Exp Brain Res (1991) 87:218 222

Experimental BrainResearch 9 Springer-Verlag1991

Research Note

Express saccades: is there a separate population in humans? M.G. Wenban-Smith and J.M. Findlay Department of Psychology,University of Durham, Durham DH1 3LE, UK Received January 22, 1991 / Accepted May 27, 1991

Summary. It is well known that the latencies of target elicited saccades are significantly reduced when the target onset is preceded shortly by the offset of a fixation point (Saslow 1967). Fischer and Boch (1983) reported the discovery that, with monkeys as subjects, in addition to the general reduction in saccade latencies previously reported, there occurred a separate population of saccades with extremely short reaction times. They termed this population "express saccades", and more recenly reported the discovery of an equivalent population of express saccades for humans (Fischer and Ramsperger 1984; Fischer 1987). In this paper, work is reported which confirms the existence of short latency visually guided saccades in humans but questions whether these form a separate population of "express saccades". The conditions used were very similar to those used by Fischer and Bocn. Key words: Eye movements - Reaction time - Express saccade - Human

Introduction In many studies of saccadic control a trial begins with the appearance of a fixation point which the subject is asked to foveate. When the target stimulus subsequently appears the fixation point may be removed. Saslow (1967) pointed out that the offset of the fixation point and the onset of the target stimulus in fact constituted two independent visual events. He showed that the relative timing of these events had a strong influence on the saccadic control processes. Saslow varied the timing of fixation point offset with respect to target onset, and measured saccadic latency with respect to the target onset. Fixation point offset could either precede (gap condition), coincide with (simultaneous condition), or follow (overlap condition) Offprint requests to: J.M. Findlay

target onset. Saslow found that a gap between fixation point offset and target onset led to shorter saccade latencies, and an overlap to longer latencies, with simultaneous presentations having an intermediate effect. In Saslow's original study saccades made with gaps of greater than 250 ms gave saccade latencies of about 130 ms, and saccades made with overlaps greater than 150 ms resulted in saccade latencies of about 240 ms. For intermediate gaps/overlaps, latencies changed monotonically, with a latency of 195 ms when fixation offset and target onset were simultaneous. The same basic procedure has been used by a number of authors since Saslow (Ross and Ross 1980; Reulen 1984; Kalesnykas and Hallett 1987; Braun and Breitmeyer 1988). Although the values of mean latency for each gap condition seem to vary considerably between different studies, and although there is large intersubject variability (e.g. Reulen 1984), the latency reduction in various gap conditions has been consistently replicated. More recently a second phenomenon of latency reduction associated with use of the gap condition has been reported. Fischer and Boch (1983) used the gap paradigm to study the latency of target elicited saccades in the monkey. In addition to the expected decrease in saccade latencies as the gap length was increased, they reported a discrete population of extremely short latency saccades that occurred when the gap exceeded about 150 ms. This population had latencies with a mean of around 70 ms, clearly separated from the population whose mean latency was about 150 ms. They called these extremely short latency saccades "Express Saccades'. There are therefore two phenomena of decreased saccadic latencies associated with the gap condition. The first of these, following Fischer's terminology, is the reduction in "regular" saccade latency, from the normal latencies observed when there is no offset of the fixation point, to the "fast regular" latencies observed when target onset is preceded, or very closely followed, by fixation point offset. The second phenomenon is the occurrence of a separate population of extremely short latency saccades, separated from the "fast regular" population, oc-

