Browsing by Author "Haas, Kyle D."

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Automated Quantitative Analysis of Nerve Fiber Conduction Velocity
    (Office of the Vice Chancellor for Research, 2015-04-17) Haas, Kyle D.; Santa Cruz Chavez, Grace; Schild, John
    The baroreflex (BRX) is essential for reliable autonomic control of arterial blood pressure. Central to BRX function is a rapid, negative feedback control of heart rate. Arterial pressure sensors known as baroreceptors (BR) encode heart rate and blood pressure information into patterns of neural discharge that is conveyed to the central nervous system via a network of sensory afferent nerve fibers. These BR fibers are broadly classified as myelinated A-fibers with diameters in the range of 1-10 μm and unmyelinated Cfibers with diameters typically less than 1 μm. Fiber diameter and conduction velocity are related with the large A-fibers being much faster (> 10 m/sec) than the smaller diameter C-fibers (< 1 m/sec). Recently, our lab has documented an additional phenotype of myelinated BR afferents termed Ah-fibers that are notably present in female; but only rarely observed in male rats. In response to an electrical stimulus, the nerve fibers produce a compound action potential (CAP) that propagates away from the stimulation site. The CAP of each fiber type is observable in the evoked waveform on account of the differing conduction velocities. As Ah-fibers have conduction velocities in the range of 10 m/sec - 2 m/sec, the resulting CAP is clearly separated in time from the faster A-fibers and much slower C-fibers. Root-mean-square analysis of these distinct time segments provides a quantitative measure of the total signal energy from each of the A-, Ah-, and C-type fibers. This project sought to create MATLAB scripts that would import nerve recording files from both male and female rats and automate the energy analysis in an efficient and reliable manner. Doing so not only facilitates the analysis of these large data files, but also reduces the possibility for biases and errors that can occur during a manual measurement of nerve activity.
  • Thumbnail Image
    Item
    Transfer learning for medication adherence prediction from social forums self-reported data
    (2018-12) Haas, Kyle D.; Ben-Miled, Zina; King, Brian; El-Sharkawy, Mohamed
    Medication non-adherence and non-compliance left unaddressed can compound into severe medical problems for patients. Identifying patients that are likely to become non-adherent can help reduce these problems. Despite these benefits, monitoring adherence at scale is cost-prohibitive. Social forums offer an easily accessible, affordable, and timely alternative to the traditional methods based on claims data. This study investigates the potential of medication adherence prediction based on social forum data for diabetes and fibromyalgia therapies by using transfer learning from the Medical Expenditure Panel Survey (MEPS). Predictive adherence models are developed by using both survey and social forums data and different random forest (RF) techniques. The first of these implementations uses binned inputs from k-means clustering. The second technique is based on ternary trees instead of the widely used binary decision trees. These techniques are able to handle missing data, a prevalent characteristic of social forums data. The results of this study show that transfer learning between survey models and social forum models is possible. Using MEPS survey data and the techniques listed above to derive RF models, less than 5% difference in accuracy was observed between the MEPS test dataset and the social forum test dataset. Along with these RF techniques, another RF implementation with imputed means for the missing values was developed and shown to predict adherence for social forum patients with an accuracy >70%. This thesis shows that a model trained with verified survey data can be used to complement traditional medical adherence models by predicting adherence from unverified, self-reported data in a dynamic and timely manner. Furthermore, this model provides a method for discovering objective insights from subjective social reports. Additional investigation is needed to improve the prediction accuracy of the proposed model and to assess biases that may be inherent to self-reported adherence measures in social health networks.