Neuromorphic model for sound source segregation

While humans can easily segregate and track a speaker's voice in a loud noisy environment, most modern speech recognition systems still perform poorly in loud background noise. The computational principles behind auditory source segregation in humans is not yet fully understood. In this dissertation, we develop a computational model for source segregation inspired by auditory processing in the brain. To support the key principles behind the computational model, we conduct a series of electro-encephalography experiments using both simple tone-based stimuli and more natural speech stimulus. Most source segregation algorithms utilize some form of prior information about the target speaker or use more than one simultaneous recording of the noisy speech mixtures. Other methods develop models on the noise characteristics. Source segregation of simultaneous speech mixtures with a single microphone recording and no knowledge of the target speaker is still a challenge. Using the principle of temporal coherence, we develop a novel computational model that exploits the difference in the temporal evolution of features that belong to different sources to perform unsupervised monaural source segregation. While using no prior information about the target speaker, this method can gracefully incorporate knowledge about the target speaker to further enhance the segregation.Through a series of EEG experiments we collect neurological evidence to support the principle behind the model. Aside from its unusual structure and computational innovations, the proposed model provides testable hypotheses of the physiological mechanisms of the remarkable perceptual ability of humans to segregate acoustic sources, and of its psychophysical manifestations in navigating complex sensory environments. Results from EEG experiments provide further insights into the assumptions behind the model and provide motivation for future single unit studies that can provide more direct evidence for the principle of temporal coherence.

Phase noise and spur reduction in an array of direct digital synthesizers

Many applications, including communications, test and measurement, and radar, require the generation of signals with a high degree of spectral purity. One method for producing tunable, low-noise source signals is to combine the outputs of multiple direct digital synthesizers (DDSs) arranged in a parallel configuration. In such an approach, if all noise is uncorrelated across channels, the noise will decrease relative to the combined signal power, resulting in a reduction of sideband noise and an increase in SNR. However, in any real array, the broadband noise and spurious components will be correlated to some degree, limiting the gains achieved by parallelization. This thesis examines the potential performance benefits that may arise from using an array of DDSs, with a focus on several types of common DDS errors, including phase noise, phase truncation spurs, quantization noise spurs, and quantizer nonlinearity spurs. Measurements to determine the level of correlation among DDS channels were made on a custom 14-channel DDS testbed. The investigation of the phase noise of a DDS array indicates that the contribution to the phase noise from the DACs can be decreased to a desired level by using a large enough number of channels. In such a system, the phase noise qualities of the source clock and the system cost and complexity will be the main limitations on the phase noise of the DDS array. The study of phase truncation spurs suggests that, at least in our system, the phase truncation spurs are uncorrelated, contrary to the theoretical prediction. We believe this decorrelation is due to the existence of an unidentified mechanism in our DDS array that is unaccounted for in our current operational DDS model. This mechanism, likely due to some timing element in the FPGA, causes some randomness in the relative phases of the truncation spurs from channel to channel each time the DDS array is powered up. This randomness decorrelates the phase truncation spurs, opening the potential for SFDR gain from using a DDS array. The analysis of the correlation of quantization noise spurs in an array of DDSs shows that the total quantization noise power of each DDS channel is uncorrelated for nearly all values of DAC output bits. This suggests that a near N gain in SQNR is possible for an N-channel array of DDSs. This gain will be most apparent for low-bit DACs in which quantization noise is notably higher than the thermal noise contribution. Lastly, the measurements of the correlation of quantizer nonlinearity spurs demonstrate that the second and third harmonics are highly correlated across channels for all frequencies tested. This means that there is no benefit to using an array of DDSs for the problems of in-band quantizer nonlinearities. As a result, alternate methods of harmonic spur management must be employed.