219 c u r r i n g u s u a l l y w i t h g a p s g r e a t e r t h a n 150 m s a l t h o u g h p r a c t i c e m a y l e a d to the g e n e r a t i o n o f express saccades w i t h s h o r t e r g a p d u r a t i o n s ( F i s c h e r et al. 1984). T h e essential c h a r a c t e r i s t i c o f express saccades is t h a t t h e y o c c u r as a s e p a r a t e d p o p u l a t i o n to the r e g u l a r saccades. T h e r e is n o t r a n s i t i o n o f i n t e r m e d i a t e latencies b e t w e e n the fast r e g u l a r a n d express saccades. T h e distrib u t i o n o f s a c c a d e latencies is b i m o d a l w i t h one p e a k c o r r e s p o n d i n g to express s a c c a d e s a n d a second, l o n g e r latency, p e a k c o r r e s p o n d i n g to fast r e g u l a r saccades. F i s c h e r a n d B o c h initially failed to find a similar b i m o d a l d i s t r i b u t i o n o f latencies w h e n h u m a n s were used as subjects in a s i m i l a r e x p e r i m e n t a l p r o c e d u r e , b u t in 1984 F i s c h e r a n d R a m s p e r g e r r e p o r t e d finding a b i m o d a l d i s t r i b u t i o n o f latencies for saccades in h u m a n s - a n "express" population with a mean latency of around 115 ms, a n d a n o r m a l p o p u l a t i o n w i t h a m e a n l a t e n c y a r o u n d 155 ms, a l t h o u g h the express p e a k d i d v a r y for different subjects b e t w e e n 115 a n d 135 m s ( F i s c h e r a n d R a m s p e r g e r 1984). T h e r e p o r t e d difference b e t w e e n the " f a s t r e g u l a r " s a c c a d e l a t e n c y p e a k a n d the " e x p r e s s " s a c c a d e l a t e n c y p e a k is c o n s i d e r a b l y less t h a n in m o n keys. W e were i n t e r e s t e d in these differences a n d felt it w o r t h w h i l e to c a r r y o u t a s t u d y e x a m i n i n g the statistics o f s a c c a d e latencies in h u m a n s u n d e r different g a p conditions. M a y f r a n k et al. (1987) v a r i e d g a p d u r a t i o n a n d r e p o r t t h a t n o express saccades o c c u r w h e n g a p d u r a t i o n is r e d u c e d to zero. A l t h o u g h m o r e recent w o r k (e.g. B r a u n a n d B r e i t m e y e r 1988) suggests t h a t a g a p b e t w e e n fixation p o i n t offset a n d t a r g e t o n s e t m a y n o t be a necessary c o n d i t i o n for the o c c u r r e n c e o f " e x p r e s s " saccades, it nevertheless r e m a i n s the simplest a n d m o s t easily cont r o l l e d m e t h o d for their elicitation. T h e g a p / o v e r l a p p a r a d i g m is t h e r e f o r e the m e t h o d o f p r e s e n t a t i o n u s e d in this investigation. D a t a h a v e b e e n collected o f s a c c a d e latencies o v e r a r a n g e o f d u r a t i o n s o f g a p s a n d overlaps. T h e p r e d i c t i o n s f o l l o w the earlier w o r k b y Fischer. A s g a p l e n g t h is i n c r e a s e d it is e x p e c t e d t h a t m e a n saccadic l a t e n c y will fall. A t s o m e p o i n t it is e x p e c t e d t h a t the s h o r t l a t e n c y p o p u l a t i o n will split into s e p a r a t e p o p u l a tions o f " f a s t r e g u l a r " a n d " e x p r e s s " saccades. A s n o t e d , in h u m a n s the m e a n latencies o f these t w o p o p u l a t i o n s m a y be close together. T h e two d i s t r i b u t i o n s w o u l d t h e n s h o w m o r e o v e r l a p a n d m i g h t be difficult to d i s t i n g u i s h f r o m the d i s t r i b u t i o n o f a single p o p u l a t i o n . It w o u l d nevertheless b e e x p e c t e d t h a t the v a r i a n c e o f the latencies s h o u l d increase as the split is reached. I n a n a l y s i n g the results t h e r e f o r e p a r t i c u l a r a t t e n t i o n will be p a i d to the d i s t r i b u t i o n o f the d a t a as the m e a n l a t e n c y decreases.

Methods The authors and one other member of the Psychology Dept. were used as subjects. The authors (MWS, male, 24 yrs and JMF, male, 47 yrs) were aware of the purpose of the experiment, the other subject (ALA, female, 29 yrs) was unfamiliar with the precise nature of the experiment. Stimuli were presented on a Phillips TP-200 monochrome monitor with phosphor P-31. Because of the use of this phosphor the fading image of the fixation point was clearly visible some time after

