First-principle data-driven models for assessment of motor disorders in Parkinson’s disease

Parkinson’s disease (PD) is an increasing neurological disorder in an aging society. The motor and non-motor symptoms of PD advance with the disease progression and occur in varying frequency and duration. In order to affirm the full extent of a patient’s condition, repeated assessments are necessary to adjust medical prescription. In clinical studies, symptoms are assessed using the unified Parkinson’s disease rating scale (UPDRS). On one hand, the subjective rating using UPDRS relies on clinical expertise. On the other hand, it requires the physical presence of patients in clinics which implies high logistical costs. Another limitation of clinical assessment is that the observation in hospital may not accurately represent a patient’s situation at home. For such reasons, the practical frequency of tracking PD symptoms may under-represent the true time scale of PD fluctuations and may result in an overall inaccurate assessment. Current technologies for at-home PD treatment are based on data-driven approaches for which the interpretation and reproduction of results are problematic.  The overall objective of this thesis is to develop and evaluate unobtrusive computer methods for enabling remote monitoring of patients with PD. It investigates first-principle data-driven model based novel signal and image processing techniques for extraction of clinically useful information from audio recordings of speech (in texts read aloud) and video recordings of gait and finger-tapping motor examinations. The aim is to map between PD symptoms severities estimated using novel computer methods and the clinical ratings based on UPDRS part-III (motor examination). A web-based test battery system consisting of self-assessment of symptoms and motor function tests was previously constructed for a touch screen mobile device. A comprehensive speech framework has been developed for this device to analyze text-dependent running speech by: (1) extracting novel signal features that are able to represent PD deficits in each individual component of the speech system, (2) mapping between clinical ratings and feature estimates of speech symptom severity, and (3) classifying between UPDRS part-III severity levels using speech features and statistical machine learning tools. A novel speech processing method called cepstral separation difference showed stronger ability to classify between speech symptom severities as compared to existing features of PD speech. In the case of finger tapping, the recorded videos of rapid finger tapping examination were processed using a novel computer-vision (CV) algorithm that extracts symptom information from video-based tapping signals using motion analysis of the index-finger which incorporates a face detection module for signal calibration. This algorithm was able to discriminate between UPDRS part III severity levels of finger tapping with high classification rates. Further analysis was performed on novel CV based gait features constructed using a standard human model to discriminate between a healthy gait and a Parkinsonian gait. The findings of this study suggest that the symptom severity levels in PD can be discriminated with high accuracies by involving a combination of first-principle (features) and data-driven (classification) approaches. The processing of audio and video recordings on one hand allows remote monitoring of speech, gait and finger-tapping examinations by the clinical staff. On the other hand, the first-principles approach eases the understanding of symptom estimates for clinicians. We have demonstrated that the selected features of speech, gait and finger tapping were able to discriminate between symptom severity levels, as well as, between healthy controls and PD patients with high classification rates. The findings support suitability of these methods to be used as decision support tools in the context of PD assessment.

Relationship between genotype and enzyme activity of glutathione S-transferases M1 and P1 in Chinese

Glutathione S-transferases (GSTs) are the major detoxifying Phase II enzyme for eliminating electrophilic compounds. Mutations in GSTM1, GSTP1 and GSTT1 in Caucasian and GSTA1 in Chinese have been found to reduce enzyme activity. However, data on the impact of common genetic polymorphisms of GSTM1 and GSTP1 on enzyme activity in Chinese is lacking. This study aimed to investigate the effect of common GSTP1 and GSTM1 polymorphisms on erythrocyte GST activity in healthy Chinese (n = 196). GSTM1 null mutation (GSTM1*0) was analyzed by a PCR-Multiplex procedure, whereas GSTP1 313A → G polymorphism (resulting in Ile105Val at codon 105) was analyzed by PCR-restriction fragment length polymorphism (RFLP) analysis. Erythrocyte GST activity was measured using 1-chloro-2,4-dinitro-bezene (CDNB) as the model substrate. The frequency of GSTM1 null genotype was 54.3% and the frequency of GSTP1-Ile/Ile, -Ile/Val, and -Val/Val genotype was 60.7%, 35.2% and 4.1%, respectively, with a frequency of 21.7% for the 105 valine allele. Age, gender and smoking did not significantly affect the erythrocyte GST activities. The mean erythrocyte GST enzyme activity for GSTP1*-Ile/Val genotype group (3.53 ± 0.63 U/g Hb) was significantly lower than that for subjects with GSTP1-Ile/Ile genotype (4.25 ± 1.07 U/g Hb, P = 0.004), while subjects with the GSTP1-Val/Val genotype had the lowest enzyme activity (2.44 ± 0.67 U/g Hb). In addition, the GST activity in carriers of GSTM1*0/GSTP1-Ile/Ile was significantly higher than that of subjects inherited GSTM1*0/GSTP1-Ile/Val or GSTM1*0/GSTP1-Val/Val. However, there is no association between GSTM1 null mutation and reduced enzyme activity. GSTP1 codon 105 mutation led to reduced erythrocyte GST activity in Chinese. A combined GSTP1 and GSTM1 null mutations also resulted in significantly reduced GST activity. Further studies are needed to explore the clinical implications of GSTM1 and GSTP1 polymorphisms.