A high-flux BEC source for mobile atom interferometers

Quantum sensors based on coherent matter-waves are precise measurement devices whose ultimate accuracy is achieved with Bose-Einstein condensates (BECs) in extended free fall. This is ideally realized in microgravity environments such as drop towers, ballistic rockets and space platforms. However, the transition from lab-based BEC machines to robust and mobile sources with comparable performance is a challenging endeavor. Here we report on the realization of a miniaturized setup, generating a flux of 4x10(5) quantum degenerate Rb-87 atoms every 1.6 s. Ensembles of 1 x 10(5) atoms can be produced at a 1 Hz rate. This is achieved by loading a cold atomic beam directly into a multi-layer atom chip that is designed for efficient transfer from laser-cooled to magnetically trapped clouds. The attained flux of degenerate atoms is on par with current lab-based BEC experiments while offering significantly higher repetition rates. Additionally, the flux is approaching those of current interferometers employing Raman-type velocity selection of laser-cooled atoms. The compact and robust design allows for mobile operation in a variety of demanding environments and paves the way for transportable high-precision quantum sensors.

Test-Retest Reliability of Resting-State Magnetoencephalography Power in Sensor and Source Space

Several studies have reported changes in spontaneous brain rhythms that could be used asclinical biomarkers or in the evaluation of neuropsychological and drug treatments in longitudinal studies using magnetoencephalography (MEG). There is an increasing necessity to use these measures in early diagnosis and pathology progression; however, there is a lack of studies addressing how reliable they are. Here, we provide the first test-retest reliability estimate of MEG power in resting-state at sensor and source space. In this study, we recorded 3 sessions of resting-state MEG activity from 24 healthy subjects with an interval of a week between each session. Power values were estimated at sensor and source space with beamforming for classical frequency bands: delta (2–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), low beta (13–20 Hz), high beta (20–30 Hz), and gamma (30–45 Hz). Then, test-retest reliability was evaluated using the intraclass correlation coefficient (ICC). We also evaluated the relation between source power and the within-subject variability. In general, ICC of theta, alpha, and low beta power was fairly high (ICC > 0.6) while in delta and gamma power was lower. In source space, fronto-posterior alpha, frontal beta, and medial temporal theta showed the most reliable profiles. Signal-to-noise ratio could be partially responsible for reliability as low signal intensity resulted inhigh within-subject variability, but also the inherent nature of some brain rhythms in resting-state might be driving these reliability patterns. In conclusion, our results described the reliability of MEG power estimates in each frequency band, which could be considered in disease characterization or clinical trials.

An identification procedure of multi-input Wiener models for the distortion analysis of nonlinear circuits

In this contribution, a system identification procedure of a two-input Wiener model suitable for the analysis of the disturbance behavior of integrated nonlinear circuits is presented. The identified block model is comprised of two linear dynamic and one static nonlinear block, which are determined using an parameterized approach. In order to characterize the linear blocks, an correlation analysis using a white noise input in combination with a model reduction scheme is adopted. After having characterized the linear blocks, from the output spectrum under single tone excitation at each input a linear set of equations will be set up, whose solution gives the coefficients of the nonlinear block. By this data based black box approach, the distortion behavior of a nonlinear circuit under the influence of an interfering signal at an arbitrary input port can be determined. Such an interfering signal can be, for example, an electromagnetic interference signal which conductively couples into the port of consideration. © 2011 Author(s).

Voice acoustic profi le of males exposed to occupational infrasound and low-frequency noise

Background: Vibroacoustic disease (VAD) is a systematic pathology characterized by the abnormal growth of extra-cellular matrices in the absence of infl ammatory processes, namely collagen and elastin, both of which are abundant in the basement membrane zone of the vocal folds. VAD can develop due to long-term exposure to infrasound and low-frequency noise (ILFN, <500 Hz). Mendes et al. (2006, 2008 and 2012) revealed that ILFN-exposed males and females presented an increased fundamental frequency (F0), decreased jitter %, and reduced maximum phonation frequency range, when compared with normative data. Temporal measures of maximum phonation time and S/Z ratio were generally reduced. Study Aims: Herein, the same voice acoustic parameters of 48 males, 36 airline pilots and 12 cabin crewmembers (age range 25-60 years) were studied, and the effects and interaction of age and years of ILFN exposure were investigated within those parameters. ILFN-exposure time (i.e. years of professional activity) ranged from 3.5 to 36 years. Materials and Methods: Spoken and sung phonatory tasks were recorded with a DA-P1 Tascam DAT and a C420III PP AKG head-worn microphone, positioned at 3 cm from the mouth. Acoustic analyses were performed using KayPENTAX Computer Speech Lab and Multi-Dimensional Voice Program. Results: Results revealed that even though pilots and cabin crewmembers were exposed to occupational environments with distinct (ILFN-rich) acoustical frequency distributions and sound pressure levels, differences in the vocal acoustic parameters were not evident. Analyzing data from both professional groups (N = 48) revealed that F0 increased signifi cantly with the number of years of professional activity. Conclusion: These results strongly suggest that the number of years of professional activity (i.e. total ILFN exposure time) had a signifi cant effect on F0. Furthermore, they may refl ect the histological changes specifi cally observed on the vocal folds of ILFN-exposed professionals.

