Learning and Long-Term Retention of Large-Scale Artificial Languages

Recovering discrete words from continuous speech is one of the first challenges facing language learners. Infants and adults can make use of the statistical structure of utterances to learn the forms of words from unsegmented input, suggesting that this ability may be useful for bootstrapping language-specific cues to segmentation. It is unknown, however, whether performance shown in small-scale laboratory demonstrations of “statistical learning” can scale up to allow learning of the lexicons of natural languages, which are orders of magnitude larger. Artificial language experiments with adults can be used to test whether the mechanisms of statistical learning are in principle scalable to larger lexicons. We report data from a large-scale learning experiment that demonstrates that adults can learn words from unsegmented input in much larger languages than previously documented and that they retain the words they learn for years. These results suggest that statistical word segmentation could be scalable to the challenges of lexical acquisition in natural language learning.     

Commentary Salicylate Trapping of -OH as a Tool for Studying Post-Ischemic Oxidative Injury in the Isolated Rat Heart

The use of salicylate as a chemical trap for -OH represents a simple and convenient alternative to the use of spin trapping techniques to study oxidative injury in isolated perfused organs. In these systems, salicylate is included in the perfusion buffer at concentrations ranging from 0.1 to 2mM depending on the detection apparatus employed. In our studies, we have used a coulometric detector, which has a theoretical efficiency of 100% as compared to 1-5% for the standard glassy carbon electrode. We have been able to generate reproducible results by inclusion of only 100 μM salicylate, a concentration demonstrated not to affect pre- or post-ischemic cardiac function. In initial studies, we observed an increase in perfusate 2,5-dihydroxybenzoic acid consistent with an early post-ischemic burst of -OH, not unlike that reported using spin trapping techniques. Since then we and others have used this technique to examine possible relationships between -OH formation and treatments that alter post-ischemic cardiac functional recovery. For example, preischemic loading of hearts with copper results in increases in postischemic dysfunction and LDH release that were associated with an increase in 2,5-dihydroxybenzoate and by inference, -OH formation. Alternatively, we have reported that the nitroxide spin label, TEMPO, reputed to be a superoxide dismutase mimetic, decreased post-ischemic arrhythmias and 2,5-dihydroxybenzoate formation. Most recently, we have observed that preischemic loading of hearts with zinc-bis-histidinate results in improved post-ischemic cardiac function and decreased LDH release; changes that were associated with decreased 2,5-dihydroxybenzoate formation. These studies indicate that under certain conditions, salicylate is a valuable alternative to spin trapping techniques to probe the role of -OH in cardiac oxidative injury, particularly when applied to the isolated perfused heart preparation.     

The aged hematopoietic system promotes hippocampal‐dependent cognitive decline

The aged systemic milieu promotes cellular and cognitive impairments in the hippocampus. Here, we report that aging of the hematopoietic system directly contributes to the pro‐aging effects of old blood on cognition. Using a heterochronic hematopoietic stem cell (HSC) transplantation model (in which the blood of young mice is reconstituted with old HSCs), we find that exposure to an old hematopoietic system inhibits hippocampal neurogenesis, decreases synaptic marker expression, and impairs cognition. We identify a number of factors elevated in the blood of young mice reconstituted with old HSCs, of which cyclophilin A (CyPA) acts as a pro‐aging factor. Increased systemic levels of CyPA impair cognition in young mice, while inhibition of CyPA in aged mice improves cognition. Together, these data identify age‐related changes in the hematopoietic system as drivers of hippocampal aging.     

Three critical factors affecting automated image species recognition performance for camera traps

Ecological camera traps are increasingly used by wildlife biologists to unobtrusively monitor an ecosystems animal population. However, manual inspection of the images produced is expensive, laborious, and time‐consuming. The success of deep learning systems using camera trap images has been previously explored in preliminary stages. These studies, however, are lacking in their practicality. They are primarily focused on extremely large datasets, often millions of images, and there is little to no focus on performance when tasked with species identification in new locations not seen during training. Our goal was to test the capabilities of deep learning systems trained on camera trap images using modestly sized training data, compare performance when considering unseen background locations, and quantify the gradient of lower bound performance to provide a guideline of data requirements in correspondence to performance expectations. We use a dataset provided by Parks Canada containing 47,279 images collected from 36 unique geographic locations across multiple environments. Images represent 55 animal species and human activity with high‐class imbalance. We trained, tested, and compared the capabilities of six deep learning computer vision networks using transfer learning and image augmentation: DenseNet201, Inception‐ResNet‐V3, InceptionV3, NASNetMobile, MobileNetV2, and Xception. We compare overall performance on “trained” locations where DenseNet201 performed best with 95.6% top‐1 accuracy showing promise for deep learning methods for smaller scale research efforts. Using trained locations, classifications with <500 images had low and highly variable recall of 0.750 ± 0.329, while classifications with over 1,000 images had a high and stable recall of 0.971 ± 0.0137. Models tasked with classifying species from untrained locations were less accurate, with DenseNet201 performing best with 68.7% top‐1 accuracy. Finally, we provide an open repository where ecologists can insert their image data to train and test custom species detection models for their desired ecological domain.     