it had been turned off when in a dark room. This problem was overcome by using a lit background brighter than the fading image. Throughout the experiment the background luminance of the monitor was 19 candelas/m E. The luminance of the target stimulus was 69 candelas/m 2. The timing of stimulus presentation was controlled using a BBC series B microcomputer. Care was taken to ensure that the stimuli could be changed within a single raster scan. Horizontal eye movements were recorded using a method based on the infrared reflectometry technique of Stark and Sandberg described by Young and Sheena (Young and Sheena 1975). The equipment used was an EM 130 eye movement monitor unit from ACS Applied Research Developments Ltd. The unit provided an analogue signal representing eye position. This signal was digitised and recorded using a Cambridge Electronic Design Alpha computer with 502 interface. The analogue signal was sampled and digitised every 2.3 ms, and the data recorded onto disc: Recording was initiated at the moment of stimulus presentation, and continued for 1.5 s. These records were later analysed using a program that determined saccade onset using a threshold for eye velocity (about 15 deg/s). The computed onset was marked and visually inspected to ensure that the program had not picked up noise on the record, and to ensure that the record was not contaminated with artifacts due to blinks etc. A simple calibration of eye movement amplitude was carried out at the beginning and end of every experimental session, and in between every experimental block. For the best recordings noise represented + / - 0 . 2 degrees. For the worst recordings it might be as high as + / - 0 . 5 degrees. Subjects were seated in a dimly lit room with the monitor at a distance of 83 cm, and their heads stabilised by use of a dental bite. Each trial began with the appearance of a fixation point in the form of a small dot surrounded by a circle (diameter 0.75 degrees). The fixation point remained on for a period varied randomly between 1 and 1.5 s in approximately 160 ms steps. This was to ensure that there was no accurate information from fixation point onset that would help predict the timing of target onset. The target stimulus consisted of a small square (size 0.75 degrees) positioned 4 degrees to the right or left of the fixation point. Target onset either preceded fixation stimulus offset (overlap condition) or followed fixation stimulus offset after a pause (gap condition). Within an experimental session subjects received trials of either the gap condition or the overlap condition, but never mixed. Each experimental session consisted of 4 blocks of 64 trials. Within a block the gap or overlap duration was varied between 0 and 300 ms in 20 ms steps, and for each gap or overlap the target was presented twice on each side of the fixation point. The order of presentation was randomised within every block of trials. Thus neither the position nor the timing of stimulus presentation was predictable. A single block lasted about 5 rain, and a complete session under 30 min. Subjects were provided with a hand held response key and initiated each trial by pressing the key. They were encouraged to respond as fast as possible once a trial was initiated, but were told they could pause between trials if they wished. Subjects were instructed to fixate the dot in the centre of the fixation point until the target appeared, and then to make an eye movement to the target as fast as possible. They were warned that the duration of the fixation point remained on would vary, and that they would sometimes be aware of a gap between fixation point offset and target onset, and so it was possible they might make eye movements in the wrong direction in anticipation of the target's appearance. They were told not to worry if this happened, and that it was better to make fast responses that were occasionally wrong than to make an effort to ensure that responses were always correct. After analysis of the calibration records the trial records were inspected. Data were discarded if the record of the saccade was not clear due to the occurrence of blinks, excessive noise or unsteady initial fixation. At this stage 3.5 % of all trials were discarded. Initial eye position, saccade latency, and saccade amplitude were then recorded for further analysis. As Kalesnykas and Hallett (1987) pointed out, in analysing the results of experiments using a gap paradigm it is essential to have a method of eliminating anticipatory saccades from the latency

220 I

200 175

I

I

I

I

I

I

I

I

240

I~ 3 S's: wrong direction

Subject

200

"'i .m

i

i 3 S's: right direction

9

9 9

JMF

9

~'"-m.i r . - ~ . ~ ' " i " " - n ~

9

160

150

9 ' "l~-iI.~i 9

120 o~

80 'i !

~. 100

40 ~ o

[]

=~

75 0

"§ ...................

