Search

Thesis: S.M., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018., Cataloged from PDF version of thesis., Includes bibliographical references (pages 73-77)., Optical coherence tomography (OCT) is a label-free optical imaging modality that allows non-invasive in-depth visualization of microscopic structures in samples. With a typical resolution of 10-15 [mu]m and a penetration of up to a few mm, OCT is widely used for medical diagnoses in fields such as ophthalmology and cardiology. However, the more common diagnostic tool in the microscopic regime of medical imaging is histology, an invasive technique requiring tissue biopsy. Its resolution can be as small as 0.2 [mu]m, allowing the visualization of subcellular structures. To help bridge this gap between OCT and histology, ultrahigh-resolution OCT systems have been developed, with resolutions on the order of 1 [mu]m. Yet their application remains limited, since they employ shorter-wavelength sources, reducing penetration in tissue. We have designed and built a spectral-domain ultrahigh-resolution, deep-penetration OCT system centered at 1290 nm with axial and lateral resolutions of 2 and 5 [mu]m, respectively. To our knowledge, this is the best axial resolution obtained for a highspeed OCT system centered this deeply in the infrared. We demonstrate imaging of the cardiac conduction system, which could eventually be used for intraoperative identification of conducting tissue. In addition, we show images of the corneo-scleral angle, which could help properly diagnose primary angle-closure glaucoma. Other potential applications are also discussed., by Liane Bernstein., S.M.

Thesis: S.M., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018., Cataloged from PDF version of thesis., Includes bibliographical references (pages 73-77)., Optical coherence tomography (OCT) is a label-free optical imaging modality that allows non-invasive in-depth visualization of microscopic structures in samples. With a typical resolution of 10-15 [mu]m and a penetration of up to a few mm, OCT is widely used for medical diagnoses in fields such as ophthalmology and cardiology. However, the more common diagnostic tool in the microscopic regime of medical imaging is histology, an invasive technique requiring tissue biopsy. Its resolution can be as small as 0.2 [mu]m, allowing the visualization of subcellular structures. To help bridge this gap between OCT and histology, ultrahigh-resolution OCT systems have been developed, with resolutions on the order of 1 [mu]m. Yet their application remains limited, since they employ shorter-wavelength sources, reducing penetration in tissue. We have designed and built a spectral-domain ultrahigh-resolution, deep-penetration OCT system centered at 1290 nm with axial and lateral resolutions of 2 and 5 [mu]m, respectively. To our knowledge, this is the best axial resolution obtained for a highspeed OCT system centered this deeply in the infrared. We demonstrate imaging of the cardiac conduction system, which could eventually be used for intraoperative identification of conducting tissue. In addition, we show images of the corneo-scleral angle, which could help properly diagnose primary angle-closure glaucoma. Other potential applications are also discussed., by Liane Bernstein., S.M.

Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017., This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections., Cataloged from student-submitted PDF version of thesis., Includes bibliographical references (page 83)., Intensive care unit (ICU) patients undergo detailed monitoring so that copious information regarding their condition is available to support clinical decision-making. Full utilization of the data depends heavily on its quantity, quality and manner of presentation to the physician at the bedside of a patient. In this thesis, we implemented a visualization system to aid ICU clinicians in collecting, processing, and displaying available ICU data. Our goals for the system are: to be able to receive large quantities of patient data from various sources, to compute complex functions over the data that are able to quantify an ICU patient's condition, to plot the data using a clean and interactive interface, and to be capable of live plot updates upon receiving new data. We made significant headway toward our goals, and we succeeded in creating a highly adaptable visualization system that future developers and users will be able to customize., by Shruthi Narayanan., M. Eng.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018., Cataloged from PDF version of thesis., Includes bibliographical references (pages 120-134)., Electroencephalography (EEG) features are known to predict neurological outcomes of patients in coma after cardiac arrest, but the association between EEG features and outcomes is time-dependent. Recent advances in machine learning allow temporally-dependent features to be learned from the EEG waveforms in a fully-automated way, allowing for faster, better-calibrated and more reliable prognostic predictions. In this thesis, we discuss three major contributions to the problem of coma prognostication after cardiac arrest: (1) the collection of the world's largest multi-center EEG database for patients in coma after cardiac arrest, (2) the development of time-dependent, interpretable, feature-based EEG models that may be used for both risk-scoring and decision support at the bedside, and (3) a careful comparison of the performance and utility of feature-based techniques to that of representation learning models that fully-automate the extraction of time-dependent features for outcome prognostication., by Mohammad Mahdi Ghassemi., Ph. D.

Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017., This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections., Cataloged from student-submitted PDF version of thesis., Includes bibliographical references (page 83)., Intensive care unit (ICU) patients undergo detailed monitoring so that copious information regarding their condition is available to support clinical decision-making. Full utilization of the data depends heavily on its quantity, quality and manner of presentation to the physician at the bedside of a patient. In this thesis, we implemented a visualization system to aid ICU clinicians in collecting, processing, and displaying available ICU data. Our goals for the system are: to be able to receive large quantities of patient data from various sources, to compute complex functions over the data that are able to quantify an ICU patient's condition, to plot the data using a clean and interactive interface, and to be capable of live plot updates upon receiving new data. We made significant headway toward our goals, and we succeeded in creating a highly adaptable visualization system that future developers and users will be able to customize., by Shruthi Narayanan., M. Eng.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018., Cataloged from PDF version of thesis., Includes bibliographical references (pages 120-134)., Electroencephalography (EEG) features are known to predict neurological outcomes of patients in coma after cardiac arrest, but the association between EEG features and outcomes is time-dependent. Recent advances in machine learning allow temporally-dependent features to be learned from the EEG waveforms in a fully-automated way, allowing for faster, better-calibrated and more reliable prognostic predictions. In this thesis, we discuss three major contributions to the problem of coma prognostication after cardiac arrest: (1) the collection of the world's largest multi-center EEG database for patients in coma after cardiac arrest, (2) the development of time-dependent, interpretable, feature-based EEG models that may be used for both risk-scoring and decision support at the bedside, and (3) a careful comparison of the performance and utility of feature-based techniques to that of representation learning models that fully-automate the extraction of time-dependent features for outcome prognostication., by Mohammad Mahdi Ghassemi., Ph. D.

Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017., This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections., Cataloged from student-submitted PDF version of thesis., Includes bibliographical references (page 83)., Intensive care unit (ICU) patients undergo detailed monitoring so that copious information regarding their condition is available to support clinical decision-making. Full utilization of the data depends heavily on its quantity, quality and manner of presentation to the physician at the bedside of a patient. In this thesis, we implemented a visualization system to aid ICU clinicians in collecting, processing, and displaying available ICU data. Our goals for the system are: to be able to receive large quantities of patient data from various sources, to compute complex functions over the data that are able to quantify an ICU patient's condition, to plot the data using a clean and interactive interface, and to be capable of live plot updates upon receiving new data. We made significant headway toward our goals, and we succeeded in creating a highly adaptable visualization system that future developers and users will be able to customize., by Shruthi Narayanan., M. Eng.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018., Cataloged from PDF version of thesis., Includes bibliographical references (pages 120-134)., Electroencephalography (EEG) features are known to predict neurological outcomes of patients in coma after cardiac arrest, but the association between EEG features and outcomes is time-dependent. Recent advances in machine learning allow temporally-dependent features to be learned from the EEG waveforms in a fully-automated way, allowing for faster, better-calibrated and more reliable prognostic predictions. In this thesis, we discuss three major contributions to the problem of coma prognostication after cardiac arrest: (1) the collection of the world's largest multi-center EEG database for patients in coma after cardiac arrest, (2) the development of time-dependent, interpretable, feature-based EEG models that may be used for both risk-scoring and decision support at the bedside, and (3) a careful comparison of the performance and utility of feature-based techniques to that of representation learning models that fully-automate the extraction of time-dependent features for outcome prognostication., by Mohammad Mahdi Ghassemi., Ph. D.

Thesis: M. Eng. in Computer Science and Molecular Biology, Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2015., This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections., Cataloged from student-submitted PDF version of thesis., Includes bibliographical references (pages 73-74)., We investigated the relationship between initial lactate levels and ICU patient outcomes using the MIMIC II (version 2.6) database. We divided ICU admissions based on their initial lactate measurement into three groups: admissions with high lactate (above 4.0 mmol/L), admissions with medium lactate (between 2.0 mmol/L and 4.0 mmol/L), and admissions with low lactate (below 2.0 mmol/L). In addition to the ICU population as a whole, we studied sepsis patients using three different criteria (Martin, Angus, and infection with SIRS). We found that increased lactate levels were associated with a higher ICU mortality, higher 30 day mortality, longer ICU length of stay, and higher SOFA and SAPS I severity scores in all ICU admissions and in all three sepsis cohorts. Sepsis patients with high initial lactate levels were the most severely ill of all the patient populations. Sepsis patients identified with the Martin criteria who had high lactate levels had the worst outcomes of the three sepsis cohorts, but had similar average severity scores. This suggests that knowing lactate levels may give predictive value in addition to severity scores. We also investigated the relationship between initial lactate, change in lactate from the first measurement to the second measurement, and ICU mortality. We found that patients with high initial lactate levels in combination with an increase in lactate level typically had poorer outcomes than patients with high initial lactate levels with a decrease in lactate level., by Hilary G. Mulholland., M. Eng. in Computer Science and Molecular Biology

Thesis: M. Eng. in Computer Science and Molecular Biology, Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2015., This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections., Cataloged from student-submitted PDF version of thesis., Includes bibliographical references (pages 73-74)., We investigated the relationship between initial lactate levels and ICU patient outcomes using the MIMIC II (version 2.6) database. We divided ICU admissions based on their initial lactate measurement into three groups: admissions with high lactate (above 4.0 mmol/L), admissions with medium lactate (between 2.0 mmol/L and 4.0 mmol/L), and admissions with low lactate (below 2.0 mmol/L). In addition to the ICU population as a whole, we studied sepsis patients using three different criteria (Martin, Angus, and infection with SIRS). We found that increased lactate levels were associated with a higher ICU mortality, higher 30 day mortality, longer ICU length of stay, and higher SOFA and SAPS I severity scores in all ICU admissions and in all three sepsis cohorts. Sepsis patients with high initial lactate levels were the most severely ill of all the patient populations. Sepsis patients identified with the Martin criteria who had high lactate levels had the worst outcomes of the three sepsis cohorts, but had similar average severity scores. This suggests that knowing lactate levels may give predictive value in addition to severity scores. We also investigated the relationship between initial lactate, change in lactate from the first measurement to the second measurement, and ICU mortality. We found that patients with high initial lactate levels in combination with an increase in lactate level typically had poorer outcomes than patients with high initial lactate levels with a decrease in lactate level., by Hilary G. Mulholland., M. Eng. in Computer Science and Molecular Biology

