Abstract:The Continuous-Time Hidden Markov Model (CT-HMM) is an attractive approach to modeling disease progression due to its ability to describe noisy observations arriving irregularly in time. However, the lack of an efficient parameter learning algorithm for CT-HMM restricts its use to very small models or requires unrealistic constraints on the state transitions. In this paper, we present the first complete characterization of efficient EM-based learning methods for CT-HMM models, as well as the first solution to decoding the optimal state transition sequence and the corresponding state dwelling time. We show that EM-based learning consists of two challenges: the estimation of posterior state probabilities and the computation of end-state conditioned statistics. We solve the first challenge by reformulating the estimation problem as an equivalent discrete time-inhomogeneous hidden Markov model. The second challenge is addressed by adapting three distinct approaches from the continuous time Markov chain (CTMC) literature to the CT-HMM domain. Additionally, we further improve the efficiency of the most efficient method by a factor of the number of states. Then, for decoding, we incorporate a state-of-the-art method from the (CTMC) literature, and extend the end-state conditioned optimal state sequence decoding to the CT-HMM case with the computation of the expected state dwelling time. We demonstrate the use of CT-HMMs with more than 100 states to visualize and predict disease progression using a glaucoma dataset and an Alzheimer's disease dataset, and to decode and visualize the most probable state transition trajectory for individuals on the glaucoma dataset, which helps to identify progressing phenotypes in a comprehensive way. Finally, we apply the CT-HMM modeling and decoding strategy to investigate the progression of language acquisition and development.

Mixture of Input-Output Hidden Markov Models for Heterogeneous Disease Progression Modeling

Learning transition times in event sequences: the Event-Based Hidden Markov Model of disease progression

Time-dependent Probabilistic Generative Models for Disease Progression

A Finite Mixture Hidden Markov Model for Intermittently Observed Disease Process with Heterogeneity and Partially Known Disease Type

A Probabilistic Disease Progression Model for Predicting Future Clinical Outcome

Modeling Disease Progression Trajectories from Longitudinal Observational Data

Disease Progression Modeling and Prediction through Random Effect Gaussian Processes and Time Transformation

Modeling disease progression in longitudinal EHR data using continuous-time hidden Markov models

Personalized Progression Modelling and Prediction in Parkinson’s Disease with a Novel Multi-Modal Graph Approach

Dynamic Prediction for Multiple Repeated Measures and Event Time Data: An Application to Parkinson's Disease

Efficient Learning and Decoding of the Continuous-Time Hidden Markov Model for Disease Progression Modeling

Predicting the Interaction Between Treatment Processes and Disease Progression by Using Hidden Markov Model

DPVis: Visual Analytics With Hidden Markov Models for Disease Progression Pathways

Bayesian mixed-effect higher-order hidden Markov models with applications to predictive healthcare using electronic health records

Hidden Markov models as recurrent neural networks: an application to Alzheimer's disease

Disease Progression Modelling and Stratification for detecting sub-trajectories in the natural history of pathologies: application to Parkinson's Disease trajectory modelling

Modeling Disease Progression via Multisource Multitask Learners: A Case Study With Alzheimer's Disease.

Toward a multimodal multitask model for neurodegenerative diseases diagnosis and progression prediction

Mixture of Coupled HMMs for Robust Modeling of Multivariate Healthcare Time Series

Modelling the Natural History of Huntington'S Disease Progression

Hierarchical Bayesian inference to model continuous phenotypical progression in Alzheimer's Disease