drift: gdeg/scc Tw&urc test: 290dcq/llBdeg. 3deq/sec

O-190dcg X-10deg S-Stat~onory texture

O-290drg X=llBdcg S=Stat>onory texture

drift:; 190 d+Q :

* Texture I

Textura t.*t

z

R..ting

dmJw+-p

/

cxture

null 10 dtg

drift drift: ;ip

: h.

ii

-a*y

_:

60

Fig. 1.

‘T.xtura tcet

o

q

_!____._____ __ tpliRLa\r;llr-Gx-+sT =a x

Smith

from Area 17 with (Stondord)

Complex

and

xx+

120

Yc-192-:4:

T-x&-m

drift:

Bar t-&i

CGI E? -r

Conp!.r

lkWdrg/290drg.

O-lEi@d~g x-28Bdrg Taxturr

Texture

null

drrft:

dr,ft:

(Sp.s,a!)

4&g/~ 4drg/..s,

S-Statronar,

8. 3&,drg t..tur.

100dmg

260dtg

;

Fig. I. Motion after-effects elicited in three complex cells showing, respectively from left to right, strong, weak and negligible adaptation and motion after-effects in response to continuously drifting, randomlytextured fields. The cell on the left represents the strongest effect we obtained from any complex cell in response to texture drift; the other two cells are more representative. In each case the upper row is for texture drifting in the cells’ preferred directions; the second row shows the control situation, with habituation to the test stimulus following exposure only to stationary texture; the lower row is for null drift. Adaptation and test velocities were as indicated; adaptation and test periods were I min each in A to F. 2 minutes each in G to 1. After-effects were tested with similarly-textured fields of optimal velocity in A to F, but with a dark bar of optimal orientation. width and velocity in G to 1, in every case sweeping back-and-forth once every 3 set throughout the test period following adaptation. Records are the averages of 5 periods of adaptation and test in A-C. IO in D-F, and 4 in G-I. 0 = preferred motion or drift: S = stationary texture: X = motion or drift in the null direction. 1057

1058

Research

4M NaCl-filled micropipettes, via a closed chamber. Cells were classified as simple or complex. with assessment of: end-stopping (hypercomplexity); minimum response field map; receptive field location; presence or absence of discrete “on” or “off” receptive field zones; disposition of light and dark discharge centres for motion; ocular dominance; orientation tuning; directional selectivity: velocity and spatial frequency preference; resting discharge level; and length summation-complex cells classified as standard (length summating), special (optimal response to short contours) or unclassified (intermediate in characteristicst(Hammond and Ahmed, 1985). Except that we used texture, rather than square-wave gratings, to adapt and test cortical cells. the experimental paradigm was identical to that described in Hammond et al. (1985). Stimuli were derived from a “Picasso” image generator (lnnisfree Inc.), extensively modified in order to inject hardware-generated random texture into the appropriate point in the circuitry. and displayed on an electrostatic CRT display under computer control. Each cell was adapted with random texture (average luminance I.1 log cd/m’, 512 x 512 elements, 100 framesjsec, IO x IO deg overall subtense), drifting continuously in either the cell’s preferred or null directions at constant velocity, for periods of 1-Z min. Responses to back-and-forth motion of the same field of texture, or a bar of optimal width and orientation, at fixed velocity (in either case I set in each direction, followed by a I set pause), were measured throughout the following I-2 min. Typically five and occasionally ten such sequences were averaged. Where bars were used, their contrast was 0.4 and they were presented against stationary texture. The results were compared with those following a l-2 min period with only stationary texture present. Effects of changing the velocity of the adapting texture were assessed; test velocity was constant throughout. Results are based on timings of spike trains (stored to I msec accuracy on computer disc), averaged response histograms and raster displays of spike firing throughout the entire adaptation’test sequence, and graphical plots of data. The preliminary findings are based on measurements from 32 complex cells-14 standard (4 end-stopped). 17 special (7 end-stopped), I unclassified end-stopped-and 3 simple cells. Figure I illustrates the full spectrum of tex-

