Novel Semi-Supervised Learning Models to Balance Data Inclusivity and Usability in Healthcare Applications

157564-Thumbnail Image.png
Description
Semi-supervised learning (SSL) is sub-field of statistical machine learning that is useful for problems that involve having only a few labeled instances with predictor (X) and target (Y) information, and abundance of unlabeled instances that only have predictor (X) information.

Semi-supervised learning (SSL) is sub-field of statistical machine learning that is useful for problems that involve having only a few labeled instances with predictor (X) and target (Y) information, and abundance of unlabeled instances that only have predictor (X) information. SSL harnesses the target information available in the limited labeled data, as well as the information in the abundant unlabeled data to build strong predictive models. However, not all the included information is useful. For example, some features may correspond to noise and including them will hurt the predictive model performance. Additionally, some instances may not be as relevant to model building and their inclusion will increase training time and potentially hurt the model performance. The objective of this research is to develop novel SSL models to balance data inclusivity and usability. My dissertation research focuses on applications of SSL in healthcare, driven by problems in brain cancer radiomics, migraine imaging, and Parkinson’s Disease telemonitoring.

The first topic introduces an integration of machine learning (ML) and a mechanistic model (PI) to develop an SSL model applied to predicting cell density of glioblastoma brain cancer using multi-parametric medical images. The proposed ML-PI hybrid model integrates imaging information from unbiopsied regions of the brain as well as underlying biological knowledge from the mechanistic model to predict spatial tumor density in the brain.

The second topic develops a multi-modality imaging-based diagnostic decision support system (MMI-DDS). MMI-DDS consists of modality-wise principal components analysis to incorporate imaging features at different aggregation levels (e.g., voxel-wise, connectivity-based, etc.), a constrained particle swarm optimization (cPSO) feature selection algorithm, and a clinical utility engine that utilizes inverse operators on chosen principal components for white-box classification models.

The final topic develops a new SSL regression model with integrated feature and instance selection called s2SSL (with “s2” referring to selection in two different ways: feature and instance). s2SSL integrates cPSO feature selection and graph-based instance selection to simultaneously choose the optimal features and instances and build accurate models for continuous prediction. s2SSL was applied to smartphone-based telemonitoring of Parkinson’s Disease patients.
Date Created
2019
Agent

Multi-Parametric MRI and Texture Analysis to Visualize Spatial Histologic Heterogeneity and Tumor Extent in Glioblastoma

128818-Thumbnail Image.png
Description

Background: Genetic profiling represents the future of neuro-oncology but suffers from inadequate biopsies in heterogeneous tumors like Glioblastoma (GBM). Contrast-enhanced MRI (CE-MRI) targets enhancing core (ENH) but yields adequate tumor in only ~60% of cases. Further, CE-MRI poorly localizes infiltrative tumor

Background: Genetic profiling represents the future of neuro-oncology but suffers from inadequate biopsies in heterogeneous tumors like Glioblastoma (GBM). Contrast-enhanced MRI (CE-MRI) targets enhancing core (ENH) but yields adequate tumor in only ~60% of cases. Further, CE-MRI poorly localizes infiltrative tumor within surrounding non-enhancing parenchyma, or brain-around-tumor (BAT), despite the importance of characterizing this tumor segment, which universally recurs. In this study, we use multiple texture analysis and machine learning (ML) algorithms to analyze multi-parametric MRI, and produce new images indicating tumor-rich targets in GBM.

Methods: We recruited primary GBM patients undergoing image-guided biopsies and acquired pre-operative MRI: CE-MRI, Dynamic-Susceptibility-weighted-Contrast-enhanced-MRI, and Diffusion Tensor Imaging. Following image coregistration and region of interest placement at biopsy locations, we compared MRI metrics and regional texture with histologic diagnoses of high- vs low-tumor content (≥80% vs <80% tumor nuclei) for corresponding samples. In a training set, we used three texture analysis algorithms and three ML methods to identify MRI-texture features that optimized model accuracy to distinguish tumor content. We confirmed model accuracy in a separate validation set.

Results: We collected 82 biopsies from 18 GBMs throughout ENH and BAT. The MRI-based model achieved 85% cross-validated accuracy to diagnose high- vs low-tumor in the training set (60 biopsies, 11 patients). The model achieved 81.8% accuracy in the validation set (22 biopsies, 7 patients).

Conclusion: Multi-parametric MRI and texture analysis can help characterize and visualize GBM’s spatial histologic heterogeneity to identify regional tumor-rich biopsy targets.

Date Created
2015-11-24
Agent

The role of tactile information in transfer of learned manipulation following changes in degrees of freedom

152548-Thumbnail Image.png
Description
Humans are capable of transferring learning for anticipatory control of dexterous object manipulation despite changes in degrees-of-freedom (DoF), i.e., switching from lifting an object with two fingers to lifting the same object with three fingers. However, the role that tactile

Humans are capable of transferring learning for anticipatory control of dexterous object manipulation despite changes in degrees-of-freedom (DoF), i.e., switching from lifting an object with two fingers to lifting the same object with three fingers. However, the role that tactile information plays in this transfer of learning is unknown. In this study, subjects lifted an L-shaped object with two fingers (2-DoF), and then lifted the object with three fingers (3-DoF). The subjects were divided into two groups--one group performed the task wearing a glove (to reduce tactile sensibility) upon the switch to 3-DoF (glove group), while the other group did not wear the glove (control group). Compensatory moment (torque) was used as a measure to determine how well the subject could minimize the tilt of the object following the switch from 2-DoF to 3-DoF. Upon the switch to 3-DoF, subjects wearing the glove generated a compensatory moment (Mcom) that had a significantly higher error than the average of the last five trials at the end of the 3-DoF block (p = 0.012), while the control subjects did not demonstrate a significant difference in Mcom. Additional effects of the reduction in tactile sensibility were: (1) the grip force for the group of subjects wearing the glove was significantly higher in the 3-DoF trials compared to the 2-DoF trials (p = 0.014), while the grip force of the control subjects was not significantly different; (2) the difference in centers of pressure between the thumb and fingers (ΔCoP) significantly increased in the 3-DoF block for the group of subjects wearing the glove, while the ΔCoP of the control subjects was not significantly different; (3) lastly, the control subjects demonstrated a greater increase in lift force than the group of subjects wearing the glove (though results were not significant). Combined together, these results suggest different force modulation strategies are used depending on the amount of tactile feedback that is available to the subject. Therefore, reduction of tactile sensibility has important effects on subjects' ability to transfer learned manipulation across different DoF contexts.
Date Created
2014
Agent