Reaction times of the eye and the hand of the monkey in a visual reach task

Neuroscience Letters, 58 (1985) 127-132

127

Elsevier Scientific Publishers Ireland Ltd.

NSL 03398

R E A C T I O N T I M E S OF T H E EYE A N D T H E H A N D OF T H E M O N K E Y IN A V I S U A L R E A C H TASK

L. ROGAL, G. REIBLE and B. FISCHER*

Department of Clinical Neurology and Neurophysiology, University of Freiburg, D-7800 Freiburg i. Br. (F.R.G.) (Received February 8th, 1985; Revised version received March 28th, 1985; Accepted April 16th, 1985)

Key words." reaction time - eye movement - reach movement - eye-hand coordination - express saccade -

monkey - learning

Two monkeys were trained to excute saccadic eye movements and reach movements with the hand from a central fixation point to a peripheral target. Reaction times for both movements were compared on a trial-by-trial basis. If the fixation point was extinguished before the target appeared (gap condition), extremely short latency saccades (85 ms) (express saccades) were obtained, that were followed by short latency reach movements (250 ms), but there was no correlation between them on a trial-by-trial basis. If the fixation point remained visible (overlap condition), very short (100 ms) and rather long (220 ms) latency saccades were observed. Long saccadiac latencies correlated strongly with the reach reaction times, Short latency saccades were followed by reach movements of reaction times longer than those observed after express saccades in the gap condition; there was no correlation between them. All reaction times varied systematically with practice.

In a visually guided reach task, where one has to reach out and touch with the hand a target in the periphery of the visual field, usually a goal-directed eye movement precedes the reach movement of the hand. A single sensory event, the occurrence of the visual target, initiates the preparation and execution of two movements, which then take place in sequence. From their experiments on human subjects, Biguer and co-workers [2] have considered the possibility that the two movements may be initiated by a single common command generator. However, they noticed also that there was almost no correlation between the reaction times of the eye and the hand on a trial-by-trial basis. The existence of goal-directed saccades after extremely short reaction times (express saccades) in monkeys (75 ms) [3] and man (100 ms) [5] opens the possibility to measure the reaction times of visually guided reach movements, when they are preceded by express saccades, and to compare them to those, which are preceded by saccades of regular reaction times (about 220 ms). The large difference in the reaction times of regular and express saccades should be reflected in a correspondingly large difference in the reaction time of the hand movement. *Author for correspondence at: Department of Clinical Neurology and Neurophysiology, Hansastrasse 9, D-7800 Freiburg i. Br., F.R.G. 0304-3940/85,/$ 03.30 © 1985 Elsevier Scientific Publishers Ireland Ltd.

