Epithelial-mesenchymal transition (EMT) is able to drive metastasis during progression of multiple cancer types, including non-small cell lung cancer (NSCLC). As resistance to immunotherapy has been associated with EMT and immune exclusion in melanoma, it is important to understand alterations to T-cell infiltration and the tumor microenvironment during EMT in lung adenocarcinoma and squamous cell carcinoma. We conducted an integrated analysis of the immune landscape in NSCLCs through EMT scores derived from a previously established 16 gene signature of canonical EMT markers. EMT was associated with exclusion of immune cells critical in the immune response to cancer, with significantly lower infiltration of CD4 T-cells in lung adenocarcinoma and CD4/CD8 T-cells in squamous cell carcinoma. EMT was also associated with increased expression of multiple immunosuppressive cytokines, including IL-10 and TGF-β. Furthermore, overexpression of targetable immune checkpoints, such as CTLA-4 and TIM-3 were associated with EMT in both NSCLCs. An association may exist between immune exclusion and EMT in NSCLC. Further investigation is merited as its mechanism is not completely understood and a better understanding of this association could lead to the development of biomarkers that could accurately predict response to immunotherapy.

Many anatomical factors, such as bone geometry and muscle condition, interact to affect human movements. This work aims to build a comprehensive musculoskeletal model and its control system that reproduces realistic human movements driven by muscle contraction dynamics. The variations in the anatomic model generate a spectrum of human movements ranging from typical to highly stylistic movements. To do so, we discuss scalable and reliable simulation of anatomical features, robust control of under-actuated dynamical systems based on deep reinforcement learning, and modeling of pose-dependent joint limits. The key technical contribution is a scalable, two-level imitation learning algorithm that can deal with a comprehensive full-body musculoskeletal model with 346 muscles. We demonstrate the predictive simulation of dynamic motor skills under anatomical conditions including bone deformity, muscle weakness, contracture, and the use of a prosthesis. We also simulate various pathological gaits and predictively visualize how orthopedic surgeries improve post-operative gaits.

The dorsomedial frontal cortex (DMFC) of monkeys has been implicated in mediating visually guided saccadic eye movements. The purpose of this study was to determine whether the DMFC has a topographic map coding final eye position, and to ascertain whether this region subserves the maintenance of eye position. The DMFC was stimulated electrically while monkeys fixated a target presented somewhere in visual space. A series of parametric tests was conducted to ascertain the best stimulation parameters to evoke saccades. Electrical stimulation typically produced contraversive saccades that converged onto a region of space, the termination zone. For some stimulation sites, however, stimulation produced ipsiversive saccades. This occurred when the termination zone was located straight ahead of the monkey. Convergence onto an orbital position was never observed during stimulation of the frontal eye fields (FEF), stimulation of which evoked fixed-vector saccades. The latency to evoke a saccade from the DMFC varied with fixation position, such that it increased monotonically the closer the fix spot was to the termination zone. Moreover, the probability of evoking a saccade from the DMFC decreased the closer the fix spot was to the termination zone. The latency for evoking a saccade and the probability of evoking a saccade from the FEF did not vary with fixation position. Horizontal head movements were not evoked from the DMFC while a monkey fixated targets presented in different positions of visual space. Moveover, changing the position of the head with respect to the body did not change the location of a termination zone with respect to the head. The DMFC was found to contain a topographic coding of termination zones, with rostral sites representing zones in extreme contralateral visual space, and caudal sites representing zones straight ahead or ipsilaterally. Furthermore, lateral sites represented zones in upper visual space, whereas medial sites represented zones in lower visual space. Once the eyes were positioned within a termination zone, further stimulation fixed the gaze and inhibited visually evoked saccades. Following release from inhibition, which occurred shortly after the end of stimulation, the saccades reached the visual target accurately. This shows that the stimulation delayed the execution of the saccades without actually aborting their execution. We conclude that the DMFC contains a map representing eye position in craniotopic coordinates, and we argue that this map is utilized to maintain eye position.