About this Research Topic
Artificial Intelligence (AI) and Machine Learning (ML) have demonstrated remarkable success in automating the diagnosis and prognosis of neurological and neurodegenerative diseases. Their application has yielded excellent performance, providing not only interpretability and explainability but also advancing our understanding of disease progression. However, typical ML algorithms used in these contexts, such as supervised learning approaches, lack the capacity to consider the temporal aspect of events, survival probability, or predict risk scores. Additionally, they struggle to handle censoring, a common occurrence when events of interest are not observed for some subjects before the study concludes.
Survival analysis, a statistical field developed to predict the time of an event while incorporating censored data, plays a pivotal role in addressing these limitations. Although widely applied, traditional methods like Cox Proportional Hazard (CPH) were designed for small datasets and face challenges with scalability to high dimensions. Recently, ML algorithms have been adapted to handle censoring efficiently and predict event times, showing optimal performance even on high-dimensional and heterogeneous data. Examples include decision trees-based approaches like Random Survival Forests, Bayesian methods, and Deep Learning models like DeepSurv.
ML for Survival analysis has shown promise in various applications within the neuroscience domain. Examples include predicting the conversion risk from Mild Cognitive Impairment (MCI) to Alzheimer’s disease (AD), estimating the time of respiratory failure for Amyotrophic Lateral Sclerosis (ALS) patients, and predicting therapy switch/failure in Multiple Sclerosis (MS). Despite these successes, the application of ML-based survival methods in neurological and neurodegenerative diseases remains limited, offering a rich yet unexplored field of research.
This Research Topic aims to collect novel or state-of-the-art works focusing on the application of AI and ML methods specifically designed for Survival Analysis. We invite contributions related to predicting the time of clinical events, assessing survival probabilities, or predicting risk scores in Neurological and Neurodegenerative diseases. Meta-analyses or systematic reviews of the literature are also welcome.
Topics of interest include but are not limited to:
- Comparison and evaluation of different ML approaches in terms of performance, scalability, and interpretability.
- Strategies for integrating diverse data types (imaging, clinical, genetic, etc.) to enhance the predictive power of survival models.
- Advanced methods for addressing censoring in survival analysis using ML, including but not limited to handling missing data and imputation strategies.
- Strategies for handling imbalanced datasets in survival analysis, considering the often uneven distribution of event and non-event instances.
- Identification of early biomarkers and critical time points in the progression of neurological diseases.
- Exploration of how predictive models can inform individualized therapeutic interventions based on patient-specific data.
Survival analysis, a statistical field developed to predict the time of an event while incorporating censored data, plays a pivotal role in addressing these limitations. Although widely applied, traditional methods like Cox Proportional Hazard (CPH) were designed for small datasets and face challenges with scalability to high dimensions. Recently, ML algorithms have been adapted to handle censoring efficiently and predict event times, showing optimal performance even on high-dimensional and heterogeneous data. Examples include decision trees-based approaches like Random Survival Forests, Bayesian methods, and Deep Learning models like DeepSurv.
ML for Survival analysis has shown promise in various applications within the neuroscience domain. Examples include predicting the conversion risk from Mild Cognitive Impairment (MCI) to Alzheimer’s disease (AD), estimating the time of respiratory failure for Amyotrophic Lateral Sclerosis (ALS) patients, and predicting therapy switch/failure in Multiple Sclerosis (MS). Despite these successes, the application of ML-based survival methods in neurological and neurodegenerative diseases remains limited, offering a rich yet unexplored field of research.
This Research Topic aims to collect novel or state-of-the-art works focusing on the application of AI and ML methods specifically designed for Survival Analysis. We invite contributions related to predicting the time of clinical events, assessing survival probabilities, or predicting risk scores in Neurological and Neurodegenerative diseases. Meta-analyses or systematic reviews of the literature are also welcome.
Topics of interest include but are not limited to:
- Comparison and evaluation of different ML approaches in terms of performance, scalability, and interpretability.
- Strategies for integrating diverse data types (imaging, clinical, genetic, etc.) to enhance the predictive power of survival models.
- Advanced methods for addressing censoring in survival analysis using ML, including but not limited to handling missing data and imputation strategies.
- Strategies for handling imbalanced datasets in survival analysis, considering the often uneven distribution of event and non-event instances.
- Identification of early biomarkers and critical time points in the progression of neurological diseases.
- Exploration of how predictive models can inform individualized therapeutic interventions based on patient-specific data.
Keywords: Artificial Intelligence, Machine Learning, Neurological survival analysis, Neurodegenerative diseases, survival analysis
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
