Filter Dissertations and Theses By:
AUDITORY PROCESSING OF COMPLEX TONES IN NEWBORN INFANTS
Year of Dissertation:
Cologne Carnival's "Alternative" Stunksitzung: Carnivalization? Meta-Carnival? Or Bakhtinian Restoration?
Year of Dissertation:
In the 1820s, Carnival in Cologne, Germany, underwent a series of reforms, ostensibly to bring the festival back to the people. Among the traditions that developed was the Sitzung, a theatrical variety-show event, with music, comic speeches and sketches, dance troupes, and various additional Carnival-related entertainments. The shows, and Carnival itself were, and largely have been since that time, mostly overseen by a Festival Committee and the official Carnival Societies it recognizes. In 1984, a group of mostly students decided to create their own version of a Sitzung, an alternative version, the Stunksitzung. From three inaugural performances, it has grown to presenting over forty performances a year to sell-out crowds of one thousand people per night and to being a popular annual television event. This dissertation considers the history of the Stunksitzung within a frame of Mikhail Bakhtin's work on Carnival. I examine over two-dozen performance pieces of the Ensemble, and compare and situate the production and its history within Cologne Carnival, in particular the broader dichotomous status of the official versus the alternative, interrogating how alternative the production is, has been, and continues to be. Ultimately, I frame the Stunksitzung within the larger context of Carnival and the particular status it holds in Cologne.
ANALYSIS AND IMPLEMENTATION OF SIGNAL PROCESSING STRATEGIES FOR A 3-D DOPPLER LIDAR WIND PROFILER
Year of Dissertation:
A heterodyne detection fiber optic based wind lidar system has been developed and tested, which benefits from unique field programmable gate array (FPGA) signal processing techniques and leverages devices from the telecommunication industry to make it particularly cost efficient. A narrow band stabilized fiber laser, polarization maintaining fiber amplifiers, acousto-optic modulators and an optical circulator comprise the transmitter which is coupled to free space using refractive optics. The collinear propagating lidar return signal that scatters off of atmospheric aerosols and in a heterodyne arrangement beats with a local oscillator and is then detected using a shot noise limited polarization maintaining balanced receiver. The system, which operates at a 20 kHz pulse repetition rate and acquires lidar return signals at 400 MSample/second, accumulates signals that are as much as 20 dB lower than the receiver noise power by using embedded programming techniques. For this reason two FPGA embedded programming approaches are considered and compared. In the first approach, the acquired return signal is gated in time and the square modulus of the fast Fourier transform is accumulated for each range gate, producing a series of power spectra as a function of range. Wind speed estimates based on numerical estimators can then be made after transferring the range gated accumulated power spectra to a host computer, enabling line of sight wind speed to be calculated as a function of range gate and stored for additional processing. In the second FPGA approach, a digital IQ demodulator and down sampler reduces the data flow requirements so that an autocorrelation matrix representing a pre-selected number of lags can be accumulated, allowing for the process of range gating to be explored on the host computer. The Fourier transform of the autocorrelation produces the power spectrum and, in the same manner as the first approach, estimates can then be made regarding the line of sight wind speed. The added feature of the second approach is that it allows for an additional capability to adjust the range gate period dynamically as the state of the atmospheric boundary layer (e.g. backscatter coefficient and stability condition) changes. A simple manual beam scanning technique is used to sample three line of sight directions and, by making suitable assumptions regarding the coherence of the averaged wind fields, the three dimensional wind field vector (representing both the horizontal wind speed and direction and the vertical wind speed and direction) is calculated and graphically displayed on time-height cross section plots. Precision in the velocity measurements is estimated to be on the order of 0.08 m/sec and the precision in the measured horizontal wind direction is estimated to be to be about 2 degrees, where both of these estimates are made assuming a relatively short 3-beam cycle time (less than 2 minutes) and a typical backscatter coefficient and atmospheric stability condition. A comparison to other observed wind information is presented which indicates that this lidar will open new doors for the practical characterization of microscale meteorology.