Note

ture adaptation and after-effects that we have elicited from complex cells. The fitted curves are informed. freehand. interpretations of the data points, based on an over-all assessment of data for all cells. In all cases the texture velocity which most strongly adapted a cell also elicited the most pronounced after-effect. The cell on the left-a direction-selective special complex cell exhibited the most profound adaptation to texture drifting in its preferred direction (A) and the strongest after-effects of any complex cell we have so far examined. Preferred drift elicited potent adaptation over some 30 set, followed by a short-lived but convincing “after-effect”. also declining more-or-less exponentially in strength over some 30 set following cessation of drift. [we draw attention to the variable firing of this cell, most evident in the latter 30 set of texture drift in A and less so during exposure to stationary texture (B). There is thus some doubt about the precise steady-state level of firing reached during adaptation by preferred texture drift in this cell, as indicated by the broken line in (A).] The strength of this effect is emphasized by comparison of texture test in (A), with that in (B) where the test period simply followed a I min period of exposure to a field of stationary texture. Texture drifting in the null direction (C) elicited weak null suppression of firing, compared with the level of firing to stationary texture (B), but no clear after-effect-the magnitude and time-course of responses in the cell’s preferred direction throughout the period of texture test were comparable to those (B) following exposure only to stationary texture, and responses to texture sweeping in the null direction were unaffected. At the other end of the scale (and relatively common) were cells such as the special complex cell illustrated in Fig. I, right, which were insensitive to adaptation, whether texture drift was in the cell’s preferred (G) or null directions (I); testing with an optimally-oriented bar showed no measurable after-effect, despite using adaptation and test periods of 2 min each. The lack of effect was also in spite of the fact that texture drifting in the cell’s preferred direction elicited potent firing (G), whereas null drift evoked equally potent depression of firing (I), manifest as sustained elevation or depression of discharge, respectively, throughout the adaptation period, compared with the level of firing to stationary texture (H). Figure I, centre. illustrates a more typical result for texture adaptation-a direction-

Research Note

selective standard

complex cell showing adaptadrift (D). and a weak-effect manifest as transient depression of firing to texture sweeping in the preferred direction. The true magnitude of this effect can be appreciated by comparison of texture test in (D) with that following exposure to stationary texture (E) or to texture null drift (F). As with the special complex cell on the left, there was no enhancement of response to texture test in the cell’s preferred direction following texture adaptation in the null direction: the habituation to texture test following null drift was comparable in strength and decay to that following exposure to stationary texture. In conclusion, texture drift evoked weak adaptation and after-effects compared with drifting square-wave gratings (Hammond et al., 1985). Qualitatively, however, results were similar. Thus texture preferred drift brought about a transient reduction in sensitivity to test-texture motion in the same (preferred) direction whereas null drift, at least in direction-selective cells, was without effect. Effects of the strength manifest by the cell of Fig. I, left, were exceptional; Fig. I, centre, was more typical. Texture adaptation and the ensuing after-effects were generally stronger in standard (lengthsummating) than in special complex cells (those lacking appreciable length summation). Indeed, many of the special complex cells-which, as a class. are particularly strongly driven by texture motion-showed no consistent adaptation or after-effects (Fig. 1, right). For the illustrated special complex cell, in particular (Fig. 1, panel I), the only effect of adaptation was the slight but noticeable consolidation of null suppression during the adaptation period, associated with a slight rise of firing in response to the test bar moving in the cell’s preferred direction during the subsequent test period. Such long-term depression of firing below the resting discharge level might plausibly induce prolonged hyperpolarization, followed by comparably slow recovery during the test period. At this stage we cannot be certain whether the weakness of textural after-effects relates primarily to stimulus class or to cell type. The results are at odds with our parallel psychophysical observations in which we have elicited powerful after-effects in response to drifting texture (Smith ef al., 1984; Smith and Hammond, 1985). Moreover, the perceived velocity of the psychophysically-observed motion after-effect is known to be much lower than the tion