Frequency of migraine and the risk of cardiovascular disease for women : limited evidence of an association

Evaluation of: Kurth T, Schurks M, Logroscino G et al. Migraine frequency and risk of cardiovascular disease in women. Neurology 73, 581–588 (2009). There is substantial evidence that migraine with aura (MA) is associated with ischemic stroke and myocardial infarction in women. The mechanisms of this association are poorly understood. Analysis of data from the Women’s Health Study, from 27,798 women over 45 years of age who were initially free of cardiovascular disease, found that women with baseline MA at a frequency of less than monthly had increased risk of major cardiovascular disease (HR: 2.28; 95% CI: 1.70–3.07) relative to women without migraine, and those who reported MA with a frequency of more than weekly had more than four-times the risk of ischemic stroke (HR: 4.25; 95% CI: 1.36–13.29) compared with those without migraine. Low numbers of outcome events in each of the frequency categories and lack of information on migraine frequency during follow-up limit the interpretation of these findings, but they suggest that frequency of migraine may be a moderating factor in the link between MA and cardiovascular disease.

Learning from large data : bias, variance, sampling, and learning curves

One of the fundamental machine learning tasks is that of predictive classification. Given that organisations collect an ever increasing amount of data, predictive classification methods must be able to effectively and efficiently handle large amounts of data. However, it is understood that present requirements push existing algorithms to, and sometimes beyond, their limits since many classification prediction algorithms were designed when currently common data set sizes were beyond imagination. This has led to a significant amount of research into ways of making classification learning algorithms more effective and efficient. Although substantial progress has been made, a number of key questions have not been answered. This dissertation investigates two of these key questions. The first is whether different types of algorithms to those currently employed are required when using large data sets. This is answered by analysis of the way in which the bias plus variance decomposition of predictive classification error changes as training set size is increased. Experiments find that larger training sets require different types of algorithms to those currently used. Some insight into the characteristics of suitable algorithms is provided, and this may provide some direction for the development of future classification prediction algorithms which are specifically designed for use with large data sets. The second question investigated is that of the role of sampling in machine learning with large data sets. Sampling has long been used as a means of avoiding the need to scale up algorithms to suit the size of the data set by scaling down the size of the data sets to suit the algorithm. However, the costs of performing sampling have not been widely explored. Two popular sampling methods are compared with learning from all available data in terms of predictive accuracy, model complexity, and execution time. The comparison shows that sub-sampling generally products models with accuracy close to, and sometimes greater than, that obtainable from learning with all available data. This result suggests that it may be possible to develop algorithms that take advantage of the sub-sampling methodology to reduce the time required to infer a model while sacrificing little if any accuracy. Methods of improving effective and efficient learning via sampling are also investigated, and now sampling methodologies proposed. These methodologies include using a varying-proportion of instances to determine the next inference step and using a statistical calculation at each inference step to determine sufficient sample size. Experiments show that using a statistical calculation of sample size can not only substantially reduce execution time but can do so with only a small loss, and occasional gain, in accuracy. One of the common uses of sampling is in the construction of learning curves. Learning curves are often used to attempt to determine the optimal training size which will maximally reduce execution time while nut being detrimental to accuracy. An analysis of the performance of methods for detection of convergence of learning curves is performed, with the focus of the analysis on methods that calculate the gradient, of the tangent to the curve. Given that such methods can be susceptible to local accuracy plateaus, an investigation into the frequency of local plateaus is also performed. It is shown that local accuracy plateaus are a common occurrence, and that ensuring a small loss of accuracy often results in greater computational cost than learning from all available data. These results cast doubt over the applicability of gradient of tangent methods for detecting convergence, and of the viability of learning curves for reducing execution time in general.

An approach to filtering duplicate RFID data streams

In a system where distributed network of Radio Frequency Identification (RFID) readers are used to collaboratively collect data from tagged objects, a scheme that detects and eliminates redundant data streams is required. To address this problem, we propose an approach that is based on Bloom filter to detect duplicate readings and filter redundant RFID data streams. We have evaluated the performance of the proposed approach and compared it with existing approaches. The experimental results demonstrate that the proposed approach provides superior performance as compared to the baseline approaches.

