ObjectiveTo describe the patterns and clinical features of toxicity related to recreational use of mephedrone and other cathinones in the UK using data collected by the National Poisons Information Service (NPIS).MethodsThe number of accesses to TOXBASE, the NPIS online poisons information database, details of consecutive cases uploaded onto TOXBASE and the number and details of telephone enquiries made to the NPIS by health professionals in the UK were collected for the period March 2009 to February 2010.ResultsOver the year of study there were 2901 TOXBASE accesses and 188 telephone enquiries relating to cathinones, the majority relating to mephedrone (TOXBASE 1664, telephone 157), with a month-on-month increase in numbers. In 131 telephone enquiries concerning mephedrone, alone or in combination with alcohol, common clinical features reported included agitation or aggression (n=32, 24%, 95% CI 18% to 33%), tachycardia (n=29, 22%, 95% CI 16% to 30%), confusion or psychosis (n=18, 14%, 95% CI 9% to 21%), chest pain (n=17, 13%, 95% CI 8% to 20%), nausea (n=15, 11%, 95% CI 7% to 18%), palpitations (n=14, 11%, 95% CI 6% to 18%), peripheral vasoconstriction (n=10, 8%, 95% CI 4% to 14%) and headache (n=7, 5%, 95% CI 2% to 11%). Convulsions were reported in four cases (3%, 95% CI 1% to 8%). One exposed person died following cardiac arrest (1%, 95% CI 0% to 4%), although subsequent investigation suggested that mephedrone was not responsible.ConclusionsToxicity associated with recreational mephedrone use is increasingly common in the UK. Sympathomimetic adverse effects are common and severe effects are also reported. Structured data collected by the NPIS may be of use in identifying trends in poisoning and in establishing toxidromes for new drugs of abuse.

No abstract

The widespread availability of broadband connections has led to an increase in the use of Internet broadcasting (webcasting). Most webcasts are archived and accessed numerous times retrospectively. In the absence of transcripts of what was said, users have difficulty searching and scanning for specific topics. This research investigates user needs for transcription accuracy in webcast archives, and measures how the quality of transcripts affects user performance in a question-answering task, and how quality affects overall user experience. We tested 48 subjects in a within-subjects design under 4 conditions: perfect transcripts, transcripts with 25% Word Error Rate (WER), transcripts with 45% WER, and no transcript. Our data reveals that speech recognition accuracy linearly influences both user performance and experience, shows that transcripts with 45% WER are unsatisfactory, and suggests that transcripts having a WER of 25% or less would be useful and usable in webcast archives.

We develop a general feature space for automatic classification of verbs into lexical semantic classes. Previous work was limited in scope by the need for manual selection of discriminating features, through a linguistic analysis of the target verb classes (Merlo and Stevenson, 2001). We instead analyze the classification structure at a higher level, using the possible defining characteristics of classes as the basis for our feature space. The general feature space achieves reductions in error rates of 42-69%, on a wider range of classes than investigated previously, with comparable performance to feature sets manually selected for the particular classification tasks. Our results show that the approach is generally applicable, and avoids the need for resource-intensive linguistic analysis for each new task.

Acute and overuse injuries are a common experience for artistic gymnasts; however, this population has unique needs when returning to their sport after an injury due to the technical demands imposed during gymnastics. We reviewed the current literature regarding return to play (RTP) in artistic gymnasts and developed four goals: 1) to define the guiding principles used to determine RTP in sports, 2) to identify factors that affect recovery progression among gymnasts, 3) to determine how different injury types affect RTP protocols, and 4) to create structured RTP protocols specific to gymnasts based on sex and body part injured. By establishing these guidelines, we hope to provide guidance to medical providers through a standardized approach for returning gymnasts to their sport.