128 Two monkeys (Macaca mulatta) were used in the experiment. They were placed in a primate chair. Their heads were fixed by a permanently implanted metal bar and eye movements were monitored by an infrared-light-sensitive device [1]. The animals viewed a light-emitting diode (LED) and a CRT screen by means of two half-silvered mirrors. The LED was used as a fixation point and the CRT was used as a stimulus display. In front of them was a plate on which the light stimuli seemed to appear. With this arrangement, the monkey could reach the stimuli without covering them with his fingers. Animals were first trained to fixate a stationary spot of light (fixation point) using the dimming paradigm [7]. In the fixation task, animals had to touch the fixation point after its appearance at the beginning of a trial. They had to hold their hand in this position for a randomly varying period of time after which the fixation point dimmed. Animals had to detect the dimming and to redraw their hand within a maximal reaction time of 700 ms to receive a water reward. To control the correct positioning of the hand, two infrared light beams were used, that crossed at the fixation point. Interruption of both beams indicated that the monkey had positioned his hand correctly. The saccade task started also with the presentation of the fixation point. A peripheral (eccentricity: 11 degrees) target (size: 1.25 × 1.25 degrees) occurred 2000 ms later and the monkeys had to detect the target's dimming. Since the dimming was made difficult to see (small step in luminance), the animals made saccades to the target. In this task, they had to hold their hand at the central fixation point throughout the trial until they detected the dimming. The saccade-and-reach task was physically identical to the saccade task, but in addition the animals had to move their hand from the central fixation point to the peripheral target. The correct position of the hand at the peripheral side was monitored by an appropriately placed capacitive detector behind the touch screen. Two paradigms were used. The first is called the gap paradigm: the fixation point was turned off some time (100 or 200 ms) before the target appeared, this paradigm is good to elicit express saccades. The second is called the overlap paradigm: the target appeared while the fixation point remained on. The command to move was the occurrence of the target in both paradigms. Saccadic reaction times (SRT) were determined automatically by detecting the beginning of the saccade after target onset. Reach reaction times (GRT) were defined by the time, when the monkey's hand left the position of the fixation point. All reaction times were measured from target onset. Altogether, we collected more than 60,000 pairs of movements from the 2 monkeys. Here, we report only the basic result that emerged from a comparison of express saccades and regular saccades preceding the corresponding reach movements. After the animals had learned the fixation task, they had to learn the saccade task using the gap paradigm with a gap duration of 100 ms. When they produced a sufficient amount of express saccades (usually after 10-14 days of daily practive, ref. 4), they had to learn the saccade-and-reach task using the gap paradigm. After quite a number of days of practice (10-20 days), the data of Fig. 1A were collected. The reaction times of express saccades were plotted against reach reaction times on a trial-by-trial basis. The figure shows a horizontal narrow strip of data points. For the visual conditions used in these experiments, the mean saccadic reaction time was 85 + 5.4 ms. The mean reaction time for the reach

129 SRT 'ms

SRT/ms

21~iPi

Gap express

- socc

Over[o.p regul socc

A

1501

5N3500" 450-

~ SRT/ms 200

2;0 36o 4;0 560 ~

Overtop: e x p r e ~

GRTIms 4007;0 8(;0 350250" 2(20"

~ socc

C GRams

t;0

2;0 3;0 4;0 ~o

~

7~

8~

15010050"

B

GRTIm,: ,6o 260 36o ~;0 s~o 6;0 76o ,~o

Fig. 1. Correlograms between saccadic (SRT) and reach (GRT) reaction times of monkey PI. A: gap paradigm. Reach movements preceded by express saccades. B: overlap paradigm. Reach movements preceded by regular saccades. C: overlap paradigm. Reach movements preceded by express saccades.

movement was 247-1-25 ms. The next step was to teach the animals to perform the visual reach task using the overlap paradigm. Again the monkeys had a number of days of practice, before the data of Fig. IB, C were collected. Part B of the figure shows that in the overlap paradigm, saccadic reaction times were broadly scattered with a mean value of 208__+35 ms, and reach reaction times had a mean value of 368 + 52 ms. Therefore, one effect of switching from the gap paradigm, where many express saccades occur, to the overlap paradigm, where many regular (long latency) saccades occur, is an almost parallel increase of mean reaction times of the eye and the hand. The second effect is that in the overlap case, the data points exhibit quite a large scatter with a tendency for eye movements of longer latencies to be followed by reach movements of correspondingly longer latencies on a trial-by-trial basis. In fact, if one plots the mean values of the SRT against those of the GRT (Fig. 2), they can be connected by a straight line of slope 1. Therefore, the changes of the mean values seem to support the hypothesis of a common command for the eye and hand movement. However, as the training of the overlap paradigm was continued, both monkeys began to produce a bimodal distribution of SRTs even though the fixation point remained on. The corresponding data have been treated separately and are plotted in Fig. 1C. The mean value of the short saccadic reaction times was 99 + 6.5 ms, but the reach reaction times are clearly shifted to the right, having a mean value of 321 + 3 3 ms. The latter two mean values are also plotted in Fig. 2. They clearly fall off the straight line. This means that in this case it is not the preceding saccade which accelerates or delays the following reach movement. To further look into the nature of the difference of the three sets of data (Fig. 1A-C), we evaluated the reaction times during the time of practice in the overlap paradigm (Fig. 3A-C). What is hardly seen in Fig. 1B became clear from several (but not all) training sessions: saccades of long reaction times are followed by reach movements of correspondingly long reaction times, such that there was a strong correlation on a trial-by-trial basis (Fig. 3A). As

