A Bayesian Developmental Approach to Robotic Goal-Based Imitation Learning

A fundamental challenge in robotics today is building robots that can learn new skills by observing humans and imitating human actions. We propose a new Bayesian approach to robotic learning by imitation inspired by the developmental hypothesis that children use self-experience to bootstrap the process of intention recognition and goal-based imitation. Our approach allows an autonomous agent to: (i) learn probabilistic models of actions through self-discovery and experience, (ii) utilize these learned models for inferring the goals of human actions, and (iii) perform goal-based imitation for robotic learning and human-robot collaboration. Such an approach allows a robot to leverage its increasing repertoire of learned behaviors to interpret increasingly complex human actions and use the inferred goals for imitation, even when the robot has very different actuators from humans. We demonstrate our approach using two different scenarios: (i) a simulated robot that learns human-like gaze following behavior, and (ii) a robot that learns to imitate human actions in a tabletop organization task. In both cases, the agent learns a probabilistic model of its own actions, and uses this model for goal inference and goal-based imitation. We also show that the robotic agent can use its probabilistic model to seek human assistance when it recognizes that its inferred actions are too uncertain, risky, or impossible to perform, thereby opening the door to human-robot collaboration.     

Strain Dependent Variation of Immune Responses to A. fumigatus: Definition of Pathogenic Species

For over a century microbiologists and immunologist have categorized microorganisms as pathogenic or non-pathogenic species or genera. This definition, clearly relevant at the strain and species level for most bacteria, where differences in virulence between strains of a particular species are well known, has never been probed at the strain level in fungal species. Here, we tested the immune reactivity and the pathogenic potential of a collection of strains from Aspergillus spp, a fungus that is generally considered pathogenic in immuno-compromised hosts. Our results show a wide strain-dependent variation of the immune response elicited indicating that different isolates possess diverse virulence and infectivity. Thus, the definition of markers of inflammation or pathogenicity cannot be generalized. The profound understanding of the molecular mechanisms subtending the different immune responses will result solely from the comparative study of strains with extremely diverse properties.     

Strategies for signal amplification in nucleic acid detection

Many aspects of molecular genetics necessitate the detection of nucleic acid sequences. Current approaches involving target amplification (in situ PCR, Primed in situ Labeling, Self-Sustained Sequence Replication, Strand Displacement Amplification), probe amplification (Ligase Chain Reaction, Padlock Probes, Rolling Circle Amplification) and signal amplification (Tyramide Signal Amplification, Branched DNA Amplification) are summarized in the present review, together with their advantages and limitations.     

Simplified procedure for measurement of serum dehydroepiandrosterone and its sulfate with gas chromatography–ion trap mass spectrometry and selected reaction monitoring

Dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEAS) are endogenous steroids that have recently been widely publicized as potential treatments for many disorders. This paper describes a gas chromatographic–ion trap mass spectrophotometric assay with selected reaction monitoring for measurement of DHEA and DHEAS levels. The hormones and internal standard (5-androsten-3β-ol-16-one methyl ester) are extracted from serum with Oasis solid-phase extraction tubes. The extracted steroids are dissolved in methanol and injected into a Finnigan GCQ ion trap mass spectrometer. In the selected reaction mode, both DHEA and DHEAS can be identified and quantified in a single injection. No derivatization or expensive deuterated internal standards are required.     

β-Amyloid (Aβ) Oligomers Impair Brain-derived Neurotrophic Factor Retrograde Trafficking by Down-regulating Ubiquitin C-terminal Hydrolase,UCH-L1

We previously found that BDNF-dependent retrograde trafficking is impaired in AD transgenic mouse neurons. Utilizing a novel microfluidic culture chamber, we demonstrate that Aβ oligomers compromise BDNF-mediated retrograde transport by impairing endosomal vesicle velocities, resulting in impaired downstream signaling driven by BDNF/TrkB, including ERK5 activation, and CREB-dependent gene regulation. Our data suggest that a key mechanism mediating the deficit involves ubiquitin C-terminal hydrolase L1 (UCH-L1), a deubiquitinating enzyme that functions to regulate cellular ubiquitin. Aβ-induced deficits in BDNF trafficking and signaling are mimicked by LDN (an inhibitor of UCH-L1) and can be reversed by increasing cellular UCH-L1 levels, demonstrated here using a transducible TAT-UCH-L1 strategy. Finally, our data reveal that UCH-L1 mRNA levels are decreased in the hippocampi of AD brains. Taken together, our data implicate that UCH-L1 is important for regulating neurotrophin receptor sorting to signaling endosomes and supporting retrograde transport. Further, our results support the idea that in AD, Aβ may down-regulate UCH-L1 in the AD brain, which in turn impairs BDNF/TrkB-mediated retrograde signaling, compromising synaptic plasticity and neuronal survival.     

Glioma IL13Rα2 Is Associated with Mesenchymal Signature Gene Expression and Poor Patient Prognosis

A major challenge for successful immunotherapy against glioma is the identification and characterization of validated targets. We have taken a bioinformatics approach towards understanding the biological context of IL-13 receptor α2 (IL13Rα2) expression in brain tumors, and its functional significance for patient survival. Querying multiple gene expression databases, we show that IL13Rα2 expression increases with glioma malignancy grade, and expression for high-grade tumors is bimodal, with approximately 58% of WHO grade IV gliomas over-expressing this receptor. By several measures, IL13Rα2 expression in patient samples and low-passage primary glioma lines most consistently correlates with the expression of signature genes defining mesenchymal subclass tumors and negatively correlates with proneural signature genes as defined by two studies. Positive associations were also noted with proliferative signature genes, whereas no consistent associations were found with either classical or neural signature genes. Probing the potential functional consequences of this mesenchymal association through IPA analysis suggests that IL13Rα2 expression is associated with activation of proinflammatory and immune pathways characteristic of mesenchymal subclass tumors. In addition, survival analyses indicate that IL13Rα2 over-expression is associated with poor patient prognosis, a single gene correlation ranking IL13Rα2 in the top ~1% of total gene expression probes with regard to survival association with WHO IV gliomas. This study better defines the functional consequences of IL13Rα2 expression by demonstrating association with mesenchymal signature gene expression and poor patient prognosis. It thus highlights the utility of IL13Rα2 as a therapeutic target, and helps define patient populations most likely to respond to immunotherapy in present and future clinical trials.