Prosodic Phrasing and Modifier Attachment in Standard Arabic Sentence Processing
Year of Dissertation:
This dissertation investigates the syntax-prosody interface in Standard Arabic, focusing on the ambiguity of a modifier (relative clause or adjective phrase) in relation to the two nouns in a complex noun phrase. Ambiguity resolution tendencies for this construction differ across languages, contrary to otherwise universal parsing tendencies. One explanation proposed is Fodor's (2000) Implicit Prosody Hypothesis: that readers mentally project onto a text a default prosodic phrasing (possibly different between languages), which then influences their syntactic ambiguity resolution. Since implicit (silent) prosody cannot be directly observed, previous research has had to infer it by analogy with overt prosody. But the phonology and orthography of SA permit use of novel methods for tapping into the silent prosody of readers. Liaison phenomena sensitive to prosodic boundaries make phonological phrasing in SA very easy to detect. Also, liaison is indicated by diacritics in the `vowelized' version of SA orthography. Thus, clear data on prosodic phrasing patterns in SA complex nominals can be related to their preferred syntactic/semantic interpretations. Six experiments are reported: three production experiments and three perception experiments. Participants in Experiment 4 silently read sentences in unvowelized orthography, and added diacritics as they thought appropriate. The inserted diacritics gave evidence of their implicit prosodic phrasing of the sentence. Experiments 5 and 6 investigated Arabic speakers' preferred overt prosodic phrasing when the modifier was forced to attach to either the lower or the higher noun, providing standards for comparison with the prosodic phrasing preferences in silent reading in Experiment 4. The orthography was put to a different use in assessing modifier interpretation under varying prosodic conditions. In Experiments 2 and 3, vowelized text was presented, establishing one or other of two relevant prosodic patterns. Participants read aloud, and then indicated their interpretation of the sentence. This provided standards for comparison with modifier interpretation (attachment preferences) in silent reading of unvowelized texts (lacking prosodic disambiguation) in Experiment 1. Results obtained from these experiments provide new information concerning the constraints that apply at the syntax-prosody interface in SA, and also support the hypothesis of an effect of implicit prosody on syntactic interpretation during silent reading.
MIDDLEWARE ROUTING ALGORITHMS COMPONENTS FOR MOBILE AD-HOC WIRELESS NETWORKS
Year of Dissertation:
In this research, we introduce middleware routing algorithms components for Mobile Ad-hoc Networks (MANETs). Unlike the conventional networks, MANET is a decentralized radio wireless network that can be established in situation where no infrastructure exists or where deployment of infrastructure is expensive or inconvenient. This inherent flexibility makes it attractive for applications such as military operations, vehicle to vehicle networks, sensor networks, etc. Hence, MANETs require special type of routing algorithms to operate efficiently in such dynamic environment (i.e., wireless channel, bandwidth constrains, nodes resources, etc ...). In this thesis, we propose an add-on generic solution to on-demand ad-hoc routing protocols to enhance the routing protocols performance with minimum control overhead. Our Solution, namely, Destination Assisted Routing Enhancement (DARE), is based on the new idea of transmitting frequent destination beacon packets. These beacon packets are able to refresh the routing cache tables and announce the destination node existence. This methodology results in dramatically minimizing the initialization (learning/optimization) connection set-time as well as the network overhead. Comparison v between the traditional Dynamic Source Routing protocol (DSR) and the DSR with DARE (DSR-DARE) are presented to show the potential of DARE middleware. Second, we propose middleware protocol components in order to improve the real-time applications at the receivers' side; we propose an algorithm that gives the receiver dynamic ability to move from one multicast session to another based on the receiver capabilities and the path conditions leading to it. Our Multicast Multi-Stream (MMS) solution is added as an extension to the traditional Protocol Independent Multicast (PIM) protocol. Then, we present cooperative video caching technique in MANETs in order to reduce the average access latency as well as enhance the video accessibility. Efficient video caching placement and replacement strategies are developed at some of the distributed intermediate nodes across the network. The simulations results have shown that the system has better video perception (i.e. Quality of Service).
Efferent-Mediated Changes in the Composite Distortion Product Otoacoustic Emissions Signal and Its Components: A Potential Tool to Investigate Auditory Processing Disorder
Year of Dissertation:
Speech & Hearing Sciences
One of the hallmarks of auditory processing disorder (APD) is difficulty listening in background noise. This difficulty maybe related to the function of the medial olivocochlear (MOC) efferent system, which is hypothesized to provide an anti-masking effect that might aid in speech processing in noise. To test this hypothesis, we investigated the efferent anti-masking hypothesis via efferent-induced changes in the baseline levels of distortion product otoacoustic emissions (DPOAE) in a group of subjects suspected of having APD with speech-in-noise deficits matched for age and gender with a control group with less speech-in-noise deficits. There was no significant difference in audiometric thresholds between the groups. We examined not only the typical composite DPOAE, but also the two major components (overlap and reflection components), which determine the overall DPOAE level. We hypothesized that the group with speech-in-noise deficits would show reduced efferent effects relative to the control group. The findings did not support the efferent anti-masking hypothesis as efferent-induced changes in the composite DPOAE signal and the overlap component did not differ significantly between the two groups, but the statistical power was low. The separation of the two DPOAE components was beneficial in detecting efferent effects at the high frequency region where the DPOAE levels were lowest, and efferent effects were variable. The mean baseline levels and SNR of composite DPOAE and the overlap component were lower in the group with speech-in-noise deficits than the control group. This difference was not significant, but the statistical power was low. In addition, no significant correlations were found between performance on speech-in-noise tests and DPOAE change due to efferent activation across groups. Factors that might explain why the efferent anti-masking hypothesis was not supported are discussed.
