Linear visuomotor transformations in midbrain superior colliculus control saccadic eye-movements

A spiking neural network model of the Superior Colliculus that is robust to changes in the spatial-temporal input

Previous studies have indicated that the location of a large neural population in the Superior Colliculus (SC) motor map specifies the amplitude and direction of the saccadic eye-movement vector, while the saccade trajectory and velocity profile are encoded by the population firing rates. We recently proposed a simple spiking neural network model of the SC motor map, based on linear summation of individual spike effects of each recruited neuron, which accounts for many of the observed properties of SC cells in relation to the ensuing eye movement. However, in the model, the cortical input was kept invariant across different saccades. Electrical microstimulation and reversible lesion studies have demonstrated that the saccade properties are quite robust against large changes in supra-threshold SC activation, but that saccade amplitude and peak eye-velocity systematically decrease at low input strengths. These features were not accounted for by the linear spike-vector summation model. Here we show that the model’s input projection strengths and intra-collicular lateral connections can be tuned to generate saccades and neural spiking patterns that closely follow the experimental results.

Speed-accuracy tradeoffs influence the main sequence of saccadic eye movements

Several studies have proposed that an optimal speed-accuracy tradeoff underlies the stereotyped relationship between amplitude, duration and peak velocity of saccades (main sequence). To test this theory, we asked 8 participants to make saccades to Gaussian-blurred spots and manipulated the task's accuracy constraints by varying target size (1, 3, and 5°). The largest targets indeed yielded more endpoint scatter (and lower gains) than the smallest targets, although this effect subsided with target eccentricity. The main sequence depended on several interacting factors: saccade latency, saccade gain and target size. Early saccades, which were faster than amplitude-matched late saccades, followed the target-size dependency one would expect from a speed-accuracy tradeoff process. They had higher peak velocities and shorter durations for larger targets than for smaller targets. For late saccades, however, the opposite was found. Deviations from the main sequence also covaried with saccade gain, in line with the idea that motor noise underlies part of the endpoint variability. Thus, our data provide partial evidence that the saccadic system weighs the detrimental effects of motor noise on saccade accuracy against movement duration and speed, but other factors also modulate the kinematics. We discuss the possible involvement of parallel saccade pathways to account for our findings.

Frontal eye field inactivation alters the readout of superior colliculus activity for saccade generation in a task-dependent manner

Saccades require a spatiotemporal transformation of activity between the intermediate layers of the superior colliculus (iSC) and downstream brainstem burst generator. The dynamic linear ensemble-coding model (Goossens and Van Opstal 2006) proposes that each iSC spike contributes a fixed mini-vector to saccade displacement. Although biologically-plausible, this model assumes cortical areas like the frontal eye fields (FEF) simply provide the saccadic goal to be executed by the iSC and brainstem burst generator. However, the FEF and iSC operate in unison during saccades, and a pathway from the FEF to the brainstem burst generator that bypasses the iSC exists. Here, we investigate the impact of large yet reversible inactivation of the FEF on iSC activity in the context of the model across four saccade tasks. We exploit the overlap of saccade vectors generated when the FEF is inactivated or not, comparing the number of iSC spikes for metrically-matched saccades. We found that the iSC emits fewer spikes for metrically-matched saccades during FEF inactivation. The decrease in spike count is task-dependent, with a greater decrease accompanying more cognitively-demanding saccades. Our results show that FEF integrity influences the readout of iSC activity in a task-dependent manner. We propose that the dynamic linear ensemble-coding model be modified so that FEF inactivation increases the gain of a readout parameter, effectively increasing the influence of a single iSC spike. We speculate that this modification could be instantiated by FEF and iSC pathways to the cerebellum that could modulate the excitability of the brainstem burst generator.

Double Stimulation in a Spiking Neural Network Model of the Midbrain Superior Colliculus

The midbrain superior colliculus (SC) is a crucial sensorimotor interface in the generation of rapid saccadic gaze shifts. For every saccade it recruits a large population of cells in its vectorial motor map. Supra-threshold electrical microstimulation in the SC reveals that the stimulated site produces the saccade vector specified by the motor map. Electrically evoked saccades (E-saccades) have kinematic properties that strongly resemble natural, visual-evoked saccades (V-saccades), with little influence of the stimulation parameters. Moreover, synchronous stimulation at two sites yields eye movements that resemble a weighted vector average of the individual stimulation effects. Single-unit recordings have indicated that the SC population acts as a vectorial pulse generator by specifying the instantaneous gaze-kinematics through dynamic summation of the movement effects of all SC spike trains. But how to reconcile the a-specific stimulation pulses with these intricate saccade properties? We recently developed a spiking neural network model of the SC, in which microstimulation initially activates a relatively small set of (~50) neurons around the electrode tip, which subsequently sets up a large population response (~5,000 neurons) through lateral synaptic interactions. Single-site microstimulation in this network thus produces the saccade properties and firing rate profiles as seen in single-unit recording experiments. We here show that this mechanism also accounts for many results of simultaneous double stimulation at different SC sites. The resulting E-saccade trajectories resemble a weighted average of the single-site effects, in which stimulus current strength of the electrode pulses serve as weighting factors. We discuss under which conditions the network produces effects that deviate from experimental results.