This paper describes a model-free method for estimating some yield metrics that are used to track integrated circuit fabrication processes. Our method uses binary probe test data at the wafer level to estimate the size, shape and location of large-area defects or clusters of defective chips. Unlike previous methods in the yield modeling literature, our approach makes extensive use of the location of failing chips to directly identify clusters. An important byproduct of this analysis is a decomposition of wafer yield that attributes defective chips to either large-or small-area defects. Simulation studies show that our procedure is superior to the time-honored windowing technique for achieving a similar breakdown. In addition, by directly estimating defect clusters, we can provide engineers with a greater understanding of the manufacturing process. It has been our experience that routine identification of the spatial signatures of clustered defects and associated root-cause analysis is a costeffective approach to yield and process improvement.

The "topic classification" systems described in the speech literature typically partition a collection of spoken messages into a small number of predefined topics. As such, they are only useful if the set of message topics does not vary over time. However, the techniques of textual information retrieval (IR) have long allowed for retrieval by arbitrary subject from a document collection. This paper describes experiments in unrestricted retrieval from a collection of radio news broadcasts. A hybrid message indexing strategy, with conventional word recognition and a fast lattice-based wordspotter, allows for the retrieval of news reports concerning any subject. The results show that retrieval can be carried out extremely quickly and that high accuracy is possible, even with errorful recognition output. THE MESSAGE RETRIEVAL PROBLEMThere is considerable interest, in the speech research community, in the automatic classification of spoken-word recordings solely by their acoustic content. Topic cZussFers achieve this by predefining a group of allowable topics, v d training weights which relate the recognition of certain words to topic membership [1] [2]. This approach guarantees good performance at the expense of flexibility. It is no trivial matter to add new topics, or split an existing one to create a finer granularity of classification. The utility of such classifiers is therefore limited.Recently, there have been attempts to integrate the wellunderstood methods of textual information retrieval (IR) [3] into an acoustic recognition front-end [4][5]. IR allows for retrieval from a text collection in response to a user's arbitrary information request (e.g. "get m items about the United Nations"). The main problem of applying such an approach to spoken messages is that the vocabulary of interest is not available in advance, and not predictable. Of the systems for which results have been published, the Cambridge VMR system [4] has until now been restricted to indexing messages using a fixed keyword set. Schauble and Glavitsch [5] [6], and Wechsler and Schauble [7], have proposed novel sub-word indexing methods, but published results have until recently been confined to simulation. This paper describes experiments in unrestricted retrieval from a collection of spoken news reports from BBC Radio 4 in the UK. Subsequent sections describe the collection of the acoustic corpora and the sequence of experiments. THE SPOKEN MESSAGE COLLECTIONSeveral factors suggested the use of speech data obtained from BBC Radio 4 news reports [8]. The typical format of these broadcasts is as follows; the newsreader first gives the headlines, and then, for each story, gives a concise summary and introduces a more in-depth, "on the spot" report from a journalist. The regular rotation of newsreaders meant that the same simall group of speakers could be used for both acoustic mcdlel training and message retrieval experiments. Also, the newsreader speech is read as opposed to spontaneous, and clearly spoken, as the newsreaders are all profession...

Networked applications have software components that reside on different computers. Email, for example, has database, processing, and user interface components that can be distributed across a network and shared by users in different locations or work groups. End-to-end performance and reliability metrics describe the software quality experienced by these groups of users, taking into account all the software components in the pipeline. Each user produces only some of the data needed to understand the quality of the application for the group, so group performance metrics are obtained by combining summary statistics that each end computer periodically (and automatically) sends to a central server. The group quality metrics usually focus on medians and tail quantiles rather than on averages. Distributed quantile estimation is challenging, though, especially when passing large amounts of data around the network solely to compute quality metrics is undesirable. This paper describes an Incremental Quantile (IQ) estimation method that is designed for performance monitoring at arbitrary levels of network aggregation and time resolution when only a limited amount of data can be transferred. Applications to both real and simulated data are provided.Comment: This paper commented in: [arXiv:0708.0317], [arXiv:0708.0336], [arXiv:0708.0338]. Rejoinder in [arXiv:0708.0339]. Published at http://dx.doi.org/10.1214/088342306000000583 in the Statistical Science (http://www.imstat.org/sts/) by the Institute of Mathematical Statistics (http://www.imstat.org