Modeling gene-environment interaction in longitudinal data : risk for neuroticism due to interaction between maternal care and the Dopamine 4 Receptor gene (DRD4)

The purpose of this study was to investigate risk for neuroticism due to the joint action of low maternal care and compromised mesocorticolimbic ‘reward’ system function linked to a variable number tandem repeat (VNTR) in the dopamine 4 receptor gene (DRD4). Data were drawn from the Victorian Adolescent Health Cohort Study, a longitudinal study of the health and well-being of 2,000 young Australians followed from adolescence to young adulthood across 8 waves from 14- to 28-years. Genetic risk was defined by carriage of at least one copy of the 7-repeat allele or derivative alleles 5, 6, and 8 (labeled 7R+). Neuroticism was assessed in adolescence and young adulthood. We observed an approximately fourfold increase in the odds of reporting neurotic symptoms in carriers of the 7R+ disposition who reported low maternal care compared with non-carriers who reported high maternal care. The percentage of risk attributable to mechanisms in which both factors played a role was 35%. Findings are discussed in terms of implications for prevention.

Frequency of Anti-HBc & HBV DNA detection in blood donors of Kerman province, Iran

Hepatitis B is a serious global infection disease and a major cause of mortality and morbidity worldwide. However, data on Occult Hepatitis B in Iran are scare. The current study assessed the frequency of Anti-HBc and HBV DNA in serum sample of healthy blood donors negative for HBsAg stratified by sex and age; and also investigated the relationship between detection of HBV-DNA and anti-HBc positivity. Since anti-HBc screening is not performed in Iranian Blood Bank, we assessed whether anti-HBc could be adopted as a screening assay for the donated blood. The study included a total of 1525 blood samples of blood donors negative for hepatitis B virus surface antigen ( 87% male with a mean age ± SD: of 31±8yr; and 13% female with a mean age ± SD of 30±6yr). Eight percent (121 out of 1525) of the blood samples with negative HBs-Ag were positive for Anti-HBc and were all from males. HBV-DNA was detected in 36 out of 121 anti-HBc+ specimens (29.7%). The study found a positive relation between anti-HBc positivity and detection of HBV-DNA in serum samples of HBs-Ag negative blood donors. Findings from this study suggest that, introducing anti HBc screening in Iran maybe very practical in order to limit the transmission risk of Occult Hepatitis B virus through blood transfusion.

An approach for removing redundant data from RFID data streams

Radio frequency identification (RFID) systems are emerging as the primary object identification mechanism, especially in supply chain management. However, RFID naturally generates a large amount of duplicate readings. Removing these duplicates from the RFID data stream is paramount as it does not contribute new information to the system and wastes system resources. Existing approaches to deal with this problem cannot fulfill the real time demands to process the massive RFID data stream. We propose a data filtering approach that efficiently detects and removes duplicate readings from RFID data streams. Experimental results show that the proposed approach offers a significant improvement as compared to the existing approaches.

An efficient one-pass method for discovering bases of recently frequent episodes over online data streams

The knowledge embedded in an online data stream is likely to change over time due to the dynamic evolution of the stream. Consequently, infrequent episode mining over an online stream, frequent episodes should be adaptively extracted from recently generated stream segments instead of the whole stream. However, almost all existing frequent episode mining approaches find episodes frequently occurring over the whole sequence. This paper proposes and investigates a new problem: online mining of recently frequent episodes over data streams. In order to meet strict requirements of stream mining such as one-scan, adaptive result update and instant result return, we choose a novel frequency metric and define a highly condensed set called the base of recently frequent episodes. We then introduce a one-pass method for mining bases of recently frequent episodes. Experimental results show that the proposed method is capable of finding bases of recently frequent episodes quickly and adaptively. The proposed method outperforms the previous approaches with the advantages of one-pass, instant result update and return, more condensed resulting sets and less space usage.

Children's consumption of energy dense nutrient poor foods, fruit and vegetables; are they related? An analysis of data from a cross sectional survey

Issue addressed: Many children consume excessive amounts of energy-dense, nutrient-poor (EDNP) or 'extra' foods and low intakes of fruit and vegetables. The aim of this study was to examine the associations between EDNP foods and ascertain whether certain EDNP foods and beverages are more likely to be eaten in association with other EDNP foods.

Methods: A cross-sectional representative population survey of children in preschool (n=764), and of school students in Years K, 2 and 4 (n=1,560) and in Years 6, 8 and 10 (n=1,685) residing in the Hunter New England region of New South Wales, Australia. Dietary data were collected using a short food frequency questionnaire. Multivariate logistic regression models examined the association between EDNP foods and fruit and vegetable intake. Data were stratified by sex and age cohort.

Results: More frequent consumption of some EDNP food types was significantly associated with more frequent consumption of other EDNP foods. Fast food and soft drinks consumption were associated with each other as well as with fried potato and salty snacks; and with lower intakes of fruit and vegetables in some but not all age groups.