~=o

0o ~

-r.................. v................... :" ................... ~".................. v ................... ,

-300

50

I-I-i"

"i

125

overlap

0

gap

300

25 200 7 --;.......~...,.j....,?.i . I ....i ......... i 160 9 9 9 9 " -.,

o

5

15

25

35

45

55

65

75

85

95

Latency/ms

Fig. 1. Initial part of latency distribution, showing the frequency of correct and incorrect saccades for the three subjects' combined data

Subject 9

.= cD

N Ill

nIiNiii

ALA lmi

20 iii 8o i 4o -i

records. These saccades may have decreased latency not because of any particularly fast guided response to the target, but due to the elicitation of a motor program determined prior to the target's appearance. When the latency of saccades is as low as 20-30 ms, it can be safely assumed that they are not target guided as efferent and afferent delays preclude such a fast reaction time. A short latency cut off cannot however be used as a general criterion by which to eliminate anticipatory saccades, as it is specifically unusually low latency saccades that are of interest. If target position is predictable, then it is impossible to eliminate all anticipatory saccades from the records. Although the characteristics of anticipatory saccades may differ from those of target elicited saccades (Smit and Van Gisbergen 1989), these are not sufficiently distinct to allow their accurate identification. However if target direction is unpredictable, then it is an easy matter to determine whether saccades of a particular latency are consistently guided in the direction of the target or not. Previous studies conclude that the computation of target position is made in parallel over the visual field (e.g. Heywood and Churcher 1980), so there is no reason to suppose that there will be an additional aspect of latency associated with the uncertainty of target direction. The distributions of saccades made in the correct and incorrect directions for the three subjects are shown in figure 1 as a function of saccade latency. From the distributions it can be seen that there is a sharp transition from the case where saccade direction is as often incorrect as correct to that where all saccades are correctly directed. No incorrect saccade was made for any subject with a latency greater than 78 ms showing that the figure of 120 ms suggested by Kalesnykas and Hallett (1987) is almost certainly too conservative. On this basis all saccades with a latency less than 80 ms were eliminated from further analysis. Saccades with outlying amplitudes (more than 2 standard deviations from the mean amplitude) were also discarded. The removal of anticipatory saccades, and of saccades of outlying amplitudes led to the elimination of a further 7.5 % of the available data. The remaining records were then analysed further.

Results and discussion

Figure 2 shows the mean saccadic latency for each condition of gap or overlap for all three subjects. Each data point is based on 32 trials. The lines between the scattered points are produced by an iterative smoothing function from a proprietary software package. As expected from the work of other authors (e.g. Saslow 1967;

.~oo 0..i

~

o

o

oomoma

cl

~176176176176

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

-300

overlap

Q

gap

300

2 0 0 --~ ~ " I ~ " ' I ' ~ ' ~ " ~ " " ' I ' ~ ,

16o ]] .E

120

"i

80

--

40

9 II'~-~.li.i

q

ii ii~ "'I'~I'II

[]

i

0

MWS

'" " "

-i

=

Subject

%

~176176

~

D

gap

300

~. -300

overlap

0

Fig. 2. Means (filled symbols) and standard deviations (open s y m bols) of saccade latencies for each of the gap and overlap conditions presented. The abscissa shows the separation in ms between the fixation point offset and target onset. Each point is based on 32 saccade trials. The curves show a smoothed regression fit to the data points

Reulen 1984) there is a transition from longer latency saccades in the overlap conditions to shorter latency saccades in the gap conditions, and this is likely to account for the overall bimodal distribution of saccade latencies when gap and overlap conditions are combined. It is worth noting that the mean latencies to the right of the graphs are well within the range of"express" latencies described by Fischer and Ramsperger (1986) for different subjects. The mean of the longer latency saccades is between the reported "regular" and "fast regular" latencies. Two possibilities must be considered. Firstly, the decrease in mean saccade latency might represent a shift from the regular population to the express population. In this case, those gap conditions that lead to saccades with intermediate mean latencies would be composed of a mixture of saccades from the "regular" and "express" populations. The standard deviations of the latency of

