[1] 8w Inactivation of the central thalamus delays self-timed saccades

[1] 62w The central thalamus transmits corollary discharge signals for eye movement control, but its role in eye movement generation remains uncertain. Inactivation of the paralaminar part of the ventrolateral thalamus delayed the initiation of contraversive saccades, particularly during a new memory-guided saccade task that required self-triggering of the movement. The results suggest that signals through the thalamus regulate the timing of self-initiated saccades.

[1] 357w Monkeys (n ¼ 3) were trained to perform the same three saccade tasks. In the visually guided saccade task (Fig. 1a, left), a central fixation point was extinguished at the time of peripheral target onset, and monkeys were required to make a saccade to the target within 400 ms. In the memory-guided saccade task (Fig. 1a, middle), a peripheral target ('cue') was flashed briefly (for 200 ms) during fixation. Monkeys were required to maintain fixation for an additional 1,000-1,200 ms and to make a saccade to the cue location immediately after the offset of the fixation point (Supplementary Fig. 1 online). Because of a delay between the offsets of the cue and the fixation point, the direction and amplitude of saccades must be programmed based on internal signals that are related to the cue location. In this task, however, the time of saccade initiation was given externally-by the offset of the fixation point. This task therefore will be referred to as an 'externally triggered' memory-guided saccade task; this is essentially identical to classic memory-guided saccade tasks described in the literature 2,3,5 . A new version of the memory-guided saccade task required self-timing (Fig. 1a, right). It began like the classic memory-guided saccade task, but the monkeys were not given an overt instruction to move; instead, they were required to make a saccade to the cue location 800-1,600 ms after the offset of the cue. The fixation point was extinguished immediately after initiation of the saccade, and the peripheral target reappeared 400 ms later and remained visible for an additional 800 ms (Supplementary Fig. 1). Monkeys were rewarded only when they generated a saccade within an 800-ms time window and when their eye position was within 51 of the peripheral target as long as it was visible. In this 'internally triggered' memory-guided saccade task, both the timing and the metrics of the saccades must be programmed based on internal signals. The location of the saccade target was 121 either to the right or the left of central fixation, and a total of six different trials (three tasks in two directions) were presented in random order within a block.

[2] 286w To examine whether signals in the central thalamus contribute to saccade generation, a small amount of a GABA agonist (muscimol hydrobromide; 5 mg ml -1 in saline; volumes ranging from 1-2 ml; mean volume ¼ 1.6 ± 0.3 ml; n ¼ 47) was injected locally (Fig. 1b; see Supplementary Methods online). The inactivation sites selected were those that had showed eye movement-related neuronal activities in separate recording experiments 6 . The monkey generated saccades within a 400-or 800-ms time window in most trials, both before and after inactivation. However, after the injection of muscimol, data for contraversive saccades became more variable and tended to show longer latencies that were statistically different from the values obtained before injection (Fig. 2a, top; Wilcoxon rank-sum test, P o 0.01). In contrast, the latency of ipsiversive saccades was shortened slightly in the internally triggered memory-guided saccade task (P o 0.01) but was not altered in the other two tasks (Fig. 2a, bottom). Inactivation effects were also evident in both the accuracy and dynamics of saccades (Fig. 2b). After inactivation, the distance between target and saccade endpoints increased significantly for all contraversive saccades and for ipsiversive externally triggered memory-guided saccades (two-tailed t-test, P o 0.01), but not for the others (P 4 0.05). This was due primarily to increased scatter, not to systematic changes in mean saccade amplitude (which ranged from -0.521 to 0.721). The peak velocity of individual saccades decreased slightly but significantly for contraversive saccades and for ipsiversive internally triggered memoryguided saccades (two-tailed t-test, P o 0.01, for the experiment shown in Fig. 2. The data from individual injection sites on the accuracy, amplitude and velocity of saccades are shown in Supplementary Figs. 2, 3 and 4 online, respectively).