Sparse time-frequency data analysis : a multi-scale approach

In this work, we further extend the recently developed adaptive data analysis method, the Sparse Time-Frequency Representation (STFR) method. This method is based on the assumption that many physical signals inherently contain AM-FM representations. We propose a sparse optimization method to extract the AM-FM representations of such signals. We prove the convergence of the method for periodic signals under certain assumptions and provide practical algorithms specifically for the non-periodic STFR, which extends the method to tackle problems that former STFR methods could not handle, including stability to noise and non-periodic data analysis. This is a significant improvement since many adaptive and non-adaptive signal processing methods are not fully capable of handling non-periodic signals. Moreover, we propose a new STFR algorithm to study intrawave signals with strong frequency modulation and analyze the convergence of this new algorithm for periodic signals. Such signals have previously remained a bottleneck for all signal processing methods. Furthermore, we propose a modified version of STFR that facilitates the extraction of intrawaves that have overlaping frequency content. We show that the STFR methods can be applied to the realm of dynamical systems and cardiovascular signals. In particular, we present a simplified and modified version of the STFR algorithm that is potentially useful for the diagnosis of some cardiovascular diseases. We further explain some preliminary work on the nature of Intrinsic Mode Functions (IMFs) and how they can have different representations in different phase coordinates. This analysis shows that the uncertainty principle is fundamental to all oscillating signals.

A survey on multi-output regression

In recent years, a plethora of approaches have been proposed to deal with the increasingly challenging task of multi-output regression. This paper provides a survey on state-of-the-art multi-output regression methods, that are categorized as problem transformation and algorithm adaptation methods. In addition, we present the mostly used performance evaluation measures, publicly available data sets for multi-output regression real-world problems, as well as open-source software frameworks.

Atmospheric calibration for sub-millimeter radio astronomy

This thesis is focused on improving the calibration accuracy of sub-millimeter astronomical observations. The wavelength range covered by observational radio astronomy has been extended to sub-millimeter and far infrared with the advancement of receiver technology in recent years. Sub-millimeter observations carried out with airborne and ground-based telescopes typically suffer from 10% to 90% attenuation of the astronomical source signals by the terrestrial atmosphere. The amount of attenuation can be derived from the measured brightness of the atmospheric emission. In order to do this, the knowledge of the atmospheric temperature and chemical composition, as well as the frequency-dependent optical depth at each place along the line of sight is required. The altitude-dependent air temperature and composition are estimated using a parametrized static atmospheric model, which is described in Chapter 2, because direct measurements are technically and financially infeasible. The frequency dependent optical depth of the atmosphere is computed with a radiative transfer model based on the theories of quantum mechanics and, in addition, some empirical formulae. The choice, application, and improvement of third party radiative transfer models are discussed in Chapter 3. The application of the calibration procedure, which is described in Chapter 4, to the astronomical data observed with the SubMillimeter Array Receiver for Two Frequencies (SMART), and the German REceiver for Astronomy at Terahertz Frequencies (GREAT), is presented in Chapters 5 and 6. The brightnesses of atmospheric emission were fitted consistently to the simultaneous multi-band observation data from GREAT at 1.2 ∼ 1.4 and 1.8 ∼ 1.9 THz with a single set of parameters of the static atmospheric model. On the other hand, the cause of the inconsistency between the model parameters fitted from the 490 and 810 GHz data of SMART is found to be the lack of calibration of the effective cold load temperature. Besides the correctness of atmospheric modeling, the stability of the receiver is also important to achieving optimal calibration accuracy. The stabilities of SMART and GREAT are analyzed with a special calibration procedure, namely the “load calibration". The effects of the drift and fluctuation of the receiver gain and noise temperature on calibration accuracy are discussed in Chapters 5 and 6. Alternative observing strategies are proposed to combat receiver instability. The methods and conclusions presented in this thesis are applicable to the atmospheric calibration of sub-millimeter astronomical observations up to at least 4.7 THz (the H channel frequency of GREAT) for observations carried out from ∼ 4 to 14 km altitude. The procedures for receiver gain calibration and stability test are applicable to other instruments using the same calibration approach as that for SMART and GREAT. The structure of the high performance, modular, and extensible calibration program used and further developed for this thesis work is presented in the Appendix C.

