Carpal tunnel syndrome: diagnostic usefulness of sonography

PURPOSE: To prospectively evaluate accuracy of sonography for diagnosis of carpal tunnel syndrome (CTS) in patients clinically suspected of having the disease in one or both hands.
MATERIALS AND METHODS: A prospective cohort of 133 patients suspected of having CTS were referred to a teaching hospital between October 2001 and June 2002 for electrodiagnostic study. One hundred twenty patients (98 women, 22 men; mean age, 49 years; range, 19–83 years) underwent sonography within 1 week after electrodiagnostic study. Radiologist was blinded to electrodiagnostic study results. Seventy-five patients had bilateral symptoms; 23 patients, right-hand symptoms; and 22 patients, left-hand symptoms (total, 195 symptomatic hands). Cross-sectional area of median nerve was measured at three levels: immediately proximal to carpal tunnel inlet, at carpal tunnel inlet, and at carpal tunnel outlet. Flexor retinaculum was used as a landmark to margins of carpal tunnel. Optimal threshold levels (determined with classification and regression tree analysis) for areas proximal to and at tunnel inlet and at tunnel outlet were used to discriminate between patients with and patients without disease. Sensitivity, specificity, and false-positive and false-negative rates were derived on the basis of final diagnosis, which was determined with clinical history and electrodiagnostic study results as reference standard.
RESULTS: For right hands, sonography had sensitivity of 94% (66 of 70); specificity, 65% (17 of 26); false-positive rate, 12% (nine of 75); and false-negative rate, 19% (four of 21) (cutoff, 0.09 cm2 proximal to tunnel inlet and 0.12 cm2 at tunnel outlet). For left hands, sensitivity was 83% (53 of 64); specificity, 73% (24 of 33); false-positive rate, 15% (nine of 62); and false-negative rate, 31% (11 of 35) (cutoff, 0.10 cm2 proximal to tunnel inlet).
CONCLUSION: Sonography is comparable to electrodiagnostic study in diagnosis of CTS and should be considered as initial test of choice for patients suspected of having CTS.

An abnormal-based approach to effectively detect DDOS attacks

Distributed Denial-of-Service (DDoS) attacks are a serious threat to the safety and security of cyberspace. In this paper we propose a novel metric to detect DDoS attacks in the Internet. More precisely, we use the function of order α of the generalized (Rényi) entropy to distinguish DDoS attacks traffic from legitimate network traffic effectively. In information theory, entropies make up the basis for distance and divergence measures among various probability densities. We design our abnormal-based detection metric using the generalized entropy. The experimental results show that our proposed approach can not only detect DDoS attacks early (it can detect attacks one hop earlier than using the Shannon metric while order  α =2, and two hops earlier than the Shannon metric while order α =10.) but can also reduce both the false positive rate and the false negative rate, compared with the traditional Shannon entropy metric approach.

Effective DDoS attacks detection using generalized entropy metric

In information theory, entropies make up of the basis for distance and divergence measures among various probability densities. In this paper we propose a novel metric to detect DDoS attacks in networks by using the function of order α of the generalized (Rényi) entropy to distinguish DDoS attacks traffic from legitimate network traffic effectively. Our proposed approach can not only detect DDoS attacks early (it can detect attacks one hop earlier than using the Shannon metric while order α=2, and two hops earlier to detect attacks while order α=10.) but also reduce both the false positive rate and the false negative rate clearly compared with the traditional Shannon entropy metric approach.

Browsing behavior mimicking attacks on popular web sites for large botnets

With the significant growth of botnets, application layer DDoS attacks are much easier to launch using large botnet, and false negative is always a problem for intrusion detection systems in real practice. In this paper, we propose a novel application layer DDoS attack tool, which mimics human browsing behavior following three statistical distributions, the Zipf-like distribution for web page popularity, the Pareto distribution for page request time interval for an individual browser, and the inverse Gaussian distribution for length of browsing path. A Markov model is established for individual bot to generate attack request traffic. Our experiments indicated that the attack traffic that generated by the proposed tool is pretty similar to the real traffic. As a result, the current statistics based detection algorithms will result high false negative rate in general. In order to counter this kind of attacks, we discussed a few preliminary solutions at the end of this paper.