Parallel activation of prospective motor plans during visually-guided sequential saccades

Behavioural evidences suggest that sequential saccades to multiple stimuli are planned in parallel. However, it remains unclear whether such parallel programming reflects concurrent processing of goals or whether multiple motor plans coexist, unfolding subsequently during execution. Here we use midway saccades, directed at intermediate locations between two targets, as a probe to address this question in a novel double-step adaptation task. The task consisted of trials where subjects had to follow the appearance of two targets presented in succession with two sequential saccades. In some trials, the second target predictably jumped to a new location during the second saccade. Initially, the second saccade was aimed at the final target's location before the jump. As subjects adapted to the target jump, saccades were aimed to the second target's new location. We tested whether the spatial distribution of midway saccades could be explained as an interaction between two concurrent saccade goals, each directed at the two target locations, or between the initial motor plan to the first target location and a prospective motor plan directed from the initial to the final target location. A shift in the midway saccades' distribution towards the jumped location of the second target following adaptation indicated that the brain can make use of prospective motor plans to guide sequential eye movements. Furthermore, we observed that the spatiotemporal pattern of endpoints of midway saccades can be well explained by a motor addition model. These results provide strong evidence of parallel activation of prospective motor plans during sequential saccades.

Saccadic and Postsaccadic Disconjugacy in Zebrafish Larvae Suggests Independent Eye Movement Control

Spontaneous eye movements of zebrafish larvae in the dark consist of centrifugal saccades that move the eyes from a central to an eccentric position and postsaccadic centripetal drifts. In a previous study, we showed that the fitted single-exponential time constants of the postsaccadic drifts are longer in the temporal-to-nasal (T->N) direction than in the nasal-to-temporal (N->T) direction. In the present study, we further report that saccadic peak velocities are higher and saccadic amplitudes are larger in the N->T direction than in the T->N direction. We investigated the underlying mechanism of this ocular disconjugacy in the dark with a top-down approach. A mathematic ocular motor model, including an eye plant, a set of burst neurons and a velocity-to-position neural integrator (VPNI), was built to simulate the typical larval eye movements in the dark. The modeling parameters, such as VPNI time constants, neural impulse signals generated by the burst neurons and time constants of the eye plant, were iteratively adjusted to fit the average saccadic eye movement. These simulations suggest that four pools of burst neurons and four pools of VPNIs are needed to explain the disconjugate eye movements in our results. A premotor mechanism controls the synchronous timing of binocular saccades, but the pools of burst and integrator neurons in zebrafish larvae seem to be different (and maybe separate) for both eyes and horizontal directions, which leads to the observed ocular disconjugacies during saccades and postsaccadic drifts in the dark.

Response normalization in the superficial layers of the superior colliculus as a possible mechanism for saccadic averaging

How does the brain decide where to look? Neuronal networks within the superior colliculus (SC) encode locations of intended eye movements. When faced with multiple targets, the relative activities of neuronal populations compete for the selection of a saccade. However, the computational principles underlying saccadic choices remain poorly understood. We used voltage imaging of slices of rat SC to record circuit dynamics of population responses to single- and dual-site electrical stimulation to begin to reveal some of the principles of how populations of neurons interact. Stimulation of two distant sites simultaneously within the SC produced two distinct peaks of activity, whereas stimulation of two nearby sites simultaneously exhibited a single, merged peak centered between the two sites. The distances required to produce merged peaks of activity corresponded to target separations that evoked averaging saccades in humans performing a corresponding dual target task. The merged activity was well accounted for by a linear weighed summation and a divisive normalization of the responses evoked by the single-site stimulations. Interestingly, the merging of activity occurred within the superficial SC, suggesting a novel pathway for saccadic eye movement choice.

Overlapping structures in sensory-motor mappings

This paper examines a biologically-inspired representation technique designed for the support of sensory-motor learning in developmental robotics. An interesting feature of the many topographic neural sheets in the brain is that closely packed receptive fields must overlap in order to fully cover a spatial region. This raises interesting scientific questions with engineering implications: e.g. is overlap detrimental? does it have any benefits? This paper examines the effects and properties of overlap between elements arranged in arrays or maps. In particular we investigate how overlap affects the representation and transmission of spatial location information on and between topographic maps. Through a series of experiments we determine the conditions under which overlap offers advantages and identify useful ranges of overlap for building mappings in cognitive robotic systems. Our motivation is to understand the phenomena of overlap in order to provide guidance for application in sensory-motor learning robots.