Climate variability of the last 1000 years in the NW Pacific: high resolution, multi-biomarker records from Lake Toyoni

The East Asian Monsoon (EAM) is an active component of the global climate system and has a profound social and economic impact in East Asia and its surrounding countries. Its impact on regional hydrological processes may influence society through industrial water supplies, food productivity and energy use. In order to predict future rates of climate change, reliable and accurate reconstructions of regional temperature and rainfall are required from all over the world to test climate models and better predict future climate variability. Hokkaido is a region which has limited palaeo-climate data and is sensitive to climate change. Instrumental data show that the climate in Hokkaido is influenced by the East Asian Monsoon (EAM), however, instrumental data is limited to the past ~150 years. Therefore down-core climate reconstructions, prior to instrumental records, are required to provide a better understanding of the long-term behaviour of the climate drivers (e.g. the EAM, Westerlies, and teleconnections) in this region. The present study develops multi-proxy reconstructions to determine past climatic and hydrologic variability in Japan over the past 1000 years and aid in understanding the effects of the EAM and the Westerlies independently and interactively. A 250-cm long sediment core from Lake Toyoni, Hokkaido was retrieved to investigate terrestrial and aquatic input, lake temperature and hydrological changes over the past 1000-years within Lake Toyoni and its catchment using X-Ray Fluorescence (XRF) data, alkenone palaeothermometry, the molecular and hydrogen isotopic composition of higher plant waxes (δD(HPW)). Here, we conducted the first survey for alkenone biomarkers in eight lakes in the Hokkaido, Japan. We detected the occurrence of alkenones within the sediments of Lake Toyoni. We present the first lacustrine alkenone record from Japan, including genetic analysis of the alkenone producer. C37 alkenone concentrations in surface sediments are 18µg C37 g−1 of dry sediment and the dominant alkenone is C37:4. 18S rDNA analysis revealed the presence of a single alkenone producer in Lake Toyoni and thus a single calibration is used for reconstructing lake temperature based on alkenone unsaturation patterns. Temperature reconstructions over the past 1000 years suggest that lake water temperatures varies between 8 and 19°C which is in line with water temperature changes observed in the modern Lake Toyoni. The alkenone-based temperature reconstruction provides evidence for the variability of the EAM over the past 1000 years. The δD(HPW) suggest that the large fluctuations (∼40‰) represent changes in temperature and source precipitation in this region, which is ultimately controlled by the EAM system and therefore a proxy for the EAM system. In order to complement the biomarker reconstructions, the XRF data strengthen the lake temperature and hydrological reconstructions by providing information on past productivity, which is controlled by the East Asian Summer monsoon (EASM) and wind input into Lake Toyoni, which is controlled by the East Asian Winter Monsoon (EAWM) and the Westerlies. By combining the data generated from XRF, alkenone palaeothermometry and the δD(HPW) reconstructions, we provide valuable information on the EAM and the Westerlies, including; the timing of intensification and weakening, the teleconnections influencing them and the relationship between them. During the Medieval Warm Period (MWP), we find that the EASM dominated and the EAWM was suppressed, whereas, during the Little Ice Age (LIA), the influence of the EAWM dominated with time periods of increased EASM and Westerlies intensification. The El Niño Southern Oscillation (ENSO) significantly influenced the EAM; a strong EASM occurred during El Niño conditions and a strong EAWM occurred during La Niña. The North Atlantic Oscillation, on the other hand, was a key driver of the Westerlies intensification; strengthening of the Westerlies during a positive NAO phase and weakening of the Westerlies during a negative NAO phase. A key finding from this study is that our data support an anti-phase relationship between the EASM and the EAWM (e.g. the intensification of the EASM and weakening of the EAWM and vice versa) and that the EAWM and the Westerlies vary independently from each other, rather than coincide as previously suggested in other studies.