130

SRT/m s

300-J

• RHO

SRT= GRT-151ms

oPI

20

Overlap

100Gap/~

00

100

2 0

t tOverlap

300

1.00 GRT/rns

Fig. 2. Mean values of SRT versus mean values of GRT computed from the data presented in Fig. I. Standard deviations are indicated by vertical bars. A line of slope 1 was fitted through the data points corresponding to the gap paradigm (express saccades) and the overlap paradigm (regular saccades). Open circles: data from monkey PI; closed circles: data from monkey RHO.

practice continued, short latency saccades occurred (Fig. 3B) and started to form a second group of data points. In a final state of training, one sees the two different groups of data (Fig. 3C), but there is almost no correlation any more because the long latency reach movements have disappeared. The plots of Fig. 3 have been selected, because they show most clearly, how closely related in time the two movements can be. Not only was there a high correlation (coefficient = 0.9) but also the slope of a line through the data was very close to 1 and the position of the line was almost the same for all 3 data sets from the 3 different sessions. In conclusion: (i) The mean values of the eye and reach reaction time vary in a 1:1 relation if one compares express saccades in the gap paradigm with regular saccades in the overlap paradigm (Fig. 2). (ii) There exists a close correlation between SRT and G R T in the overlap paradigm on a trial-by-trial basis (Fig. 3). These two results are in agreement with the hypothesis of a c o m m o n m o t o r c o m m a n d generator. However, the short latency saccades obtained in the overlap paradigm, having only slightly longer reaction times than the express saccades, are followed by reach movements of longer reaction time. This observation rejects the hypothesis of a c o m m o n c o m m a n d generator. Instead, we suggest that the eye and reach movements are prepared independently

131

PI 2078 6. session (each session N=200)

SRT/ms 500. 400

A

300 200 100SRT/ms 500-

/~,

,I,h_ . . . . . . . . . 300

100

500

700

GRT/ms PI 2083 10.session

400300-

B

200 100i

SRT/ms

100

m

3oo

i

500

70o

500-

GRT/ms PI 2088 21.session

400 300

C

200 I. 100 " -

10o

300

500

' 700

' GRT/ms

Fig. 3. Correlograms from three selected sessions of monkey PI. Overlap paradigm. The straight line of slope 1 is the same for all three plots as well as for that of Fig. 2. Histograms of the saccadic reaction times are plotted along the vertical axis, those of the reach reaction times along the horizontal axis.

and in parallel, with the exception of a final coordination process which is dominated by the saccade system: it synchronized the two motor commands only in cases where the preparation of the saceades takes longer than the preparation of the reach movement. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) Sonderforschungsbereich 'Hirnforschung und Sinnesphysiology' (SFB 70, Tp B7). 1 Bach, M., Bouis, D. and Fischer, B., An accurate and linear infrared oculometer, J. Neurosci. Methods, 9 (1983) 9-14. 2 Biguer, B., Jeannerod, M. and Prablanc, C., The coordination of eye, head, and arm movements during reaching at a single visual target, Exp. Brain Res., 46 (1982) 301-304.

132 3 Fischer, B. and Boch, R., Saccadic eye movements after extremely short reaction times in the monkeys, Brain Res., 260 (1983) 21-26. 4 Fischer, B., Boch, R. and Ramsperger, E., Express-saccades of the monkey: effects of daily training on probability of occurrence and reaction time, Exp. Brain Res., 55 (1984) 232- 242. 5 Fischer, B. and Ramsperger, E., Human express-saccades: extremely short reaction times of goal-directed eye movements, Exp. Brain Res., 57 (1984) 191 195. 6 Fischer, B. and Ramsperger, E., Human express-saccades: effects of daily practice and randomization. Hum. Neurobiol., in press. 7 Poggio, G.F. and Fischer, B., Binocular interaction and depth sensitivity of striate and prestriate cortical neurons of the behaving rhesus monkey, J. Neurophysiol., 40 (1977) 1392-1405.