Thesis: M. Eng. in Computer Science and Molecular Biology, Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2015., This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections., Cataloged from student-submitted PDF version of thesis., Includes bibliographical references (pages 73-74)., We investigated the relationship between initial lactate levels and ICU patient outcomes using the MIMIC II (version 2.6) database. We divided ICU admissions based on their initial lactate measurement into three groups: admissions with high lactate (above 4.0 mmol/L), admissions with medium lactate (between 2.0 mmol/L and 4.0 mmol/L), and admissions with low lactate (below 2.0 mmol/L). In addition to the ICU population as a whole, we studied sepsis patients using three different criteria (Martin, Angus, and infection with SIRS). We found that increased lactate levels were associated with a higher ICU mortality, higher 30 day mortality, longer ICU length of stay, and higher SOFA and SAPS I severity scores in all ICU admissions and in all three sepsis cohorts. Sepsis patients with high initial lactate levels were the most severely ill of all the patient populations. Sepsis patients identified with the Martin criteria who had high lactate levels had the worst outcomes of the three sepsis cohorts, but had similar average severity scores. This suggests that knowing lactate levels may give predictive value in addition to severity scores. We also investigated the relationship between initial lactate, change in lactate from the first measurement to the second measurement, and ICU mortality. We found that patients with high initial lactate levels in combination with an increase in lactate level typically had poorer outcomes than patients with high initial lactate levels with a decrease in lactate level., by Hilary G. Mulholland., M. Eng. in Computer Science and Molecular Biology

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1986., MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING, Bibliography: leaves 238-239., by Scott David Greenwald., M.S.

Thesis (M.Eng. and S.B.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2002., Includes bibliographical references (leaf 31)., by Christine Lieu., M.Eng.and S.B.

Thesis (M.Eng. and S.B.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2001., Includes bibliographical references (leaf 40)., by Jack V. Chung., M.Eng.and S.B.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2000., Includes bibliographical references (leaves 59-60)., by Yi Zhang., S.M.

Thesis (M.Eng. and S.B.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2001., Includes bibliographical references (leaf 40)., by Jack V. Chung., M.Eng.and S.B.

Thesis (M.Eng. and S.B.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2002., Includes bibliographical references (leaf 31)., by Christine Lieu., M.Eng.and S.B.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2000., Includes bibliographical references (leaves 59-60)., by Yi Zhang., S.M.

Thesis (M.Eng. and S.B.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2001., Includes bibliographical references (leaf 40)., by Jack V. Chung., M.Eng.and S.B.

Thesis (M.Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, September 2001., August 13, 2001.", Includes bibliographical references (leaf 40, first group)., by Emily A. Liu., M.Eng.

Thesis (M.Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, September 2001., August 13, 2001.", Includes bibliographical references (leaf 40, first group)., by Emily A. Liu., M.Eng.

Thesis (M.Eng. and S.B.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2002., Includes bibliographical references (leaf 31)., by Christine Lieu., M.Eng.and S.B.

