Channel Contribution In Deep Learning Based Automatic Sleep Scoring - How Many Channels Do We Need?

Machine learning based sleep scoring methods aim to automate the process of annotating polysomnograms with sleep stages. Although sleep signals of multiple modalities and channels should contain more information according to sleep guidelines, most multi-channel multi-modal models in the literature showed only a little performance improvement compared to single-channel EEG models and sometimes even failed to outperform them. In this paper, we investigate whether the high performance of single-channel EEG models can be attributed to specific model features in their deep learning architectures and to which extent multi-channel multi-modal models take the information from different channels of modalities into account. First, we transfer the model features from single-channel EEG models, such as combinations of small and large filters in CNNs, to multi-channel multi-modal models and measure their impacts. Second, we employ two explainability methods, the layer-wise relevance propagation as post-hoc and the embedded channel attention network as intrinsic interpretability methods, to measure the contribution of different channels on predictive performance. We find that i) single-channel model features can improve the performance of multi-channel multi-modal models and ii) multi-channel multi-modal models focus on one important channel per modality and use the remaining channels to complement the information of the focused channels. Our results suggest that more advanced methods for aggregating channel information using complementary information from other channels may improve sleep scoring performance for multi-channel multi-modal models.

STQS: Interpretable multi-modal Spatial-Temporal-seQuential model for automatic Sleep scoring.

Sleep scoring is an important step for the detection of sleep disorders and usually performed by visual analysis. Since manual sleep scoring is time consuming, machine-learning based approaches have been proposed. Though efficient, these algorithms are black-box in nature and difficult to interpret by clinicians. In this paper, we propose a deep learning architecture for multi-modal sleep scoring, investigate the model's decision making process, and compare the model's reasoning with the annotation guidelines in the AASM manual. Our architecture, called STQS, uses convolutional neural networks (CNN) to automatically extract spatio-temporal features from 3 modalities (EEG, EOG and EMG), a bidirectional long short-term memory (Bi-LSTM) to extract sequential information, and residual connections to combine spatio-temporal and sequential features. We evaluated our model on two large datasets, obtaining an accuracy of 85% and 77% and a macro F1 score of 79% and 73% on SHHS and an in-house dataset, respectively. We further quantify the contribution of various architectural components and conclude that adding LSTM layers improves performance over a spatio-temporal CNN, while adding residual connections does not. Our interpretability results show that the output of the model is well aligned with AASM guidelines, and therefore, the model's decisions correspond to domain knowledge. We also compare multi-modal models and single-channel models and suggest that future research should focus on improving multi-modal models.

Automatic Process Comparison for Subpopulations: Application in Cancer Care.

Processes in organisations, such as hospitals, may deviate from the intended standard processes, due to unforeseeable events and the complexity of the organisation. For hospitals, the knowledge of actual patient streams for patient populations (e.g., severe or non-severe cases) is important for quality control and improvement. Process discovery from event data in electronic health records can shed light on the patient flows, but their comparison for different populations is cumbersome and time-consuming. In this paper, we present an approach for the automatic comparison of process models that were extracted from events in electronic health records. Concretely, we propose comparing processes for different patient populations by cross-log conformance checking, and standard graph similarity measures obtained from the directed graph underlying the process model. We perform a user study with 20 participants in order to obtain a ground truth for similarity of process models. We evaluate our approach on two data sets, the publicly available MIMIC database with the focus on different cancer patients in intensive care, and a database on breast cancer patients from a Dutch hospital. In our experiments, we found average fitness to be a good indicator for visual similarity in the ZGT use case, while the average precision and graph edit distance are strongly correlated with visual impression for cancer process models on MIMIC. These results are a call for further research and evaluation for determining which similarity or combination of similarities is needed in which type of process model comparison.

Post-Structuring Radiology Reports of Breast Cancer Patients for Clinical Quality Assurance.

Hospitals often set protocols based on well defined standards to maintain the quality of patient reports. To ensure that the clinicians conform to the protocols, quality assurance of these reports is needed. Patient reports are currently written in free-text format, which complicates the task of quality assurance. In this paper, we present a machine learning based natural language processing system for automatic quality assurance of radiology reports on breast cancer. This is achieved in three steps: we i) identify the top-level structure (headings) of the report, ii) classify the report content into the top-level headings, and iii) convert the free-text detailed findings in the report to a semi-structured format (post-structuring). Top level structure and content of report were predicted with an F1 score of 0.97 and 0.94, respectively, using Support Vector Machine (SVM) classifiers. For automatic structuring, our proposed hierarchical Conditional Random Field (CRF) outperformed the baseline CRF with an F1 score of 0.78 versus 0.71. The determined structure of the report is represented in semi-structured XML format of the free-text report, which helps to easily visualize the conformance of the findings to the protocols. This format also allows easy extraction of specific information for other purposes such as search, evaluation, and research.