There is growing consensus that accurate and efficient face recognition is mediated by a neural circuit composed of a posterior "core" and an anterior "extended" set of regions. Here, we characterize the distributed face network in human individuals with congenital prosopagnosia (CP)-a lifelong impairment in face processing-relative to that of matched controls. Using functional magnetic resonance imaging, we first uncover largely normal activation patterns in the posterior core face patches in CP. We also document normal activity of the amygdala (emotion processing) as well as normal or even enhanced functional connectivity between the amygdala and the core regions. Critically, in the same individuals, activation of the anterior temporal cortex (identity processing) is reduced and connectivity between this region and the posterior core regions is disrupted. The dissociation between the neural profiles of the anterior temporal lobe and amygdala was evident both during a task-related face scan and during a resting state scan, in the absence of visual stimulation. Taken together, these findings elucidate selective disruptions in neural circuitry in CP and offer an explanation for the known differential difficulty in identity versus emotional expression recognition in many individuals with CP.

MicroRNAs modulate cellular phenotypes by inhibiting expression of mRNA targets. In this study, we have shown that the muscle-specific microRNAs miR-133a-1 and miR-133a-2 are essential for multiple facets of skeletal muscle function and homeostasis in mice. Mice with genetic deletions of miR-133a-1 and miR-133a-2 developed adult-onset centronuclear myopathy in type II (fast-twitch) myofibers, accompanied by impaired mitochondrial function, fast-to-slow myofiber conversion, and disarray of muscle triads (sites of excitation-contraction coupling). These abnormalities mimicked human centronuclear myopathies and could be ascribed, at least in part, to dysregulation of the miR-133a target mRNA that encodes dynamin 2, a GTPase implicated in human centronuclear myopathy. Our findings reveal an essential role for miR-133a in the maintenance of adult skeletal muscle structure, function, bioenergetics, and myofiber identity; they also identify a potential modulator of centronuclear myopathies.

Researchers from multiple fields have sought to understand how sex moderates human social behavior. While over 50 years of research has revealed differences in cooperation behavior of males and females, the underlying neural correlates of these sex differences have not been explained. A missing and fundamental element of this puzzle is an understanding of how the sex composition of an interacting dyad influences the brain and behavior during cooperation. Using fNIRS-based hyperscanning in 111 same- and mixed-sex dyads, we identified significant behavioral and neural sex-related differences in association with a computer-based cooperation task. Dyads containing at least one male demonstrated significantly higher behavioral performance than female/female dyads. Individual males and females showed significant activation in the right frontopolar and right inferior prefrontal cortices, although this activation was greater in females compared to males. Female/female dyad’s exhibited significant inter-brain coherence within the right temporal cortex, while significant coherence in male/male dyads occurred in the right inferior prefrontal cortex. Significant coherence was not observed in mixed-sex dyads. Finally, for same-sex dyads only, task-related inter-brain coherence was positively correlated with cooperation task performance. Our results highlight multiple important and previously undetected influences of sex on concurrent neural and behavioral signatures of cooperation.

Clustered regularly interspaced short palindromic repeats (CRISPR)-associated (Cas)9 genomic editing has revolutionized the generation of mutant animals by simplifying the creation of null alleles in virtually any organism. However, most current approaches with this method require zygote injection, making it difficult to assess the adult, tissue-specific functions of genes that are widely expressed or which cause embryonic lethality when mutated. Here, we describe the generation of cardiac-specific Cas9 transgenic mice, which express high levels of Cas9 in the heart, but display no overt defects. In proof-of-concept experiments, we used Adeno-Associated Virus 9 (AAV9) to deliver single-guide RNA (sgRNA) that targets the Myh6 locus exclusively in cardiomyocytes. Intraperitoneal injection of postnatal cardiac-Cas9 transgenic mice with AAV9 encoding sgRNA against Myh6 resulted in robust editing of the Myh6 locus. These mice displayed severe cardiomyopathy and loss of cardiac function, with elevation of several markers of heart failure, confirming the effectiveness of this method of adult cardiac gene deletion. Mice with cardiac-specific expression of Cas9 provide a tool that will allow rapid and accurate deletion of genes following a single injection of AAV9-sgRNAs, thereby circumventing embryonic lethality. This method will be useful for disease modeling and provides a means of rapidly editing genes of interest in the heart.

Abstract-Automatic decision-making approaches, such as reinforcement learning (RL), have been applied to (partially) solve the resource allocation problem adaptively in the cloud computing system. However, a complete cloud resource allocation framework exhibits high dimensions in state and action spaces, which prohibit the usefulness of traditional RL techniques. In addition, high power consumption has become one of the critical concerns in design and control of cloud computing systems, which degrades system reliability and increases cooling cost. An effective dynamic power management (DPM) policy should minimize power consumption while maintaining performance degradation within an acceptable level. Thus, a joint virtual machine (VM) resource allocation and power management framework is critical to the overall cloud computing system. Moreover, novel solution framework is necessary to address the even higher dimensions in state and action spaces.In this paper, we propose a novel hierarchical framework for solving the overall resource allocation and power management problem in cloud computing systems. The proposed hierarchical framework comprises a global tier for VM resource allocation to the servers and a local tier for distributed power management of local servers. The emerging deep reinforcement learning (DRL) technique, which can deal with complicated control problems with large state space, is adopted to solve the global tier problem. Furthermore, an autoencoder and a novel weight sharing structure are adopted to handle the high-dimensional state space and accelerate the convergence speed. On the other hand, the local tier of distributed server power managements comprises an LSTM based workload predictor and a model-free RL based power manager, operating in a distributed manner. Experiment results using actual Google cluster traces show that our proposed hierarchical framework significantly saves the power consumption and energy usage than the baseline while achieving no severe latency degradation. Meanwhile, the proposed framework can achieve the best trade-off between latency and power/energy consumption in a server cluster.