221 the saccades in these intermediate gap conditions should be significantly greater than the corresponding standard deviations in the conditions leading to only either "regular" or "express" saccades. The second possibility is that the intermediate latencies represent the transition from "regular" to "fast regular" saccades. In this case, express saccades occur for the same gap conditions that lead to "fast regular" saccades, and the "fast regular" population should be inspected more closely for bimodality. In order to test the first possibility in which the intermediate gap condition populations show bimodality it was necessary to study the relationship between the mean latency in each condition and the standard deviation of the mean for the condition. If bimodality occurs for intermediate mean latencies, then an augmented standard deviation would be expected in the transition region. Figure 2 also shows the standard deviation with each gap duration for the three subjects. In no case is there any systematic increase in standard deviation in the tr~/nsition region. Rather the standard deviation decreases in a similar manner to the mean latency. The second possibility, that the transition from longer latency to shorter latency saccades is the transition from "regular" to "fast regular" saccades, and that express saccades would only occur after this transition, was therefore also tested. Figure 3 shows, for each subject, a representation of the latency distributions with each gap duration. Subject MWS produces saccades with a very fast mean latency but with no indication of bimodality. Neither, for this subject, is there any increase in standard deviation during the transition period. It must be concluded for this subject that the short latency saccades represent speeded regular saccades and that there is no evidence whatsoever of a separate population of express saccades. The distributions for the other subjects do show some sign of bimodality following the transition region, particularly in the case of JMF. If the data for the all three subjects is considered, there appears to be a common peak at around 90-110 ms, and a possible second peak at longer latencies for A L A and JMF. However it seems difficult tO maintain that the fast peak represents express saccades for some subjects but not for others. A more parsimonious interpretation might be that the fast latencies represent an irreducible minimum latency for saccade production under optimum conditions. The optimum conditions would be an appropriate gap in the eliciting stimulus, together with perhaps an optimum state of readiness. Slower saccades produced under optimum gap conditions would reflect some failure to maintain internal readiness. Adopting this approach it might be possible to reconcile the results in this paper with those which have shown bimodal distributions. The suggested optimum state of readiness could depend in part on central control and it is possible that for some subjects modulation of internal readiness operates to allow a brief momentary peak which gives rise to very short latency (express) saccades but is not always adequate to trigger the eye movement. If the saccadic eye movement is not triggered by this early

500 t: 19

I I

J

:.". . . . .

if:

Subject

18

JMF

17

~!;::-i::::;?:~ ....... ::i~" .i:' :

E

"..-.:~-,'-'::-'-.:'!:i:i:i::-.:'::.~';.i:.~:

16

:

= "[.i.:-?ii:. ii:iii-.ii~:i;...%~;~:::':-i ! o

~:

9

- 9:~-;i:'~ :..::.::.::":~:':?;::':':~i.:Kj::-"v'.:i:. ":.:':~.- : .,.'<'..:" ::~.'s..:-.,-&:.. ,.'..: ":~:.:::: ~-:.: : : : . .v:~l~.:;.~:."..',:;~'::,'~':.:,

15 ....

.'..',:- .~-.','~-:

- ..:.:-.:::-...'~;:" ..:."~--,~.:--:.'.:-:.-' . ~..~'.-"~~:-.;. :: <.'::

_

<'-:'.."':~$:':..'::5"::':'..'-'.

.::..:

.

.

.

.

.

-'~-:

. .:iiii..-..!..:...

14

. ".:..".!.:.:.~':.'-.'.:-'.:':':."-"

.

,,.

.,.,,,

. , ,.,,.... ..

15

Iz ::~! :.'.S~'..';','.'!:'~ !: ! ~::-"" '"5

overlap

-300

0

gap

11 300

tO

m

500 12 11 10

Subject A L A

i!::, ;i,ii:il ii

(t)

E

.;~..~:;.~?',:....~::-:~,~......... ..:~, 9~::. :.."i~:." :U-":':;~.'...e':..':

'::.'.?.g.

!i!.:..:... ::i..:....i:'..

s

"'i~:'.,~,::.~:........,:,.:..:. ::,i: . :.: ~.~. ,~, ,. :~..:.:.> :9 .';'L. :. :" ".,~" :.iz,].: :,~i:: :~': :.~ '~;:"~'"~(-.,'-" :.'.:.:."~ :.-.- :.;,: -~

~,t.t~- ..~'~'i,;,.$.~~~-~:.:..':,~.~..-.-:~.~~.,.:~ ~..~.,~ ~'~~.~"-:~,~.-'~,~.~.' :: -.~~~~ 9.. :..-

80

,

,

,

,

-300

i

i

i

,

i i i

,

i

i

overlap

,

i

,

: : : ! ~q~i~!.-~.:

,

,

,

0

,

,

,

~:...:.