Thesis (M.Eng. and S.B.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2001., Includes bibliographical references (leaf 40)., by Jack V. Chung., M.Eng.and S.B.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2009., This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections., Cataloged from student-submitted PDF version of thesis., Includes bibliographical references (p. 91-93)., A signal quality assessment scheme for the photoplethysmogram waveform recorded by a pulse oximeter has been created. The signal quality algorithm uses statistical methods on time-series and spectral analysis to locate high-frequency segments of the photoplethysmogram waveform. A photoplethysmogram pulse onset detector has been implemented for heart rate estimation. Application of the signal quality metric and photoplethysmogram pulse onset detector are demonstrated in an algorithm which suppresses false electrocardiogram critical arrhythmia alarms issued by bedside monitors in hospital intensive care units., by Anagha Vishwas Deshmane., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2009., Includes bibliographical references (p. 115-118)., The effect of signal quality on the accuracy of cardiac output (CO) estimation from arterial blood pressure (ABP) was evaluated using data from the Multi-Parameter Intelligent Patient Monitoring for Intensive Care (MIMIC) II database. Thermodilution CO (TCO) was the gold standard, and a total of 121 records with 1,497 TCO measurements were used. Six lumped-parameter and systolic area CO estimators were tested, using ABP features and a robust heart rate (HR) estimate. Signal quality indices for ABP and HR were calculated using previously described metrics. For retrospective analysis, results showed that the Liljestrand estimator yielded the lowest error for all levels of signal quality and for any single estimator when using five or more calibration points. Increasing signal quality decreased error and only marginally reduced the amount of available data, as a signal quality level of 90% preserved sufficient data for almost continuous CO estimation. At the recommended signal quality thresholds, the lowest gross root mean square normalized error (RMSNE) was found to be 15.4% (or 0.74 L/min) and average RMSNE was 13.7% (0.71 L/min). Based on these results, a linear combination (LC) of the six CO estimation methods was developed and proved superior to all other methods when up to 13 TCO calibration values were used. The clinical utility of the CO estimates were examined by correlating changes in four vasoactive medication doses with corresponding changes in estimated resistance, which was derived from mean ABP and estimated CO., (cont.) Both the Liljestrand estimator and the LC estimator were used to estimate CO. Regression analysis failed to show a clear correlation between dose level and estimated resistance for either estimator except for neosynephrine, revealing the limitations of current SQI methods in ensuring signal fidelity. Examples of types of non-physiological or artifactual ABP waveforms are shown, and a potential damping detection method is proposed., by Tiffany Chen., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2008., Includes bibliographical references (p. 71-73)., Biomarkers have become a growing field for disease research and diagnostics. For the infectious, deadly disease, tuberculosis (TB), the discovery of effective TB biomarkers in the human breath may enable the development of a cheap, fast, and accurate TB diagnostic to fight against this epidemic. The Charles Stark Draper Laboratory is employing a differential mobility spectrometer (DMS), in order to identify key biomarkers for TB. Biomarker candidates appear as peaks in the DMS sensor data at particular retention times and DC compensation voltages. Possible biomarkers are identified through well-known classification techniques, K-nearest neighbors and support vector machines. This research is focused on the development of a toolbox that will aid in the discovery and in-depth understanding of the TB breath biomarkers. This thesis demonstrates the algorithms and capabilities of peak quantification, characterization, RIP-based alignment, and siloxane-based alignment on protein and bacterial sensor data, including those from initial Mycobacterium tuberculosis samples. The results show that siloxane-based alignment with high-intensity peak landmarks effectively increase DMS peak stabilities and signal-to-noise ratios. These new tools will be important for the analysis of specific TB biomarker candidates as our group at the Charles Stark Draper Laboratory receive more samples of M. tuberculosis., by Amy Yuen-Lee Tang., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2008., Includes bibliographical references (p. 71-73)., Biomarkers have become a growing field for disease research and diagnostics. For the infectious, deadly disease, tuberculosis (TB), the discovery of effective TB biomarkers in the human breath may enable the development of a cheap, fast, and accurate TB diagnostic to fight against this epidemic. The Charles Stark Draper Laboratory is employing a differential mobility spectrometer (DMS), in order to identify key biomarkers for TB. Biomarker candidates appear as peaks in the DMS sensor data at particular retention times and DC compensation voltages. Possible biomarkers are identified through well-known classification techniques, K-nearest neighbors and support vector machines. This research is focused on the development of a toolbox that will aid in the discovery and in-depth understanding of the TB breath biomarkers. This thesis demonstrates the algorithms and capabilities of peak quantification, characterization, RIP-based alignment, and siloxane-based alignment on protein and bacterial sensor data, including those from initial Mycobacterium tuberculosis samples. The results show that siloxane-based alignment with high-intensity peak landmarks effectively increase DMS peak stabilities and signal-to-noise ratios. These new tools will be important for the analysis of specific TB biomarker candidates as our group at the Charles Stark Draper Laboratory receive more samples of M. tuberculosis., by Amy Yuen-Lee Tang., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2006., Includes bibliographical references (p. 62-64)., This paper presents a de-identification study at the Harvard-MIT Division of Health Science and Technology (HST) to automatically de-identify confidential patient information from text medical records used in intensive care units (ICUs). Patient records are a vital resource in medical research. Before such records can be made available for research studies, protected health information (PHI) must be thoroughly scrubbed according to HIPAA specifications to preserve patient confidentiality. Manual de-identification on large databases tends to be prohibitively expensive, time-consuming and prone to error, making a computerized algorithm an urgent need for large-scale de-identification purposes. We have developed an automated pattern-matching deidentification algorithm that uses medical and hospital-specific information. The current version of the algorithm has an overall sensitivity of around 0.87 and an approximate positive predictive value of 0.63. In terms of sensitivity, it performs significantly better than 1 person (0.81) but not quite as well as a consensus of 2 human de-identifiers (0.94). The algorithm will be published as open-source software, and the de-identified medical records will be incorporated into HST's Multi-Parameter Intelligent Monitoring for Intensive Care (MIMIC II) physiologic database., by Ishna Neamatullah., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2006., Includes bibliographical references (p. 73-74)., Cardiac output (CO) is a cardinal parameter of cardiovascular state, and a fundamental determinant of global oxygen delivery. Historically, measurement of CO has been limited to critically-ill patients, using invasive indicator-dilution methods such as thermodilution via Swan-Ganz lines, which carry risks. Over the past century, the premise that CO could be estimated by analysis of the arterial blood pressure (ABP) waveform has captured the attention of many investigators. This approach of estimating CO is minimally invasive, cheap, and can be done continuously as long as ABP waveforms are available. Over a dozen different methods of estimating CO from ABP waveforms have been proposed and some are commercialized. However, the effectiveness of this approach is nebular. Performance validation studies in the past have mostly been conducted on a small set of subjects under well-controlled laboratory conditions. It is entirely possible that there will be circumstances in real world clinical practice in which CO estimation produces inaccurate results. In this thesis, our goals are to (1) build a computational system that estimates CO using 11 of the established methods; (2) evaluate and compare the performance of the CO estimation methods on a large set clinical data, using the simultaneously available thermodilution CO measurements as gold-standard; and (3) design and evaluate an algorithm that identifies and eliminates ABP waveform segments of poor quality. Out of the 11 CO estimation methods studied, there is one method (Liljestrand method) that is clearly more accurate than the rest. Across our study population of 120 subjects, the Liljestrand method has an error distribution with a 1 standard deviation error of 0.8 L/min, which is roughly twice that of thermodilution CO. These results suggest that although CO estimation methods may not generate the most precise values, they are still useful for detecting significant (>1 L/min) changes in CO., by James Xin Sun., M.Eng.