Grassland area determines beetle assemblage dissimilarity from surrounding floodplain forest

Patch size is known to affect biodiversity in fragmented landscapes, but is usually examined in systems where the surrounding matrix habitat is unfavourable. We examined beetle diversity in a floodplain ecosystem that is characterised by naturally occurring grassland patches within a dominant matrix of contrasting yet habitable forest. We asked whether differences in the beetle assemblage between grassland and forest vegetation depended on the area of the grassland patch, which is a function of its flooding frequency and duration: smaller grasslands tend to be higher on the floodplain and are flooded less often and for shorter periods than larger grasslands. We found a negative relationship between grassland area and beetle abundance and species richness, and a positive relationship between grassland area and compositional dissimilarity from the surrounding forest. As expected, we found an overall difference in composition between forest and grassland assemblages, with five beetle species more common in the grasslands. Our study indicates that floodplain grasslands not only support beetle assemblages that are distinct from the surrounding forest, but that assemblages from the larger grasslands are compositionally more distinct than those from smaller grasslands. A likely cause of this pattern is the reduced edge effects and greater environmental contrast between forest and large grasslands that may be exposed to greater variation in local climate. Ongoing changes to flood regimes and potential encroachment of forest plants may decrease grassland area in the future, which may reduce spatial heterogeneity in the insect community in this unique floodplain ecosystem.     

The spatial scaling of beta diversity

Beta diversity is an important concept used to describe turnover in species composition across a wide range of spatial and temporal scales, and it underpins much of conservation theory and practice. Although substantial progress has been made in the mathematical and terminological treatment of different measures of beta diversity, there has been little conceptual synthesis of potential scale dependence of beta diversity with increasing spatial grain and geographic extent of sampling. Here, we evaluate different conceptual approaches to the spatial scaling of beta diversity, interpreted from ‘fixed’ and ‘varying’ perspectives of spatial grain and extent. We argue that a ‘sliding window’ perspective, in which spatial grain and extent covary, is an informative way to conceptualize community differentiation across scales. This concept more realistically reflects the varying empirical approaches that researchers adopt in field sampling and the varying scales of landscape perception by different organisms. Scale dependence in beta diversity has broad implications for emerging fields in ecology and biogeography, such as the integration of fine‐resolution ecogenomic data with large‐scale macroecological studies, as well as for guiding appropriate management responses to threats to biodiversity operating at different spatial scales.     

CD109 Plays a Role in Osteoclastogenesis

Osteoclasts are large multinucleated cells that arise from the fusion of cells from the monocyte/macrophage lineage. Osteoclastogenesis is mediated by macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor-kB ligand (RANKL) and involves a complex multistep process that requires numerous other elements, many of which remain undefined. The primary aim of this project was to identify novel factors which regulate osteoclastogenesis. To carry out this investigation, microarray analysis was performed comparing two pre-osteoclast cell lines generated from RAW264.7 macrophages: one that has the capacity to fuse forming large multinucleated cells and one that does not fuse. It was found that CD109 was up-regulated by>17-fold in the osteoclast forming cell line when compared to the cell line that does not fuse, at day 2 of the differentiation process. Results obtained with microarray were confirmed by RT-qPCR and Western blot analyses in the two cell lines, in the parental RAW264.7 cell line, as well as primary murine monocytes from bone marrow. A significant increase of CD109 mRNA and protein expression during osteoclastogenesis occurred in all tested cell types. In order to characterize the role of CD109 in osteoclastogenesis, CD109 stable knockdown cell lines were established and fusion of osteoclast precursors into osteoclasts was assessed. It was found that CD109 knockdown cell lines were less capable of forming large multinucleated osteoclasts. It has been shown here that CD109 is expressed in monocytes undergoing RANKL-induced osteoclastogenesis. Moreover, when CD109 expression is suppressed in vitro, osteoclast formation decreases. This suggests that CD109 might be an important regulator of osteoclastogenesis. Further research is needed in order to characterize the role played by CD109 in regulation of osteoclast differentiation.