[3] 546w Similar effects of inactivation were found for multiple sites in four thalami of the three monkeys. Among 47 sites tested, inactivation at 30 sites significantly altered the latency of saccades in one or both directions (Wilcoxon rank-sum test, P o 0.01). The amounts of muscimol injected into the effective and non-effective sites were not statistically different from one another (two-tailed t-test, P ¼ 0.86). In the internally triggered task, thalamic inactivation effectively lengthened the latency of contraversive saccades (Wilcoxon rank-sum test, P o 0.01) for 26 of 47 sites, whereas the latency of ipsiversive saccades was unchanged (17 sites) or even decreased (9 sites, P o 0.01) (see Fig. 3a). In the other tasks, the effect of inactivation was much smaller, but it was statistically significant for contraversive saccades in most sites (23 and 19 sites for the externally triggered task and the visually guided task, respectively) and for ipsiversive saccades in some sites (7 and 4 sites, respectively). On average, the effect on latency was greater for contraversive saccades and was greatest for the internally triggered memory-guided saccades (Fig. 3b). In all three protocols, differences in median latencies were statistically different from 0 for contraversive saccades (two-tailed t-test, P o 10 À5 ) but not for ipsiversive saccades (P 4 0.05). Injection of saline into the effective sites (n ¼ 7; mean volume ¼ 1.8 ± 0.2 ml) did not result in a significant change in latency for any of the six different tasks (P 4 0.1). A one-way analysis of variance (ANOVA) showed that the changes in median latencies after muscimol injection were significantly different across tasks for contraversive saccades (P o 0.01) but not for ipsiversive saccades (P ¼ 0.09). For contraversive saccades, a post-hoc comparison (Scheffe ´) revealed that the values for the internally triggered task were statistically different from those for the other two tasks (P o 10 À5 ) but that the values for these two tasks were not different from each other (P 4 0.05). Thus, thalamic inactivation did not produce a fixed shift in saccade timing but, instead, had more profound effects on self-timed, contraversive saccades. The present results showed that the latency of contraversive saccades was altered by inactivating the rostral portion of the caudal division of the ventrolateral nucleus and the adjacent area X of the thalamus. This study also showed that inactivation of the mediodorsal thalamus causes negligible deficits in saccade generation, consistent with the previous study 7 (Fig. 1b). The inactivation effects of the ventrolateral thalamus were much more profound in tasks that required monkeys to internally determine both the goals and the timing of saccades. These findings provide new insights into the roles of the central thalamus in eye movement control. Signals in the central thalamus are used for the internal monitoring of eye movements 7 , but they do not contribute to the initiation of visually guided eye movements 6,7 (Supplementary Fig. 5 online plots injection sites in this and the previous study that examined effects on smooth pursuit 6 ). Other studies suggested that signals in the ventrolateral 8 and mediodorsal 9 thalamus link spatial information about target location and saccade goals. The present data imply an additional role of the central thalamus, one that regulates the timing of self-initiated saccades.

[1] 113w Although signals through the motor thalamus are crucial for the generation of somatic movements 1 , their roles in the generation of eye movement remain unknown. Previous studies have demonstrated that neurons in the paralaminar part of the ventrolateral thalamus modulate their activity during both the preparation and the execution of saccades [2][3][4] . To investigate the possible roles of these signals in saccade generation, the present study examined the effects of local inactivation. All experimental protocols were approved by the Animal Care and Use Committee of the Hokkaido University School of Medicine and were in accordance with the US National Institutes of Health guide for the care and use of laboratory animals.

[2] 156w The central thalamus has a pivotal role in regulating the activity in both the cerebral cortex and the basal ganglia. Given that the diameter of muscimol spread could be up to a few millimeters 10 , it is likely that the present experiments disrupted signals in multiple thalamic pathways. Inactivation of the ventrolateral thalamus and the adjacent rostral intralaminar nuclei could modulate neuronal processes within the reciprocal thalamocortical loops 11 and/or the local basal ganglia-thalamostriatal circuitry 12 . In addition, thalamic inactivation could also disturb signals from the basal ganglia and cerebellum to the cerebral cortex 13 . Previous studies have suggested that the latter pathways are involved in the timing of somatic movements and in time perception 14,15 : thus, the delay of self-timed saccades seen here possibly resulted from the inactivation of these pathways. The signals through the thalamus to the cortex might regulate the timing of decisions for internally triggered movements, including saccades.

[3] 10w Note: Supplementary information is available on the Nature Neuroscience website.