Includes bibliographical references (leaves 87-89)., Thesis (M. Eng.)--Harvard-MIT Division of Health Sciences and Technology, 2007., (cont.) The EWS models were then tested in a forward, casual manner on a random cohort of 500 ICU patients to mimic the patients' stay in the unit. The model with the highest performance achieved a sensitivity of 0.85 and a positive predictive value (PPV) of 0.70. Of the 35 episodes of hypotension despite fluid resuscitation present in the random patient dataset, the model provided early warnings for 29 episodes with a mean early warning time of 582 ± 355 minutes., Early goal-directed therapy (EGDT) in severe sepsis and septic shock has shown to provide substantial benefits in patient outcomes. However, these preventive therapeutic interventions are contingent upon an early detection or suspicion of the underlying septic etiology. Detection of sepsis in the early stages can be difficult, as the initial pathogenesis can occur while the patient is still displaying normal vital signs. This study focuses on developing an early warning system (EWS) to provide clinicians with a forewarning of an impending hypotensive crisis-thus allowing for EGDT intervention. Research was completed in three main stages: (1) generating an annotated septic shock dataset, (2) constructing multivariate logistic regression EWS models using the annotated dataset, and (3) testing the EWS models in a forward, causal manner on a random cohort of patients to simulate performance in a real-life ICU setting. The annotated septic shock dataset was created using the Multi-parameter Intelligent Monitoring for Intensive Care II (MIMIC II) database. Automated pre-annotations were generated using search criteria designed to identify two patient types: (1) sepsis patients who do not progress to septic shock, and (2) sepsis patient who progress to septic shock. Currently, manual review by expert clinicians to verify the pre-annotations has not been completed. Six separate EWS models were constructed using the annotated septic shock dataset. The multivariate logistic regression EWS models were trained to differentiate between 107 high-risk sepsis patients of whom 39 experienced a hypotensive crisis and 68 who remained stable. The models were tested using 7-fold cross validation; the mean area under the receiver operating characteristic (ROC) curve for the best model was 0.940 ± 0.038., by Dewang Shavdia., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2006., Includes bibliographical references (p. 73-74)., Cardiac output (CO) is a cardinal parameter of cardiovascular state, and a fundamental determinant of global oxygen delivery. Historically, measurement of CO has been limited to critically-ill patients, using invasive indicator-dilution methods such as thermodilution via Swan-Ganz lines, which carry risks. Over the past century, the premise that CO could be estimated by analysis of the arterial blood pressure (ABP) waveform has captured the attention of many investigators. This approach of estimating CO is minimally invasive, cheap, and can be done continuously as long as ABP waveforms are available. Over a dozen different methods of estimating CO from ABP waveforms have been proposed and some are commercialized. However, the effectiveness of this approach is nebular. Performance validation studies in the past have mostly been conducted on a small set of subjects under well-controlled laboratory conditions. It is entirely possible that there will be circumstances in real world clinical practice in which CO estimation produces inaccurate results. In this thesis, our goals are to (1) build a computational system that estimates CO using 11 of the established methods; (2) evaluate and compare the performance of the CO estimation methods on a large set clinical data, using the simultaneously available thermodilution CO measurements as gold-standard; and (3) design and evaluate an algorithm that identifies and eliminates ABP waveform segments of poor quality. Out of the 11 CO estimation methods studied, there is one method (Liljestrand method) that is clearly more accurate than the rest. Across our study population of 120 subjects, the Liljestrand method has an error distribution with a 1 standard deviation error of 0.8 L/min, which is roughly twice that of thermodilution CO. These results suggest that although CO estimation methods may not generate the most precise values, they are still useful for detecting significant (>1 L/min) changes in CO., by James Xin Sun., M.Eng.

Includes bibliographical references (leaves 87-89)., Thesis (M. Eng.)--Harvard-MIT Division of Health Sciences and Technology, 2007., (cont.) The EWS models were then tested in a forward, casual manner on a random cohort of 500 ICU patients to mimic the patients' stay in the unit. The model with the highest performance achieved a sensitivity of 0.85 and a positive predictive value (PPV) of 0.70. Of the 35 episodes of hypotension despite fluid resuscitation present in the random patient dataset, the model provided early warnings for 29 episodes with a mean early warning time of 582 ± 355 minutes., Early goal-directed therapy (EGDT) in severe sepsis and septic shock has shown to provide substantial benefits in patient outcomes. However, these preventive therapeutic interventions are contingent upon an early detection or suspicion of the underlying septic etiology. Detection of sepsis in the early stages can be difficult, as the initial pathogenesis can occur while the patient is still displaying normal vital signs. This study focuses on developing an early warning system (EWS) to provide clinicians with a forewarning of an impending hypotensive crisis-thus allowing for EGDT intervention. Research was completed in three main stages: (1) generating an annotated septic shock dataset, (2) constructing multivariate logistic regression EWS models using the annotated dataset, and (3) testing the EWS models in a forward, causal manner on a random cohort of patients to simulate performance in a real-life ICU setting. The annotated septic shock dataset was created using the Multi-parameter Intelligent Monitoring for Intensive Care II (MIMIC II) database. Automated pre-annotations were generated using search criteria designed to identify two patient types: (1) sepsis patients who do not progress to septic shock, and (2) sepsis patient who progress to septic shock. Currently, manual review by expert clinicians to verify the pre-annotations has not been completed. Six separate EWS models were constructed using the annotated septic shock dataset. The multivariate logistic regression EWS models were trained to differentiate between 107 high-risk sepsis patients of whom 39 experienced a hypotensive crisis and 68 who remained stable. The models were tested using 7-fold cross validation; the mean area under the receiver operating characteristic (ROC) curve for the best model was 0.940 ± 0.038., by Dewang Shavdia., M.Eng.