Conclusion: The positive associations found between EDNP foods point towards the existence of a high-risk group of children who frequently consume a variety of EDNP foods and drinks.

Glutamine repeat variants in human RUNX2 associated with decreased femoral neck BMD, broadband ultrasound attenuation and target gene transactivation

RUNX2 is an essential transcription factor required for skeletal development and cartilage formation. Haploinsufficiency of RUNX2 leads to cleidocranial displaysia (CCD) a skeletal disorder characterised by gross dysgenesis of bones particularly those derived from intramembranous bone formation. A notable feature of the RUNX2 protein is the polyglutamine and polyalanine (23Q/17A) domain coded by a repeat sequence. Since none of the known mutations causing CCD characterised to date map in the glutamine repeat region, we hypothesised that Q-repeat mutations may be related to a more subtle bone phenotype. We screened subjects derived from four normal populations for Q-repeat variants. A total of 22 subjects were identified who were heterozygous for a wild type allele and a Q-repeat variant allele: (15Q, 16Q, 18Q and 30Q). Although not every subject had data for all measures, Q-repeat variants had a significant deficit in BMD with an average decrease of 0.7SD measured over 12 BMD-related parameters (p = 0.005). Femoral neck BMD was measured in all subjects (−0.6SD, p = 0.0007). The transactivation function of RUNX2 was determined for 16Q and 30Q alleles using a reporter gene assay. 16Q and 30Q alleles displayed significantly lower transactivation function compared to wild type (23Q). Our analysis has identified novel Q-repeat mutations that occur at a collective frequency of about 0.4%. These mutations significantly alter BMD and display impaired transactivation function, introducing a new class of functionally relevant RUNX2 mutants.

Multi-target tracking via space transformations using a single frequency continuous wave radar

This paper investigates the location and velocity estimation problem involving multiple targets using the phase difference and frequency shift of the returned Doppler modulated signal. The minimal receiver configuration that addresses the data association and missing information problem is presented for the case of linear arrays. Non-linearly modeled Doppler radar measurements are used to obtain an accurate estimate of the target dynamics progressively in a linear framework utilizing a recently developed robust state estimation approach.

StenoCipher to provide data confidentiality and tampered data recovery for RFID tag

Radio Frequency Identification (RFID) is an emerging wireless object identification technology with many potential applications such as supply chain management, personnel tracking and healthcare. However, security vulnerabilities of the RFID system have been a serious concern for its wide adoption in many applications. Although much work has been done to provide privacy and anonymity, little focus has been given to ensure RFID data confidentiality, integrity and to address the tampered data recovery problem. To this end, we propose a lightweight stenographic-based approach to ensure RFID data confidentiality and integrity as well as the recovery of tampered RFID data.

Enhancing RFID data quality and reliability

This thesis addressed the problem of data quality, reliability and energy consumption of networked Radio Frequency Identification systems for business intelligence applications decision making processes. The outcome of the research substantially improved the accuracy and reliability of RFID generated data as well as energy depletion thus prolonging RFID system lifetime.

DSM-5 mixed specifier for manic episodes: evaluating the effect of depressive features on severity and treatment outcome using asenapine clinical trial data

Background:
To describe the frequency of mixed specifier as proposed in DSM-5 in bipolar I patients with manic episodes, and to evaluate the effect of mixed specifier on symptom severity and treatment outcome.

Methods:
This post-hoc analysis used proxies for DSM-5 mixed features specifier by using MADRS or PANSS items.

Results:
Of the 960 patients analysed, 34%, 18% and 4.3% of patients, respectively, had ≥3 depressive features with mild (score ≥1 for MADRS items and ≥2 for PANSS item), moderate (score ≥2 MADRS, ≥3 PANSS) and severe (score ≥3 MADRS, ≥4 PANSS) symptoms. In patients with ≥3 depressive features and independent of treatment: MADRS remission (score ≤12) rate decreased with increasing severity (61–43%) and YMRS remission (score ≤12) was similar for mild and moderate patients (36–37%), but higher for severe (54%). In asenapine-treated patients, the MADRS remission rate was stable regardless of baseline depressive symptom severity (range 64–67%), whereas remission decreased with increasing severity with olanzapine (63–38%) and placebo (49–25%). Reduction in YMRS was significantly greater for asenapine compared with placebo at day 2 across the 3 severity cut-offs and continued to decrease throughout the treatment period. The difference between olanzapine and placebo was statistically significant in mild and moderate patients.

Limitations:
Results are from post-hoc analyses.

Conclusions:
These analyses support the validity of proposed DSM-5 criteria. They confirm that depressive features are frequent in bipolar patients with manic episodes. With increasing baseline severity of depressive features, treatment outcome was poorer with olanzapine and placebo, but remained stable with asenapine.