Feature reduction to speed up malware classification

In statistical classification work, one method of speeding up the process is to use only a small percentage of the total parameter set available. In this paper, we apply this technique both to the classification of malware and the identification of malware from a set combined with cleanware. In order to demonstrate the usefulness of our method, we use the same sets of malware and cleanware as in an earlier paper. Using the statistical technique Information Gain (IG), we reduce the set of features used in the experiment from 7,605 to just over 1,000. The best accuracy obtained in the former paper using 7,605 features is 97.3% for malware versus cleanware detection and 97.4% for malware family classification; on the reduced feature set, we obtain a (best) accuracy of 94.6% on the malware versus cleanware test and 94.5% on the malware classification test. An interesting feature of the new tests presented here is the reduction in false negative rates by a factor of about 1/3 when compared with the results of the earlier paper. In addition, the speed with which our tests run is reduced by a factor of approximately 3/5 from the times posted for the original paper. The small loss in accuracy and improved false negative rate along with significant improvement in speed indicate that feature reduction should be further pursued as a tool to prevent algorithms from becoming intractable due to too much data.

Rapid detection of SMARCB1 sequence variation using high resolution melting

Background : Rhabdoid tumors are rare cancers of early childhood arising in the kidney, central nervous system and other organs. The majority are caused by somatic inactivating mutations or deletions affecting the tumor suppressor locus SMARCB1 [OMIM 601607]. Germ-line SMARCB1 inactivation has been reported in association with rhabdoid tumor, epitheloid sarcoma and familial schwannomatosis, underscoring the importance of accurate mutation screening to ascertain recurrence and transmission risks. We describe a rapid and sensitive diagnostic screening method, using high resolution melting (HRM), for detecting sequence variations in SMARCB1. Methods : Amplicons, encompassing the nine coding exons of SMARCB1, flanking splice site sequences and the 5' and 3' UTR, were screened by both HRM and direct DNA sequencing to establish the reliability of HRM as a primary mutation screening tool. Reaction conditions were optimized with commercially available HRM mixes. Results : The false negative rate for detecting sequence variants by HRM in our sample series was zero. Nine amplicons out of a total of 140 (6.4%) showed variant melt profiles that were subsequently shown to be false positive. Overall nine distinct pathogenic SMARCB1 mutations were identified in a total of 19 possible rhabdoid tumors. Two tumors had two distinct mutations and two harbored SMARCB1 deletion. Other mutations were nonsense or frame-shifts. The detection sensitivity of the HRM screening method was influenced by both sequence context and specific nucleotide change and varied from 1: 4 to 1:1000 (variant to wild-type DNA). A novel method involving digital HRM, followed by re-sequencing, was used to confirm mutations in tumor specimens containing associated normal tissue. Conclusions : This is the first report describing SMARCB1 mutation screening using HRM. HRM is a rapid, sensitive and inexpensive screening technology that is likely to be widely adopted in diagnostic laboratories to facilitate whole gene mutation screening.

Detecting contaminated birthdates using generalized additive models.

Erroneous patient birthdates are common in health databases. Detection of these errors usually involves manual verification, which can be resource intensive and impractical. By identifying a frequent manifestation of birthdate errors, this paper presents a principled and statistically driven procedure to identify erroneous patient birthdates.

