Frontal eye field inactivation reduces saccade preparation in the superior colliculus, but does not alter how preparatory activity relates to saccade latency

A neural correlate for saccadic reaction times (SRTs) in the gap saccade task is the level of preparatory activity in the intermediate layers of the superior colliculus (iSC) just before visual target onset: greater levels of iSC preparatory activity precede shorter SRTs. The frontal eye fields (FEF) are one likely source of such iSC preparatory activity, since FEF preparatory activity is also inversely related to SRT. To better understand the FEF’s role in saccade preparation, and the way in which such preparation relates to SRT, in two male rhesus monkeys we examined iSC preparatory activity during unilateral reversible cryogenic inactivation of the FEF. FEF inactivation increased contralesional SRTs, and lowered ipsilesional iSC preparatory activity. FEF inactivation also reduced fixation-related activity in the rostral iSC. Importantly, the distributions of SRTs generated with or without FEF inactivation overlapped, enabling us to conduct a novel population-level analyses examining iSC preparatory activity just before generation of SRT-matched saccades. These analyses revealed no change during FEF inactivation in the relationship between iSC preparatory activity and SRT-matched saccades across a range of SRTs, even for the occasional express saccade. Thus, while our results emphasize that the FEF has an overall excitatory influence on preparatory activity in the iSC, the communication between the iSC and downstream oculomotor brainstem is unaltered for SRT-matched saccades, suggesting that the integration of preparatory and visual signals in the SC just before saccade initiation is largely independent of the FEF for saccades generated in this task. Significance statement How does the brain decide when to move? Here, we investigate the role of two oculomotor structures, the superior colliculus (SC) and frontal eye fields (FEF), in dictating visually-guided saccadic reaction times (SRTs). In both structures, higher levels of preparatory activity precede shorter SRTs. Here, we show that FEF inactivation increases SRTs and decreases SC preparatory activity. Surprisingly, a population-level analysis of SC preparatory activity showed a negligible impact of FEF inactivation, providing one examines SRT-matched saccades. Thus, while the FEF is one source of preparatory input to the SC, it is not a critical source, and it is not involved in the integration of preparatory activity and visual signals that precedes saccade initiation in simple visually-guided saccade tasks.

understand the FEF's role in saccade preparation, and the way in which such 48 preparation relates to SRT, in two male rhesus monkeys we examined iSC preparatory 49 activity during unilateral reversible cryogenic inactivation of the FEF. FEF inactivation 50 increased contralesional SRTs, and lowered ipsilesional iSC preparatory activity. FEF 51 inactivation also reduced fixation-related activity in the rostral iSC. Importantly, the 52 distributions of SRTs generated with or without FEF inactivation overlapped, enabling 53 us to conduct a novel population-level analyses examining iSC preparatory activity just 54 before generation of SRT-matched saccades. These analyses revealed no change 55 during FEF inactivation in the relationship between iSC preparatory activity and SRT-56 matched saccades across a range of SRTs, even for the occasional express saccade. 57 Thus, while our results emphasize that the FEF has an overall excitatory influence on 58 preparatory activity in the iSC, the communication between the iSC and downstream 59 oculomotor brainstem is unaltered for SRT-matched saccades, suggesting that the 60 integration of preparatory and visual signals in the SC just before saccade initiation is 61 largely independent of the FEF for saccades generated in this task.

