Stability criteria for unsupervised temporal association networks

A biologically realizable, unsupervised learning rule is described for the online extraction of object features, suitable for solving a range of object recognition tasks. Alterations to the basic learning rule are proposed which allow the rule to better suit the parameters of a given input space. One negative consequence of such modifications is the potential for learning instability. The criteria for such instability are modeled using digital filtering techniques and predicted regions of stability and instability tested. The result is a family of learning rules which can be tailored to the specific environment, improving both convergence times and accuracy over the standard learning rule, while simultaneously insuring learning stability.

Tennis, incidence of URTI and salivary IgA

Tennis played at an elite level requires intensive training characterized by repeated bouts of brief intermittent high intensity exercise over relatively long periods of time (1 - 3 h or more). Competition can place additional stress on players. The purpose of this study was to investigate the temporal association between specific components of tennis training and competition, the incidence of upper respiratory tract infections (URT1), and salivary IgA, in a cohort of seventeen elite female tennis players. Timed, whole unstimulated saliva samples were collected before and after selected 1-h training sessions at 2 weekly intervals, over 12 weeks. Salivary IgA concentration was measured by ELISA and IgA secretion rate calculated (mug IgA x ml(-1) x ml saliva x min(-1)). Players reported URTI symptoms and recorded training and competition in daily logs. Data analysis showed that higher incidence of URTI was significantly associated with increased training duration and load, and competition level, on a weekly basis. Salivary IgA secretion rate (S-IgA) dropped significantly after 1 hour of tennis play. Over the 12-week period, pre-exercise salivary IgA concentration and secretion rate were directly associated with the amount of training undertaken during the previous day and week (p < 0.05). However, the decline in S-IgA after 1 h of intense tennis play was also positively related to the duration and load of training undertaken during the previous day and week (p < 0.05). Although exercise-induced suppression of salivary IgA may be a risk factor, it could not accurately predict the occurrence of URTI in this cohort of athletes.

Spatial and temporal variation in algal-assemblage structure in isolated dryland river waterholes, Cooper Creek and Warrego River, Australia

The scale at which algal biodiversity is partitioned across the landscape, and the biophysical processes and biotic interactions which shape these communities in dryland river refugia was studied on two occasions from 30 sites in two Australian dryland rivers. Despite the waterholes studied having characteristically high levels of abiogenic turbidity, a total of 186 planktonic microalgae, 253 benthic diatom and 62 macroalgal species were recorded. The phytoplankton communities were dominated by flagellated cryptophytes, euglenophytes and chlorophytes, the diatom communities by cosmopolitan taxa known to tolerate wide environmental conditions, and the macroalgal communities by filamentous cyanobacteria. All algal communities showed significant differences between catchments and sampling times, with a suite of between 5 and 12 taxa responsible for similar to 50% of the observed change. In general, algal assemblage patterns were poorly correlated with the measured environmental variables. Phytoplankton and diatom assemblage patterns were weakly correlated with several waterhole geomorphic measures, whereas macroalgal assemblage patterns showed some association with variability in ionic concentration.