Malicious node detection in OppNets using hash chain technique

Opportunistic Networks aim to set a reliable networks where the nodes has no end-To-end connection and the communication links often suffer from frequent disruption and long delays. The design of the OppNets routing protocols is facing a serious challenges such as the protection of the data confidentiality and integrity. OppNets exploit the characteristics of the human social, such as similarities, daily routines, mobility patterns and interests to perform the message routing and data sharing. Packet dropping attack is one of the hardest attacks in Opportunistic Networks as both the source nodes and the destination nodes have no knowledge of where or when the packet will be dropped. In this paper, we present a new malicious nodes detection technique against packet faking attack where the malicious node drops one or more packets and instead of them injects new fake packets. We have called this novel attack in our previous works a packet faking attack. Each node in Opportunistic Networks can detect and then traceback the malicious nodes based on a solid and powerful idea that is, hash chain techniques. In our hash chain based defense techniques we have two phases. The first phases is to detect the attack, and the second phases is to find the malicious nodes. We have compared our approach with the acknowledgement based mechanisms and the networks coding based mechanism which are well known approaches in the literature. In our simulation, we have achieved a very high node detection accuracy and low false negative rate.

Applying both positive and negative selection to supervised learning for anomaly detection

This paper presents a novel approach of applying both positive selection and negative selection to supervised learning for anomaly detection. It first learns the patterns of the normal class via co-evolutionary genetic algorithm, which is inspired from the positive selection, and then generates synthetic samples of the anomaly class, which is based on the negative selection in the immune system. Two algorithms about synthetic generation of the anomaly class are proposed. One deals with data sets containing a few anomalous samples; while the other deals with data sets containing no anomalous samples at all. The experimental results on some benchmark data sets from UCI data set repertory show that the detection rate is improved evidently, accompanied by a slight increase in false alarm rate via introducing novel synthetic samples of the anomaly class. The advantages of our method are the increased ability of classifiers in identifying both previously known and innovative anomalies, and the maximal degradation of overfitting phenomenon.

Automated attention flags in chronic disease care planning

Objectives: To assess the value of computerised decision support in the management of chronic respiratory disease by comparing agreement between three respiratory specialists, general practitioners (care coordinators), and decision support software.
Methods: Care guidelines for two chronic obstructive pulmonary disease projects of the SA HealthPlus Coordinated Care Trial were formulated. Decision support software, Care Plan On-Line (CPOL), was created to represent the intent of these guidelines via automated attention flags to appear in patients' electronic medical records. For a random sample of 20 patients with care plans, decisions about the use of nine additional services (eg,.smoking cessation, pneumococcal vaccination) were compared between the respiratory specialists, the patients' GPs and the CPOL attention flags.
Results: Agreement among the specialists was at the lower end of moderate (intraclass correlation coefficient [ICC], 0.48; 95% CI, 0.39-0.56), with a 20% rate of contradictory decisions. Agreement with recommendations of specialists was moderate to poor for GPs (le, 0.49; 95% CI, 0.33-0.66) and moderate to good for CPOL (K, 0.72; 95% CI, 0.55-0.90). CPOL agreement with GPs was moderate to poor (le, 0.41; 95% CI, 0.24-0.58). GPs were less likely than specialists or CPOL to decide in favour of an additional service (P< 0.001). CPOL was 87% accurate as an indicator of specialist decisions. It gave a 16% false-positive rate according to specialist decisions, and flagged 61% of decisions where GPs said No and specialists said Yes.
Conclusions: Automated decision support may provide GPs with improved access to the intent of guidelines; however, further investigation is required.

Automated identification of lung nodules

A system that can automatically detect nodules within lung images may assist expert radiologists in interpreting the abnormal patterns as nodules in 2D CT lung images. A system is presented that can automatically identify nodules of various sizes within lung images. The pattern classification method is employed to develop the proposed system. A random forest ensemble classifier is formed consisting of many weak learners that can grow decision trees. The forest selects the decision that has the most votes. The developed system consists of two random forest classifiers connected in a series fashion. A subset of CT lung images from the LIDC database is employed. It consists of 5721 images to train and test the system. There are 411 images that contained expert- radiologists identified nodules. Training sets consisting of nodule, non-nodule, and false-detection patterns are constructed. A collection of test images are also built. The first classifier is developed to detect all nodules. The second classifier is developed to eliminate the false detections produced by the first classifier. According to the experimental results, a true positive rate of 100%, and false positive rate of 1.4 per lung image are achieved.