to preferred

1059

adapting velociry (Smith er al., 1984) and most strongly influenced by adaptation velocities higher than that of the test pattern (Thompson. 1981). In cortical neurones, however. textural after-effects were always weak, even though we endeavoured to maximize them by selecting drift velocities close to the optimum for each cell and by employing test velocities lower than the velocity used for adaptation. In the early stages of the investigation we were uncertain Lvhat time-course or strength of adaptation and after-effect to expect. Thus initially we adapted and tested for 2 min each, despite which adaptation was negligible, and after-effects were weak and short-lived (Fig. I, right). More recently we therefore reduced the adaptation and test periods to 1 min each (Fig. 1, left and centre), having verified that adaptation of duration sufficient to promote steadystate conditions and a plateau level of firing also evoked a maximal-strength after-effect. Parallel psychophysical observations (Smith et al., 1984; Smith and Hammond, 1985) suggest that after-effects are optimal when the class of stimulus is similar for adaptation and test sequences, i.e. texture/texture or bar/bar, rather than texture/bar. It must be acknowledged that in the cell of Fig. I (right), as in 14 others investigated-including all 3 of the reported simple cells-we adapted with texture but tested the after-effect with a sweeping bar. On the other hand, negligible after-effects were also obtained from many of the 17 cells which were both adapted and tested with texture. At the neuronal level, therefore, we have no indication, overall, that the weak after-effects evoked by are because of an adaptationtexture stimulus/test-stimulus specificity. AcknorYedgemenrs-Supported by grants G8318037N and G8008334N from the Medical Research Council to P.H. D. J. Scott provided expert technical assistance. Additional backing came from: Smith & Nephew Pharmaceuticals Ltd; Smith & Nephew hledical Ltd; Smith & Nephew Textiles Ltd; Monoject Division of Sherwood Medical; Abbott Laboratories Ltd; Roche Products Ltd; Imperial Chemical Industries Ltd: The Wellcome Foundation Ltd; CNSRespiratory Division of Astra Pharmaceuticals Ltd.

REFERENCES Barlow H. B. and Hill R. M. (1963) Evidence for a physiological explanation of the Waterfall phenomenon and figural after-effects. IVarure 200, 1345-l 347. Hammond P. (1980) A semi-chronic preparation for cortical recording. J. Ph~sid., Land. 298, 3-4P.

Research

I060 Hammond

P. and Ahmed

complex

B. ( 1985) Length

cells in cat strtate

“special ” “standard”

cortex:

summatton

a reappraisal

.Vruroscience

classilication.

of

of the

15,

639-649. P. and

lumtnance

MacKay

gradient

D.

?.l. (1953)

lntluence

of

reversal on simple cells in feline striate

J. PhFsiol.. Land. 337, 69-87.

Hammond Motion

P., Mouat

Hammond

G.

after-effects

S. V. and

in cat striate

P. and Smith

cells in cat striate

Smtth

cortex

A.

T. (1985)

elicited

A. T. (1983) Sensitivity

cortex

to relative

by mov-

ofcomplex

motion.

Brain

Rrs.

301. ‘87-798. of

1.. Ftorent(ni

A. and Bisti S. (1973) Seural adaptation

to

gratings.

correlate

Scwnce

182.

10361038 ,Wo\shon

Smith

h.

The

A.

and

Lennie

after-etfects.

P. (1979)

Pattern-selective

corncal

neurones.

.Va~ure

Thompson

of

278,

specifictty

Rrr. 60. 71-78.

21. J. and Hammond

background

motton

on

P. (198-t) the

motion

V~siun Rrs. 21. lO75-IOY2.

)

P. (I951

C’eiocit>

adaptation

to mo\(ng

sequently

seen making

Vautin

P. (1985) The pattern E.rpi Brain

T.. Mussel*hite

tnhuence

R. G. and Berkley

cells in cat visual cortex correlates

af:er-eHects:

stirnull st(mult.

the eHects of

on the perceptton

of sub-

I~~.rion Rrs. 21. 337-345.

hf. h. (1977) Responses of stngle to prolonged

of visual

sttmulus

after-effrcts.

movement:

J. .Veurophy.vrol.

40, 1051-1065. van

der

Heydt

Clokement J.

v~suai

Smtth A. T. and Hammond

neural

perceptual

in

85%8jl.

after-effect.

E.rpl Brcrin. Rex. 60, II 1116.

ing gratings.

Maffet

adaptation

of velocit)

Hammond cortex.

Note

Biol.

R..

Hannk

after-eHects

116. 248 -254

P. and

In the wsual

Adorjani cortex.

C.

(1975)

.-lrclrs 11~11.