A Circuit Model of the Neural Network for Time-Optimal Control of Horizontal Saccadic Eye Movements

This dissertation research focused on computational systems neuroscience for modeling the major midbrain structures in the control of conjugate goal-directed horizontal human and monkey saccadic eye movements. Given the complexity of the saccade generator in a large-scale spiking neural network (SNN), a combination of extant behavioral, neurophysiological, and anatomical studies needs to be primarily referenced. These studies provide abundant evidence that an SNN is well suited to evoke the properties of the firing patterns of the premotor neurons during the pulse and slide phases of innervation in a saccade. However, none of the studies have presented a demonstration of the neural circuits reproducing electrophysiological responses in a network of neurons at both premotor and motor levels. This work investigated an integrative systems approach to address the challenges involved in the implementation of the saccade dynamics from the local neural circuit computations in the midbrain. In summary, the main contributions of this work are: (1) computational modeling of neural circuits in the midbrain (specifically, premotor neural sites in the paramedian pontine reticular formation and motoneurons); and (2) implementing systems models of sensory-motor integration by using a first-order time-optimal neural controller. Having dealt with the characteristics of human and monkey saccades, the extension of the neural modeling approach was examined to consider the case of human saccades with a glissade. In this context, considerable attention was paid to comparing the simulated glissades with those of the normal saccades for three different magnitudes. Computational modeling of the glissades is advantageous because it allows investigation of one of the widely-reported clinical oculomotor version dysfunctions. In particular, glissade characteristics contribute to detection of the anomalies in the saccade dynamics as a consequence of diffuse traumatic brain injury (DTBI). They also provoke further studies as to determine the origin of the damage to the midbrain in diagnosis of DTBI. Simulation findings confirmed that the anomalies are manifested due to an unplanned post-inhibitory rebound burst firing in the antagonist motoneurons as a source of coordination error in returning to tonic firing rates. The research results are demonstrated by a standalone graphical user interface that integrates all of the computational models