,

,

,

gap

,

,

500

li ~-~0

300

1E

Subject M W S

lz m

E

-

IC G)

::!iiil;&,:, .....~-:~~: ". ,."..:. ~:.'.! ! :",-::';:[.",*,.:~ ~;,'r ":~!::"

80

i

-500

,

i

i

J

i

i

J

i

overlap

i

i

i

i

,

i

O

i

i

,

-.~z..'.::.:.~.

"'::~i!::.-':-':" "" i

,

i

i

i

gap

i

i

,

~l

,

500

o

Fig. 3. Contour plots of the distributions of saccade latencies for each subject. In the critical regions, the contour plot allows an evaluation of the shape of the latency distributions. However some outlying isolated data points do not appear on the plot

peak, a longer latency (regular) saccade occurs. This provides an alternative account to the suggestion sometimes made that separate anatomical pathways are involved in the production of different types of saccades and also an account which might allow the effects of practice to be more readily incorporated.

222

Conclusions

Although the experimental technique described has been used on many previous occasions, and many of the basic findings have been previously published, it is only recently that the phenomenon of express saccades has been reported. It was possible that previous work had not observed such saccades because their unusually low latencies led to the assumption that they were anticipatory, and that they had thus been excluded from analysis. In this work particular care was taken to look for such express saccades, and relate their occurrence to the data previously published using similar techniques. Using the techniques described, saccades with short latencies, similar to the latencies previously described as "express" were observed. In one subject, these saccades clearly did not form a separate population from the longer latency saccades of the overlap condition. The transition appeared to correspond to the change from regular to fast regular saccades but at much shorter latencies than previously reported. The distributions for the other two subjects were more scattered and showed some suggestion of bimodality but overall the results do not support the existence of a separate population that can be described as express saccades. References Becker W, Jfirgens R (1979) An analysis of the saccadic system by means of double step stimuli. Vision Res 19:967-983 Braun D, Breitmeyer BG (1988) Relationship between directed visual attention and saccadic reaction times. Exp Brain Res 73 : 546-552 Fischer B (1987) The preparation of visually guided saccades. Rev Physiol Biochem Pharmacol 106:1-35

Fischer B, Boch R (1983) Saccadic eye movements after extremely short reaction times in the monkey. Brain Res 260:21-26 Fischer B, Breitmeyer B (1987) Mechanisms of visual attention revealed by saccadic eye movements. Neuropsychologia 25 : 73-83 Fischer B, Ramsperger E (1984) Human express saccades: extremely short reaction times of goal directed eye movements. Exp Brain Res 57:191-195 Fischer B, Ramsperger E (1986) Human express saccades: effects of randomization and daily practice. Exp Brain Res 64: 569578 Fischer B, Boch O, Ramsperger E (1984) Express-saccades of the monkey: effects of daily training on the probability of occurrence and reaction time. Exp Brain Res 55:232-242 Heywood S, Churcher J (1980) Structure of the visual array and saccadic latency: implications for oculomotor control. Q J Exp Psychol 32:335-341 Kalesnykas RP, Hallett PE (1987) The differentiation of visually guided and anticipatory saccades in gap and overlap paradigms. Exp Brain Res 68:115-121 Mayfrank L, Kimmig H, Fischer B (1987) The role of attention in the preparation of visually guided saccadic eye movements in man. In: Eye movements from physiology to cognition. O'Regan JK, L6vy-Schoen A (eds) Elsevier, Amsterdam pp 37-45 Reulen JPH (1984) Latency of visually evoked saccadic eye movements. I. Saccadic latency and the facilitation model. Biol Cybern 50:251-262 Ross ER, Ross SM (1980) Saccade latency and warning signals: stimulus onset, offset and change as warning events. Percept Psychophys 27: 251-257 Saslow MG (1967a) Effects of components of displacement-step stimuli upon latency for saccadic eye movement. J Opt Soc Am [A] 57:1024-1029 Smit AC, Van Gisbergen JAM (1989) A short-latency transition in saccade dynamics during square-wave tracking and its significance for the differentiation of visually-guided and predictive saccades. Exp Brain Res 76:64-74 Young RL, Sheena D (1975) Eye-movement measurement techniques. Am Psycho1 30:315-330