1. There are indications that the execution of behavioral sequences in
volves the basal ganglia. In this study we examined the role of the ca
udate nucleus in the construction, storage, and execution of spatial p
lans. 2. Two monkeys (Macaca mulatta) were trained to perform sequence
s of saccades and arm movements. The animals had to remember the order
of illumination, variable from one sequence to another, of three fixe
d spatial targets. After a delay, they had to visually orient toward,
and press each target in the same order. Six different sequences were
executed on the basis of the order of illumination of the targets. Sin
gle cell activity was recorded from the four caudate nuclei of the two
monkeys. 3. Neural activity was analyzed in each sequence during 10 d
ifferent periods: the instruction period in which the targets were ill
uminated, the three orientation periods toward the different targets,
the three postsaccadic periods, and the three periods of target pressi
ng. Statistical comparisons were made to detect differences between th
e different sequences with respect to activity in each period (sequenc
e specificity). 4. A total of 2,100 neurons were studied, of which 387
were task related. The task-related cells were found in both the head
and the body of the caudate nucleus. 5. During central fixation, anti
cipatory activity (n = 81) preceded onset of specific events. Four gro
ups were considered: 1) neurons (n = 46) anticipating offset of the ce
ntral fixation point, 2) neurons (n = 7) anticipating the illumination
of any target, regardless of its spatial position or order of present
ation (rank), 3) neurons (n = 17) anticipating the illumination of the
first target, regardless of its spatial position, and 4) neurons (n =
11) anticipating the illumination of a given target, regardless of it
s rank. 6. Phasic visual responses to target onset were observed in 48
cells. The cells responded primarily to the contralateral and upper t
argets. In a majority (n = 35), visual responses were modulated by the
rank of the target(s). Many cells (n = 20) responded only if the corr
esponding target was first; other cells responded only if the target w
as second or if it had complex time relationships with the other targe
ts. 7. The responses of the cells to the same instruction stimuli repe
ated twice in a row, and under the condition that the animal did not b
ehaviorally use the first instruction in between, were tested. More th
an one-third of the tested cells (n = 14) did not respond, or responde
d very weakly, to the second instruction. 8. Visuomotor activity of in
dividual caudate cells depended on the arm used to perform the target-
pressing task. 9. Orientation-related activity (n = 110) was present w
hen the animal made a saccade or a saccade and an arm movement toward
a target. Many neurons (n = 48) were sequence specific. Their activity
did not depend only on the spatial parameters of the current orientat
ion, but also on the location and existence of the next target, i.e.,
on the characteristics of the spatial plan at the time of the orientat
ion. 10. Postsaccadic activation was observed in 103 cells. In a major
ity (n = 80), the activation occurred after saccades toward the upper
and/or the ipsilateral target(s). Many neurons (n = 45) were sequence
specific. Their activity appears as a cognitive phenomenon directly re
lated to the selection of the successive targets. 11. Activity associa
ted with the target pressing was observed in 77 cells. Thirty-eight (3
8) cells displayed a sequence specificity characterized by a preferenc
e for the press at a given rank. The rank specificity is discussed in
terms of anticipation and control of the visual feedbacks of the targe
t presses. 12. A phasic activity (n = 21) followed an incorrect press.
13. Activity related to reward (n = 60) was a phasic discharge that o
ccurred after the monkey's last target press and until the reward, and
/or after the reward. 14. Taken together, the data are compatible with
the hypothesis that the caudate nucleus participates in the construct
ion, but not in the storage, of the spatial plans. Under control of th
e spatial plans, the caudate nucleus would be instrumental in running
the sequences. It would select the successive targets for the oculomot
or apparatus as sequences progress. It would participate in the execut
ion of the saccades and arm movements and would control the environmen
tal changes that action generates.