63
Significance statement: 64 How does the brain decide when to move? Here, we investigate the role of two Introduction 79 The time to respond to a behavioral event can be highly variable. Even for simple 80 visually-guided saccades, saccadic reaction times (SRTs) can range from those 81 approaching the minimal sensory-to-motor delays in the case of express saccades to 82 others that are two to three times longer (Saslow 1967;Fischer et al. 1984; Paré and 83 Munoz 1996). One neural correlate of SRT variability in this task is the level of low-  The goal of the current study is to investigate how the relationship between preparatory importance of FEF integrity on iSC preparatory activity. Second, and more importantly, 99 this approach will allow us to examine the relationship between iSC preparatory activity 100 and SRT when the oculomotor system is in an altered state. FEF inactivation increases 101 SRTs, but how such increased SRTs relate to preparatory iSC activity is unknown. Are 102 SRTs increased simply because FEF inactivation decreases preparatory activity? If so, 103 then the relationship between iSC preparatory activity and SRT would simply shift to 104 higher SRTs, so that preparatory activity would remain the same for saccades of 105 matched SRTs. Alternatively, perhaps more preparatory activity is required to generate 106 a saccade of a given SRT, perhaps due to the loss of FEF signaling along parallel 107 circuits to the oculomotor brainstem that bypass the iSC (Raybourn and Keller 1977;108 given the reciprocal relationship between fixation-related and preparatory activity in the 110 rostral and caudal iSC, respectively (Dorris and Munoz 1995;Dorris et al. 1997;Munoz 111 and Istvan 1998; Munoz  across the recorded population for SRT-matched saccades. We found that FEF 125 inactivation increased SRTs, and decreased both preparatory activity in the caudal iSC 126 and fixation-related activity in the rostral iSC. However, we found that the level of iSC 127 preparatory activity did not change when we examined the subset of SRT-matched 128 saccades. These results show that the relationship between iSC preparatory activity 129 and SRT is unaltered during FEF inactivation, so that a saccade of the same SRT, 130 including those generated at express-saccade latencies, can be generated providing 131 other non-FEF sources compensate for the loss FEF signaling.   An experimental dataset consisted of pre-cooling, peri-cooling, and post-cooling 167 sessions, with each session containing 60-120 correct trials (requiring between 180-360 168 trials total). After the pre-cooling session, the cooling pumps were turned on allowing 169 the flow of chilled methanol through the lumen of the cryoloops. The peri-cooling 170 session was initiated when cryoloop temperature reached and stayed stable at 3°C.
Once sufficient data was collected for the peri-cooling session, cooling pumps were 172 turned off, which allowed the cryoloop temperature to rapidly return towards body 173 temperature. When cryoloop temperature exceeded 35˚C, the post-cooling session was 174 initiated. To control for time dependent factors like satiation, we pooled data from pre-175 cooling and post-cooling session and termed this the FEFwarm condition. Data 176 collected during FEF inactivation was termed as coming from the FEFcool condition.  study employing an immediate response task similar to ours did not find any correlation 202 between preparatory activity with saccade metric or kinematic (Basso and Wurtz 1998). 203 Also, we (in this study) and others have also shown that the overall level of preparatory 204 activity for ipsi-and contraversive saccades is not different when their probability of

298
We report single unit activity from the left iSC of 2 monkeys (DZ and OZ) while the left 299 FEF was reversibly inactivated by cryogenic means, when animals were engaged in the   ipsiversive SRTs (Fig. 1A, C). The failure to observe ipsiversive SRT increases may be 318 due to differences in target configuration and behavioral task (e.g., 2 potential targets Wilcoxon rank sum test) and across all sessions also showed only a small but 329 significant increase from 1.04±0.30° to 1.27±0.34° (p=2.42e-08, z-val=-5.5785;

330
Wilcoxon sign rank test). Next, we compared peak saccade velocity across session and              The SRT ranges studied in the above population analysis did not include express 538 saccades, which we defined as SRTs between 70-120ms in accordance to previous 539 studies (Schiller et al. 1987;Paré and Munoz 1996;Dorris et al. 1997;Sparks et al. 2000). As shown by the SRT distributions in Fig. 5A & C, both monkeys still generated 541 express saccades during FEF inactivation, although their incidence was reduced. To 542 analyze such rare saccades, we adopted the following SRT-matching logic. First, we 543 identified those rare trials where a contraversive express saccade was generated during 544 FEF inactivation. Using the SRT of this "FEF cool" express saccade, we then searched 545 the FEF warm data recorded from the same iSC neuron for trials where a matching 546 "FEF warm" contraversive express saccade was generated with a SRT within ± 3ms. In summary, express saccades during FEF inactivation, although reduced in overall 561 incidence, were associated with higher baseline activity but an unchanged final level of 562 activity at the end of PREP epoch. lateral intraparietal area (LIP)) or subcortical (e.g., basal ganglia) inputs to the iSC, with 598 the net effect of decreasing excitatory inputs to the iSC. Regardless of the exact 599 mechanism, our core findings of reduced preparatory-and fixation-related activity in the 600 caudal and rostral iSC, respectively, relates to our recent work showing that FEF 601 inactivation reduces visual-, delay-period, and saccade-related activity (Peel et al. 2017). This recent work also correlated increases in SRT during FEF inactivation with 603 delays in the onset of saccade-related accumulation during a delayed-saccade task. In 604 the current study we did not find any changes in the onset of preparatory activity during 605 FEF inactivation. As outlined below, our data suggests that SRT in the gap saccade 606 task is largely dictated by the magnitude of preparatory activity attained just before 607 arrival of the visual transient, rather than the time at which such activity starts to 608 accumulate.