Spam filtering for network traffic security on a multi-core environment

This paper presents an innovative fusion-based multi-classifier e-mail classification on a ubiquitous multicore architecture. Many previous approaches used text-based single classifiers to identify spam messages from a large e-mail corpus with some amount of false positive tradeoffs. Researchers are trying to prevent false positive in their filtering methods, but so far none of the current research has claimed zero false positive results. In e-mail classification false positive can potentially cause serious problems for the user. In this paper, we use fusion-based multi-classifier classification technique in a multi-core framework. By running each classifier process in parallel within their dedicated core, we greatly improve the performance of our multi-classifier-based filtering system in terms of running time, false positive rate, and filtering accuracy. Our proposed architecture also provides a safeguard of user mailbox from different malicious attacks. Our experimental results show that we achieved an average of 30% speedup at an average cost of 1.4 ms. We also reduced the instances of false positives, which are one of the key challenges in a spam filtering system, and increases e-mail classification accuracy substantially compared with single classification techniques.

Distinguishing DDoS attacks from flash crowds using probability metrics

Both Flash crowds and DDoS (Distributed Denial-of-Service) attacks have very similar properties in terms of internet traffic, however Flash crowds are legitimate flows and DDoS attacks are illegitimate flows, and DDoS attacks have been a serious threat to internet security and stability. In this paper we propose a set of novel methods using probability metrics to distinguish DDoS attacks from Flash crowds effectively, and our simulations show that the proposed methods work well. In particular, these mathods can not only distinguish DDoS attacks from Flash crowds clearly, but also can distinguish the anomaly flow being DDoS attacks flow or being Flash crowd flow from Normal network flow effectively. Furthermore, we show our proposed hybrid probability metrics can greatly reduce both false positive and false negative rates in detection.

Identifying co-regulating microrna groups

Background: Current miRNA target prediction tools have the common problem that their false positive rate is high. This renders identification of co-regulating groups of miRNAs and target genes unreliable. In this study, we describe a procedure to identify highly probable co-regulating miRNAs and the corresponding co-regulated gene groups. Our procedure involves a sequence of statistical tests: (1) identify genes that are highly probable miRNA targets; (2) determine for each such gene, the minimum number of miRNAs that co-regulate it with high probability; (3) find, for each such gene, the combination of the determined minimum size of miRNAs that co-regulate it with the lowest p-value; and (4) discover for each such combination of miRNAs, the group of genes that are co-regulated by these miRNAs with the lowest p-value computed based on GO term annotations of the genes.
Results: Our method identifies 4, 3 and 2-term miRNA groups that co-regulate gene groups of size at least 3 in human. Our result suggests some interesting hypothesis on the functional role of several miRNAs through a "guilt by association" reasoning. For example, miR-130, miR-19 and miR-101 are known neurodegenerative diseases associated miRNAs. Our 3-term miRNA table shows that miR-130/19/101 form a co-regulating group of rank 22 (p-value =1.16 × 10^-2). Since miR-144 is co-regulating with miR-130, miR-19 and miR-101 of rank 4 (p-value = 1.16 × 10^-2) in our 4-term miRNA table, this suggests hsa-miR-144 may be neurodegenerative diseases related miRNA. Conclusions: This work identifies highly probable co-regulating miRNAs, which are refined from the prediction by computational tools using (1) signal-to-noise ratio to get high accurate regulating miRNAs for every gene, and (2) Gene Ontology to obtain functional related co-regulating miRNA groups. Our result has partly been supported by biological experiments. Based on prediction by TargetScanS, we found highly probable target gene groups in the Supplementary Information. This result might help biologists to find small set of miRNAs for genes of interest rather than huge amount of miRNA set.