Includes bibliographical references (leaves 87-89)., Thesis (M. Eng.)--Harvard-MIT Division of Health Sciences and Technology, 2007., (cont.) The EWS models were then tested in a forward, casual manner on a random cohort of 500 ICU patients to mimic the patients' stay in the unit. The model with the highest performance achieved a sensitivity of 0.85 and a positive predictive value (PPV) of 0.70. Of the 35 episodes of hypotension despite fluid resuscitation present in the random patient dataset, the model provided early warnings for 29 episodes with a mean early warning time of 582 ± 355 minutes., Early goal-directed therapy (EGDT) in severe sepsis and septic shock has shown to provide substantial benefits in patient outcomes. However, these preventive therapeutic interventions are contingent upon an early detection or suspicion of the underlying septic etiology. Detection of sepsis in the early stages can be difficult, as the initial pathogenesis can occur while the patient is still displaying normal vital signs. This study focuses on developing an early warning system (EWS) to provide clinicians with a forewarning of an impending hypotensive crisis-thus allowing for EGDT intervention. Research was completed in three main stages: (1) generating an annotated septic shock dataset, (2) constructing multivariate logistic regression EWS models using the annotated dataset, and (3) testing the EWS models in a forward, causal manner on a random cohort of patients to simulate performance in a real-life ICU setting. The annotated septic shock dataset was created using the Multi-parameter Intelligent Monitoring for Intensive Care II (MIMIC II) database. Automated pre-annotations were generated using search criteria designed to identify two patient types: (1) sepsis patients who do not progress to septic shock, and (2) sepsis patient who progress to septic shock. Currently, manual review by expert clinicians to verify the pre-annotations has not been completed. Six separate EWS models were constructed using the annotated septic shock dataset. The multivariate logistic regression EWS models were trained to differentiate between 107 high-risk sepsis patients of whom 39 experienced a hypotensive crisis and 68 who remained stable. The models were tested using 7-fold cross validation; the mean area under the receiver operating characteristic (ROC) curve for the best model was 0.940 ± 0.038., by Dewang Shavdia., M.Eng.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2007., Includes bibliographical references (leaves 210-219)., Intensive Care Unit (ICU) patients are physiologically fragile and require vigilant monitoring and support. The myriad of data gathered from biosensors and clinical information systems has created a challenge for clinicians to assimilate and interpret such large volumes of data. Physiologic measurements in the ICU are inherently noisy, multidimensional, and can readily fluctuate in response to therapeutic interventions as well as evolving pathophysiologic states. ICU patient monitoring systems may potentially improve the efficiency, accuracy and timeliness of clinical decision-making in intensive care. However, the aforementioned characteristics of ICU data can pose a significant signal processing and pattern recognition challenge---often leading to false and clinically irrelevant alarms. We have developed a temporal database of several thousand ICU patient records to facilitate research in advanced monitoring systems. The MIMIC-II database includes high-resolution physiologic waveforms such as ECG, blood pressures waveforms, vital sign trends, laboratory data, fluid balance, therapy profiles, and clinical progress notes over each patient's ICU stay. We quantitatively and qualitatively characterize the MIMIC-II database and include examples of clinical studies that can be supported by its unique attributes. We also introduce a novel algorithm for identifying "similar" temporal patterns that may illuminate hidden information in physiologic time series. The discovery of multi-parameter temporal patterns that are predictive of physiologic instability may aid clinicians in optimizing care. In this thesis, we introduce a novel temporal similarity metric based on a transformation of time series data into an intuitive symbolic representation., (cont.) The symbolic transform is based on a wavelet decomposition to characterize time series dynamics at multiple time scales. The symbolic transformation allows us to utilize classical information retrieval algorithms based on a vector-space model. Our algorithm is capable of assessing the similarity between multi-dimensional time series and is computationally efficient. We utilized our algorithm to identify similar physiologic patterns in hemodynamic time series from ICU patients. The results of this thesis demonstrate that statistical similarities between different patient time series may have meaningful physiologic interpretations in the detection of impending hemodynamic deterioration. Thus, our framework may be of potential use in clinical decision-support systems. As a generalized time series similarity metric, the algorithms that are described have applications in several other domains as well., by Mohammed Saeed., Ph.D.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2006., Includes bibliographical references (p. 73-74)., Cardiac output (CO) is a cardinal parameter of cardiovascular state, and a fundamental determinant of global oxygen delivery. Historically, measurement of CO has been limited to critically-ill patients, using invasive indicator-dilution methods such as thermodilution via Swan-Ganz lines, which carry risks. Over the past century, the premise that CO could be estimated by analysis of the arterial blood pressure (ABP) waveform has captured the attention of many investigators. This approach of estimating CO is minimally invasive, cheap, and can be done continuously as long as ABP waveforms are available. Over a dozen different methods of estimating CO from ABP waveforms have been proposed and some are commercialized. However, the effectiveness of this approach is nebular. Performance validation studies in the past have mostly been conducted on a small set of subjects under well-controlled laboratory conditions. It is entirely possible that there will be circumstances in real world clinical practice in which CO estimation produces inaccurate results. In this thesis, our goals are to (1) build a computational system that estimates CO using 11 of the established methods; (2) evaluate and compare the performance of the CO estimation methods on a large set clinical data, using the simultaneously available thermodilution CO measurements as gold-standard; and (3) design and evaluate an algorithm that identifies and eliminates ABP waveform segments of poor quality. Out of the 11 CO estimation methods studied, there is one method (Liljestrand method) that is clearly more accurate than the rest. Across our study population of 120 subjects, the Liljestrand method has an error distribution with a 1 standard deviation error of 0.8 L/min, which is roughly twice that of thermodilution CO. These results suggest that although CO estimation methods may not generate the most precise values, they are still useful for detecting significant (>1 L/min) changes in CO., by James Xin Sun., M.Eng.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1982., MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING, Bibliography: leaves 94-97., by Daniel Michael Chernoff., M.S.