Clustering Categorical Data Using Data Summaries and Spectral Techniques
Year of Dissertation:
Cluster analysis is an active area of research with applications in various fields including information retrieval, social sciences, bioinformatics, object recognition, and image segmentation (Jain et al., 1999). However, most algorithms are intended for numerical (continuous) data where proximity among data objects is naturally defined by virtue of their numerical properties. Although these algorithms can be used on categorical data, they are not designed to handle data properties typically found in this data type such as high dimensionality and lack of inherent relationships among attribute values. During the past decade, several algorithms have been designed for categorical data such as K-modes (Huang, 1998), STIRR (Gibson et al., 1998), CACTUS (Ganti et al., 1999), ROCK (Guha et al., 1999), COOLCAT (Barbara et al., 2002), LIMBO (Andritsos et al., 2004), CLICKS (Zaki et al., 2007), and others. Some of these algorithms exploit attribute relationships through data summaries such as attributes occurrence and co-occurrence frequencies while others use information entropy and links among data objects. In this thesis, we focus on using data summaries and spectral analysis to detect clustering structure in categorical data. Spectral techniques provide a relaxed solution to the discrete clustering problem which has been shown to be NP-hard (Drineas et al., 2004). Formulating the clustering problem as a graph partitioning problem and then finding the minimum normalized cut leads to a solution based on eigenvectors of the similarity matrix (i.e. Laplacian matrix). Spectral methods have been used in various algorithms and have been shown to find non-linearly separable clusters. Equally important, spectral analysis encompasses techniques for handling high-dimensional data since input data is projected into a lower-dimensional space where all computation/comparisons can be performed. Our approach is to extend spectral techniques to data summaries which are relatively less expensive to compute than data object similarity matrix for very large data sets. Our goal is to combine the benefits of spectral analysis with the relative low cost of computing data summaries. We have developed three algorithms for clustering categorical data using data summaries. Two of them use spectral techniques. Our test results on standard data sets and synthetic data sets show that our algorithms are competitive with current spectral and non-spectral algorithms for categorical data. Our algorithms provide a solution to the categorical data clustering problem that produces quality clustering and is scalable to large data sets.
Psychophysical and electrophysiological assessment of early visual processing and emotion recognition deficits in schizophrenia
Year of Dissertation:
Previous data suggest that patients with schizophrenia have preferential magnocellular (M) versus parvocellular (P) visual dysfunction. The goal of Experiment 1 was to characterize M–stream impairment in the patient population using a novel approach. Contrast thresholds at varying luminance levels were investigated. M– and P–biased responses were examined by using scotopic and photopic luminance conditions, respectively. Patients exhibited contrast threshold deficits during scotopic conditions, indicative of M–stream dysfunction. Further, the pattern of contrast threshold responses at photopic levels indicated relatively preserved patient P–pathway functionality. Experiment 2 used separate behavioral and electrophysiological paradigms to investigate contributions of low level visual pathway dysfunction in patients to emotion perceptual processing deficits. Contrast response curves for the dorsal (P1) and ventral (N170) pathways were elicited in response to contrast manipulated emotional faces. Results showed that the dorsal P1's pattern of response was impaired in patients while their N170 contrast response curves remained intact. Contributions of visual pathway dysfunction to impaired emotion recognition and affect–related processing, as indexed by the P250 amplitude, were then assessed. P250 activity in patients was reduced at all contrasts. Overall, across groups, the P1 component predicted both P250 amplitude and emotion recognition ability. Taken together, these data support the hypothesis that emotion recognition deficits in patients result from M/dorsal stream dysfunction. Experiment 3 examined the spatial–temporal oscillatory dynamics of schizophrenia patients during processing of complex visual stimuli. FFT spectrum activity and underlying generators of the delta, theta and alpha oscillatory frequencies were first assessed. Activity in the pre and post–stimulus intervals, and the ratio between them (event-related de/synchronization: ERD/ERS) were also evaluated. FFT data revealed controls had significantly greater alpha band activation in posterior electrodes as compared to patients. Conversely, patients exhibited greater theta–band activation over anterior electrodes versus controls. Topographical analysis suggested patients had abnormal underlying neural generators that gave rise to impaired theta and alpha–band activity. Further, instantaneous delta–band activity was significantly greater in patients during pre and post–stimulus intervals, possibly reflecting an overall generalized deficit. Finally, the data revealed that patients had a substantially reduced alpha ERD, highlighting their impairment in low level visual cortical gating mechanisms during processing of visual sensory inputs.