Gastrointestinal stromal tumors: Arbutus unedo L. as a source of novel chemotherapeutics & mechanisms behind metastasis

Gastrointestinal stromal tumors (GIST) are mesenchymal neoplasms frequently caused by a gain of function mutation in KIT or PDGFRα, two tyrosine kinase receptors (TKR). For this reason, they are successfully treated with imatinib, a tyrosine kinase inhibitor (TKI). However, the therapy is typically long-term ineffective due to imatinib resistance, which represents the main issue in the clinic of GISTs. Although numerous efforts have been made in the last two decades to develop novel therapies for imatinib-resistant GISTs, the approvals of multi-target TKIs have only improved the clinical outcomes modestly. Emblematic is the recent failure of ripretinib in the phase III INTRIGUE trial, decisively marking the end of the paradigm only based on the central role of KIT secondary mutations in imatinib resistance, and the consequent seeking of multi-target TKIs as the solution. Consistent with this clinical result, preclinical studies have revealed numerous mechanisms of resistance that are not targetable with multi-target TKIs, indicating that imatinib resistance is more multifaceted than initially hypothesized and explaining the modest efficacy of these latter. In this scenario, the absence of drugs capable of long-term counteracting the rise of imatinib-resistant subclones unavoidably leads to progressive disease and metastasis. In particular, the onset of metastases remarkably impacts the median overall survival and determines the most GIST-related deaths. Therefore, new therapy proposals are needed. Here, we present two project lines investigating novel strategies to counteract imatinib-resistant GISTs.

Exploration of X-ray, optical and radio emission of the radio-galaxy 3C 277.3

This thesis explores the X-ray nuclear and extended properties of the radio galaxy 3C 277.3, where a recent optical observation performed with the multi-unit spectroscopic explorer (MUSE) has revealed star-forming regions triggered by the propagation of non-thermal plasma in the intergalactic medium. This work aims to study the nuclear engine and its environment and, possibly, discover signatures of non-thermal plasma-gas interaction at high energies. 3C 277.3 was observed with the Chandra satellite five times from 2010 to 2014 for a total of about 200 ks. Data in the Chandra public archive were retrieved and analyzed. When necessary, the different pointings were combined to improve the signal-to-noise ratio. A detailed analysis of the Chandra image (obtained by combining all the observations) has revealed several emission regions. In addition to a bright nucleus, two jet knots and the northern hot spot were clearly detected by overlapping the X-ray data to a VLA map of the source at 1.4 GHz. An X-ray spectral analysis was performed for all these structures. Finally, the X-ray image was over-imposed on the MUSE data.

Caffeine As An Indicator Of Estrogenic Activity In Source Water.

Caffeine has already been used as an indicator of anthropogenic impacts, especially the ones related to the disposal of sewage in water bodies. In this work, the presence of caffeine has been correlated with the estrogenic activity of water samples measured using the BLYES assay. After testing 96 surface water samples, it was concluded that caffeine can be used to prioritize samples to be tested for estrogenic activity in water quality programs evaluating emerging contaminants with endocrine disruptor activity.

Doehlert Design-desirability Function Multi-criteria Optimal Separation Of 17 Phenolic Compounds From Extra-virgin Olive Oil By Capillary Zone Electrophoresis.

In Brazil, the consumption of extra-virgin olive oil (EVOO) is increasing annually, but there are no experimental studies concerning the phenolic compound contents of commercial EVOO. The aim of this work was to optimise the separation of 17 phenolic compounds already detected in EVOO. A Doehlert matrix experimental design was used, evaluating the effects of pH and electrolyte concentration. Resolution, runtime and migration time relative standard deviation values were evaluated. Derringer's desirability function was used to simultaneously optimise all 37 responses. The 17 peaks were separated in 19min using a fused-silica capillary (50μm internal diameter, 72cm of effective length) with an extended light path and 101.3mmolL(-1) of boric acid electrolyte (pH 9.15, 30kV). The method was validated and applied to 15 EVOO samples found in Brazilian supermarkets.