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2007., Includes bibliographical references (leaves 210-219)., Intensive Care Unit (ICU) patients are physiologically fragile and require vigilant monitoring and support. The myriad of data gathered from biosensors and clinical information systems has created a challenge for clinicians to assimilate and interpret such large volumes of data. Physiologic measurements in the ICU are inherently noisy, multidimensional, and can readily fluctuate in response to therapeutic interventions as well as evolving pathophysiologic states. ICU patient monitoring systems may potentially improve the efficiency, accuracy and timeliness of clinical decision-making in intensive care. However, the aforementioned characteristics of ICU data can pose a significant signal processing and pattern recognition challenge---often leading to false and clinically irrelevant alarms. We have developed a temporal database of several thousand ICU patient records to facilitate research in advanced monitoring systems. The MIMIC-II database includes high-resolution physiologic waveforms such as ECG, blood pressures waveforms, vital sign trends, laboratory data, fluid balance, therapy profiles, and clinical progress notes over each patient's ICU stay. We quantitatively and qualitatively characterize the MIMIC-II database and include examples of clinical studies that can be supported by its unique attributes. We also introduce a novel algorithm for identifying "similar" temporal patterns that may illuminate hidden information in physiologic time series. The discovery of multi-parameter temporal patterns that are predictive of physiologic instability may aid clinicians in optimizing care. In this thesis, we introduce a novel temporal similarity metric based on a transformation of time series data into an intuitive symbolic representation., (cont.) The symbolic transform is based on a wavelet decomposition to characterize time series dynamics at multiple time scales. The symbolic transformation allows us to utilize classical information retrieval algorithms based on a vector-space model. Our algorithm is capable of assessing the similarity between multi-dimensional time series and is computationally efficient. We utilized our algorithm to identify similar physiologic patterns in hemodynamic time series from ICU patients. The results of this thesis demonstrate that statistical similarities between different patient time series may have meaningful physiologic interpretations in the detection of impending hemodynamic deterioration. Thus, our framework may be of potential use in clinical decision-support systems. As a generalized time series similarity metric, the algorithms that are described have applications in several other domains as well., by Mohammed Saeed., Ph.D.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2006., "October 14, 2005.", Includes bibliographical references (leaves 63-67)., Focused ultrasound continues to be investigated as an exciting technique for several medical therapy applications including tissue ablation. Phased array transducers serve as important devices for delivering the high intensity focused ultrasound to the targeted tissue. These arrays often produce undesired secondary foci called grating lobes that can result in heating of non-targeted healthy tissue. Grating lobes can be eliminated by setting the element-spacing between adjacent elements to 2 or less. The aim of this research is to construct and evaluate a 2-D planar phased array with close to /2 element-spacing. A 7x7 element array with element-spacing of 0.762mm was constructed using 1,3-piezocomposite material. Acoustic field measurements of the array focused at three different locations were taken to demonstrate its ability to scan the focus in a volume. In addition, a maximum average intensity of 5.54W/cm2 was measured for the unfocused array using radiation force measurements., by Danish Suleiman Khatri., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2006., "October 14, 2005.", Includes bibliographical references (leaves 63-67)., Focused ultrasound continues to be investigated as an exciting technique for several medical therapy applications including tissue ablation. Phased array transducers serve as important devices for delivering the high intensity focused ultrasound to the targeted tissue. These arrays often produce undesired secondary foci called grating lobes that can result in heating of non-targeted healthy tissue. Grating lobes can be eliminated by setting the element-spacing between adjacent elements to 2 or less. The aim of this research is to construct and evaluate a 2-D planar phased array with close to /2 element-spacing. A 7x7 element array with element-spacing of 0.762mm was constructed using 1,3-piezocomposite material. Acoustic field measurements of the array focused at three different locations were taken to demonstrate its ability to scan the focus in a volume. In addition, a maximum average intensity of 5.54W/cm2 was measured for the unfocused array using radiation force measurements., by Danish Suleiman Khatri., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2005., Includes bibliographical references (p. 117-120)., Formatting data and executing time-oriented queries on a massive, multi-parameter clinical information database poses significant computational challenges. The challenges encountered in converting high-resolution waveform and trend signals in the MIMIC II (Multi-parameter Intelligent Monitoring for Intensive Care II) database from an error-prone proprietary format to a stable open-source WFDB (Waveform Database) format is presented in the first half of this thesis. The design and implementation of a search engine that is capable of executing time-series queries on clinical information in the MIMIC II database such as lab results, medications, and nurse-verified values from bedside monitors is presented in the second half of this thesis. The search engine employs simple algorithms with little storage overhead to identify time periods in patient records that satisfy time series criteria based on thresholds and gradients of unevenly-sampled measurements. Results from queries executed on the search engine to detect physiological events of clinical interest were presented. Case studies on patient records returned as hits for queries were performed to review the strengths and limitations of the search engine., by Tin Htet Kyaw., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2006., "October 14, 2005.", Includes bibliographical references (leaves 63-67)., Focused ultrasound continues to be investigated as an exciting technique for several medical therapy applications including tissue ablation. Phased array transducers serve as important devices for delivering the high intensity focused ultrasound to the targeted tissue. These arrays often produce undesired secondary foci called grating lobes that can result in heating of non-targeted healthy tissue. Grating lobes can be eliminated by setting the element-spacing between adjacent elements to 2 or less. The aim of this research is to construct and evaluate a 2-D planar phased array with close to /2 element-spacing. A 7x7 element array with element-spacing of 0.762mm was constructed using 1,3-piezocomposite material. Acoustic field measurements of the array focused at three different locations were taken to demonstrate its ability to scan the focus in a volume. In addition, a maximum average intensity of 5.54W/cm2 was measured for the unfocused array using radiation force measurements., by Danish Suleiman Khatri., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2005., Includes bibliographical references (leaves 67-70)., Medical researchers are legally required to protect patients' privacy by removing personally identifiable information from medical records before sharing the data with other researchers. Different computer-assisted methods are evaluated for removing and replacing protected health information (PHI) from free-text nursing notes collected in the hospital intensive care unit. A semi-automated method was developed to allow clinicians to highlight PHI on the screen of a tablet PC and to compare and combine the selections of different