THE MARITAL STATE: PERSONAL STATUS LAWS, DISCOURSES OF REFORM, AND SECULARISM IN LEBANON
Year of Dissertation:
An important aspect of the modern Lebanese state is the arrangement of personal status laws, which consigns matters of marriage and its consequences to the several Islamic, Christian, and Jewish religious authorities. With the absence of civil jurisdictions, some individuals choose to get married under the civil laws of countries, such as France, Cyprus, and Turkey. Recurrent attempts to make civil marriage in Lebanon legal have proven to be controversial and ended ultimately in failure. The problem of marriage has accompanied the system of personal status since the formation of the Lebanese state under French Mandate after the collapse of the Ottoman Empire. This dissertation aims to offer an account of what is at stake in marriage. Based on ethnographic and archival research in Lebanon, it analyzes the terms of the controversies over legal reform, opinions about civil marriage, as well as the decisions of the Maronite Catholic, Sunni Islamic, and civil courts in matters of personal status. It argues that at stake in marriage is the very assumption upon which the modern Lebanese state rests, namely, that Lebanon consists essentially of a variety of religious communities each possessing a distinctive personal status. The formal articulation of that status is the several religious personal status jurisdictions that oversee marriage. This assumption gives rise to a specific configuration whereby marriage, religious communities, and the state, are interconnected. Rather than adopt a perspective that sees in the problem of marriage an opposition between secularism and religion, this study seeks its conditions in tensions internal to the secular itself, in the ambiguities between moral autonomy and religious belonging, freedom and equality, religion and law.
Real-time Measurement of Glial Progenitor Chemotactic Migration
Year of Dissertation:
Gliomas are the most commonly diagnosed form of central nervous system tumors, occurring primarily in adults. Like many malignant cancers, gliomas pathologically exhibit very aggressive spreading and lead to an average diagnosed survival expectancy of twelve months. This prognosis is due in large part to the uncontrolled division and migration of malignant tumor cells within healthy brain, which makes complete surgical resection impossible. Gliomas are known to contain numerous genotypic and phenotypic alterations that affect cell proliferation and survival. Previous research has indicated that both gliomas and their precursor cells exhibit distinct migration patterns in brain tissue, which may be induced by specific cytokines and their concentration gradients. Here, we investigated the migration of four brain tumor cell (BTCs) lines (U-87 MG; U-251 MG; Daoy; and XFMPDGF) and three RCAS-infected glial progenitor cell (GPCs) populations (GPCLacZ, GPCPDGF, and GPCkRas) toward various growth factors, including but not limited to: EGF, HGF/SF, PDGF-BB, and TGF-α. Mouse neural cultures of nestin expressing glial progenitor cells that have been engineered to display the Avian Leukosis Virus (ALV) target receptor, tv-a, were infected with the RCAS viral vectors RCAS-LacZ, RCAS-PDGF-B, and RCAS-kRas, and are referred to as Ntv-a glial progenitor cells (Ntv-a-GPCs). Upon successful infection, the Ntv-a-GPCs were capable of constitutively producing β-galactosidase, PDGF-B, and kRas, respectively. Additionally, when these viral vectors were injected into the brains of transgenic Ntv-a mouse pups, only the RCAS-PDGF-B viral titer induced gliomas (Dai et al, 2001; Holland et al, 2000). For the purpose of this work it is relevant to note that the Ntv-a cell lines used, with the exception of the XFMPDGF cells, were never injected into the mouse brain. Conversely, the XFMPDGF cell population represents an excised tumor from the brain of a ≤ 12-week-old Ntv-a transgenic mouse having the Ink4a-Arf -/- genetic background. After cranial injection with a RCAS-PDGF-B viral titer, tumors of the oligodendroglioma classification developed within 1-12 weeks in 100% of the injected mice (Tchougounova et al, 2007). We use these oligodendrogliomas as our mouse tumor model. Additionally, Daoy medulloblastomas were used to model the migration of the most common childhood human brain tumor, while the U-87 MG and U-251 MG were used to model glioblastomas, the most common adult brain tumor. The design and characterization of our stand-alone bridged µLane system demonstrated sustained steady-state concentration gradients over 2-3 days while enabling the diffusivity measurement of 0.82 ± 0.01 X 10-6 cm2 / s for Dextran molecules having a molecular weight similar to growth factors examined here. A modified version of this device was then used to examine the migratory response of human MB-derived Daoy, glioblastomas, and neonatal mouse glial progenitors, which has led to the characterization of differential invasion patterns that each cell line utilizes when exposed to various growth factor concentration gradients. This data is fundamental for understanding how BTCs and GPCs use cytokines to communicate with each other and alter cellular functions, specifically migration. Additionally, we demonstrate EGF-induced Akt activation in the migration of MB-derived Daoy cells, and the induced recruitment GPCs towards picomolar concentrations of both HGF and TGF-α. We suggest based on our results that glial progenitors consisting of varying genetic background can migrate in very distinct patterns and may contribute to the selectivity of glioma recruitment/or progenitor sorting.