Search

Tropical dry forests are among the most threatened ecosystems in the world, and those occurring in the insular Caribbean are particularly vulnerable. Climate change represents a significant threat for the Caribbean region and for small islands like Jamaica. Using the Hellshire Hills protected area in Jamaica, a simple model was developed to project future abundance of arthropods and lizards based on current sensitivities to climate variables derived from rainfall and temperature records. The abundances of 20 modelled taxa were predicted more often by rainfall variables than temperature, but both were found to have strong impacts on arthropod and lizard abundance. Most taxa were projected to decrease in abundance by the end of the century under drier and warmer conditions. Where an increase in abundance was projected under a low emissions scenario, this change was reduced or reversed under a high emissions climate change scenario. The validation process showed that, even for a small population, there was reasonable skill in predicting its annual variability. Results of this study show that this simple model can be used to identify the vulnerability of similar sites to the effects of shifting climate and, by extension, their conservation needs.

An early clinical diagnosis of neurodegenerative diseases is complex and not always possible due to overlapping characteristics between the disorders. Biomarkers are necessary and functional brain imaging may provide a solution. Generally, machine learning can aid the diagnosis, but the decision process of some methods cannot always be understood. Furthermore, machine learning requires data. Therefore, functional brain scans from several neuroimaging centers are combined into a single dataset. We show that this leads to unwanted variation in the data that can inflate the performance of machine learning.Learning Vector Quantization is a type of machine learning that produces a prototypical representation (prototypes) of the classes in the data. Additionally, it weights the input space based on its relevance to the classification task. In one application example, we train a model on steroid measurements from patients with a benign or malignant adrenocortical tumor. In this case, the obtained models were directly interpretable and helped to decide between different measuring technologies.Due to the complex nature of the brain data, the models trained to diagnose neurodegenerative diseases are not directly understandable. Nonetheless, we show that prototypes and relevances can be reconstructed in the imaging space, increasing the interpretability of the models significantly. Additionally, we can produce easy-to-understand representations of the data that can visualize the diagnosis and progression over time of patients, leading to actionable scenarios. Lastly, we present a novel method to deal with center-related, unwanted variance in the data.

BACKGROUND AND OBJECTIVE: Neurodegenerative diseases like Parkinson's disease often take several years before they can be diagnosed reliably based on clinical grounds. Imaging techniques such as MRI are used to detect anatomical (structural) pathological changes. However, these kinds of changes are usually seen only late in the development. The measurement of functional brain activity by means of [18F]fluorodeoxyglucose positron emission tomography (FDG-PET) can provide useful information, but its interpretation is more difficult. The scaled sub-profile model principal component analysis (SSM/PCA) was shown to provide more useful information than other statistical techniques. Our objective is to improve the performance further by combining SSM/PCA and prototype-based generalized matrix learning vector quantization (GMLVQ).METHODS: We apply a combination of SSM/PCA and GMLVQ as a classifier. In order to demonstrate the combination's validity, we analyze FDG-PET data of Parkinson's disease (PD) patients collected at three different neuroimaging centers in Europe. We determine the diagnostic performance by performing a ten times repeated ten fold cross validation. Additionally, discriminant visualizations of the data are included. The prototypes and relevance of GMLVQ are transformed back to the original voxel space by exploiting the linearity of SSM/PCA. The resulting prototypes and relevance profiles have then been assessed by three neurologists.RESULTS: One important finding is that discriminative visualization can help to identify disease-related properties as well as differences which are due to center-specific factors. Secondly, the neurologist assessed the interpretability of the method and confirmed that prototypes are similar to known activity profiles of PD patients.CONCLUSION: We have shown that the presented combination of SSM/PCA and GMLVQ can provide useful means to assess and better understand characteristic differences in FDG-PET data from PD patients and HCs. Based on the assessments by medical experts and the results of our computational analysis we conclude that the first steps towards a diagnostic support system have been taken successfully.

The file attached to this record is the author's final peer reviewed version. The Publisher's final version can be found by following the DOI link Sleep positions have an impact on sleep quality and therefore need to be further analyzed. Current research on position tracking includes only the four basic positions. In the context of wearable devices, energy efficiency is still an open issue. This research presents a way to detect eight positions with higher granularity under energy efficient constraints. Generalized Matrix Learning Vector Quantization is used, as it is a fast and appropriate method for environments with limited computation resources, and has not been seen for this kind of application before. The overall model trained on individuals performs with an averaged accuracy of 99.8%, in contrast to an averaged accuracy of 83.62% for grouped datasets. Real world application gives an accuracy of around 98%. The results show that energy efficiency will be feasible, as performance stays similar for lower sampling rate. This is a step towards a mobile solution which gives more insight in person's sleep behaviour.

This study documents impacts of the small Indian mongoose (Herpestes auropunctatus) on a threatened endemic fauna occurring in a biodiversity hotspot within a hotspot, the Hellshire Hills, Jamaica. We analyzed the stomach contents of 217 mongooses and supplemented this information with behavioural observations. The mongoose’s diet consists primarily of invertebrates and lizards, but bird feathers, mammal hair, and a small number of seeds were also recorded. Invertebrates and lizards accounted for 93% of identified prey items. Of special concern were the remains of threatened species such as the recently re-discovered blue-tailed galliwasp (Celestus duquesneyi), indicating that the mongoose may represent a considerable threat to this poorly known taxon. Dietary analyses did not reveal remains of the Critically Endangered Jamaican iguana (Cyclura collei), yet field observations confirmed that the mongoose is a potent predator of hatchling iguanas. Ignoring issues of sample size, this suggests that the analysis of stomach contents alone may mask important demographic impacts attributable to the mongoose (or other predator species). In other words, rare and endangered species may not be detected in diet samples, but the impact of predation may be of demographic significance for effected prey taxa. This study supports previous arguments concerning the negative impact of the mongoose on endemic insular species, and underscores the utility of employing field observations of mongoose foraging behaviour to provide important insights into the conservation implications of predation by non-native predators.

Fur seals were eliminated by sealers at Heard Island soon after its discovery in the 1850s. The first recorded breeding of Antarctic fur seals (Arctocephalus gazella) since sealing was reported in early 1963 (two pups). The most recent survey of the Heard Island fur-seal population was undertaken between November 2000 and March 2001, when 1,012 Antarctic fur-seal pups were born. This represents a fourfold increase since the last complete census in 1987/1988 (13 years), when 248 births were recorded. Pup estimates and counts available for eight breeding seasons since 1962/1963 suggest the population has been increasing at between 12 and 20% per year. Based on pup production, the breeding population is estimated to number approximately 4,100 seals. The number of fur seals on Heard Island peaked in late February/early March at 29,256 indicating that, in addition to the breeding population, a significant number of seals born elsewhere haul out on the island. Most of these are moulting sub-adult and adult males. As in 1987/1988, only one subantarctic fur-seal pup (A. tropicalis) was observed, suggesting this species is not colonising the island, as has been speculated.