experts reading the same notes. Expert adjudication demonstrated that inter-human variability was high, with few false positives and many false negatives. A preliminary automated de-identification algorithm generated few false negatives but many false positives. A second automated algorithm was developed using the successful portions of the first algorithm and incorporating other heuristic methods to improve overall performance. A large de-identified collection of nursing notes was re-identified with realistic surrogate (but unprotected) dates, serial numbers, names, and phrases to form a "gold standard" reference database of over 2600 notes (approximately 340,000 words) with over 1800 labeled instances of PHI. This gold standard database of nursing notes and the Java source code used to evaluate algorithm performance will be made freely available on the Physionet web site in order to facilitate the development and validation of future de-identification algorithms., by Margaret Douglass., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2005., Includes bibliographical references (leaves 67-70)., Medical researchers are legally required to protect patients' privacy by removing personally identifiable information from medical records before sharing the data with other researchers. Different computer-assisted methods are evaluated for removing and replacing protected health information (PHI) from free-text nursing notes collected in the hospital intensive care unit. A semi-automated method was developed to allow clinicians to highlight PHI on the screen of a tablet PC and to compare and combine the selections of different experts reading the same notes. Expert adjudication demonstrated that inter-human variability was high, with few false positives and many false negatives. A preliminary automated de-identification algorithm generated few false negatives but many false positives. A second automated algorithm was developed using the successful portions of the first algorithm and incorporating other heuristic methods to improve overall performance. A large de-identified collection of nursing notes was re-identified with realistic surrogate (but unprotected) dates, serial numbers, names, and phrases to form a "gold standard" reference database of over 2600 notes (approximately 340,000 words) with over 1800 labeled instances of PHI. This gold standard database of nursing notes and the Java source code used to evaluate algorithm performance will be made freely available on the Physionet web site in order to facilitate the development and validation of future de-identification algorithms., by Margaret Douglass., M.Eng.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2005., Includes bibliographical references (leaves 67-70)., Medical researchers are legally required to protect patients' privacy by removing personally identifiable information from medical records before sharing the data with other researchers. Different computer-assisted methods are evaluated for removing and replacing protected health information (PHI) from free-text nursing notes collected in the hospital intensive care unit. A semi-automated method was developed to allow clinicians to highlight PHI on the screen of a tablet PC and to compare and combine the selections of different experts reading the same notes. Expert adjudication demonstrated that inter-human variability was high, with few false positives and many false negatives. A preliminary automated de-identification algorithm generated few false negatives but many false positives. A second automated algorithm was developed using the successful portions of the first algorithm and incorporating other heuristic methods to improve overall performance. A large de-identified collection of nursing notes was re-identified with realistic surrogate (but unprotected) dates, serial numbers, names, and phrases to form a "gold standard" reference database of over 2600 notes (approximately 340,000 words) with over 1800 labeled instances of PHI. This gold standard database of nursing notes and the Java source code used to evaluate algorithm performance will be made freely available on the Physionet web site in order to facilitate the development and validation of future de-identification algorithms., by Margaret Douglass., M.Eng.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1994., Includes bibliographical references (leaves 165-169)., by David Tak-Wai Hau., M.S.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2005., Includes bibliographical references (leaves 69-71)., Heart rate variability (HRV), the beat-to-beat fluctuation of the heart rate, is a non-invasive test that measures the autonomic regulation of the heart. Assessment of HRV has been shown to predict the risk of mortality in patients after an acute myocardial infarction. Recently, the Krieger lab at MIT developed genetically engineered double knockout (dKO) mice that develop coronary artery disease accompanied by spontaneous myocardial infarctions and die at a very young age. This thesis investigated whether HRV could function as a prognostic indicator in the dKO mouse. A novel method for estimating physiological state of the mouse from the electrocardiogram using an innovative activity index was developed in order to compare HRV variables at different times while controlling for physiologic state. Traditional time and frequency domain variables were used to assess the prognostic power of HRV. Results have shown that none of the HRV variables were helpful in predicting mortality in the dKO mice. Mean heart rate showed some prognostic power, but it was not consistent in all the dKO mice. Finally, the activity index developed in this thesis provided a reliable metric for activity in mice as validated by a camera with motion detection., by Laurence Zapanta., S.M.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1994., Includes bibliographical references (leaves 165-169)., by David Tak-Wai Hau., M.S.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2005., Includes bibliographical references (leaves 67-70)., Medical researchers are legally required to protect patients' privacy by removing personally identifiable information from medical records before sharing the data with other researchers. Different computer-assisted methods are evaluated for removing and replacing protected health information (PHI) from free-text nursing notes collected in the hospital intensive care unit. A semi-automated method was developed to allow clinicians to highlight PHI on the screen of a tablet PC and to compare and combine the selections of different experts reading the same notes. Expert adjudication demonstrated that inter-human variability was high, with few false positives and many false negatives. A preliminary automated de-identification algorithm generated few false negatives but many false positives. A second automated algorithm was developed using the successful portions of the first algorithm and incorporating other heuristic methods to improve overall performance. A large de-identified collection of nursing notes was re-identified with realistic surrogate (but unprotected) dates, serial numbers, names, and phrases to form a "gold standard" reference database of over 2600 notes (approximately 340,000 words) with over 1800 labeled instances of PHI. This gold standard database of nursing notes and the Java source code used to evaluate algorithm performance will be made freely available on the Physionet web site in order to facilitate the development and validation of future de-identification algorithms., by Margaret Douglass., M.Eng.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Aeronautics and Astronautics, 1988., Bibliography: leaves 200-206., by David Alan Israel., Ph.D.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1997., Includes bibliographical references (leaves 115-119)., by Ruilin Zhao., M.S.