Predicting Outcome in Patients with Brain Injury: Differences between Machine Learning versus Conventional Statistics

Artificial neural network analysis of factors affecting functional independence recovery in patients with lumbar stenosis after neurosurgery treatment: An observational cohort study

Background and aim

Lumbar spinal stenosis (LSS) is a leading cause of low back pain and lower limbs pain often associated with functional impairment which entails the loss or the impairment of independence in older adults. Conservative treatment is effective in a small percentage of patients, while a significant percentage undergo surgery, even if often without a complete resolution of clinical symptoms and motor deficits. The aim of the study is to identify clinical and demographic prognostic factors characterising the patients who would benefit most from surgical treatment in relation to the functional independence recovery using an innovative approach based on an artificial neural network.

Methods

Adult patients with LSS and indication of neurosurgical treatment were enrolled in the study. Clinical evaluation was performed in the preoperative-phase (into the 48 h before surgery) and after two months. Clinical battery investigated the motor, functional, cognitive, behavioural, and pain status. Demographics and clinical characteristics were analysed via Artificial Neural Network (ANN) using 24 input variables, 2 hidden layers and a single final output layer to predict the outcome. ANN results were compared with those of a multiple linear regression.

Results

108 patients were included in the study and 90 of them [66.5 ± 12.8 years; 27.8 % F] were submitted to surgery treatment and completed longitudinal evaluation. Statistically significant improvement was recorded in all clinical scales comparing pre- and post-surgery. The ANN results showed a prediction ability up to 81 %. Disability, functional limitations, and pain concerning clinical assessment and stature, onset and age about demographic characteristics are the main variables impacting on surgical outcome.

Conclusions

ANN can support clinical decision making, using clinical and demographic characteristics of patients with LSS identifying the characteristics of those who might benefit more from the surgical treatment in terms of global functional recovery.

Machine learning prediction models for in-hospital postoperative functional outcome after moderate-to-severe traumatic brain injury

AIM

This study aims to utilize machine learning (ML) and logistic regression (LR) models to predict surgical outcomes among patients with traumatic brain injury (TBI) based on admission examination, assisting in making optimal surgical treatment decision for these patients.

METHOD

We conducted a retrospective review of patients hospitalized in our department for moderate-to-severe TBI. Patients admitted between October 2011 and October 2022 were assigned to the training set, while patients admitted between November 2022 and May 2023 were designated as the external validation set. Five ML algorithms and LR model were employed to predict the postoperative Glasgow Outcome Scale (GOS) status at discharge using clinical and routine blood data collected upon admission. The Shapley (SHAP) plot was utilized for interpreting the models.

RESULTS

A total of 416 patients were included in this study, and they were divided into the training set (n = 396) and the external validation set (n = 47). The ML models, using both clinical and routine blood data, were able to predict postoperative GOS outcomes with area under the curve (AUC) values ranging from 0.860 to 0.900 during the internal cross-validation and from 0.801 to 0.890 during the external validation. In contrast, the LR model had the lowest AUC values during the internal and external validation (0.844 and 0.567, respectively). When blood data was not available, the ML models achieved AUCs of 0.849 to 0.870 during the internal cross-validation and 0.714 to 0.861 during the external validation. Similarly, the LR model had the lowest AUC values (0.821 and 0.638, respectively). Through repeated cross-validation analysis, we found that routine blood data had a significant association with higher mean AUC values in all ML and LR models. The SHAP plot was used to visualize the contributions of all predictors and highlighted the significance of blood data in the lightGBM model.

CONCLUSION

The study concluded that ML models could provide rapid and accurate predictions for postoperative GOS outcomes at discharge following moderate-to-severe TBI. The study also highlighted the crucial role of routine blood tests in improving such predictions, and may contribute to the optimization of surgical treatment decision-making for patients with TBI.

Machine learning models for predicting in-hospital outcomes after non-surgical treatment among patients with moderate-to-severe traumatic brain injury

AIM

This study aims to develop prediction models for in-hospital outcomes after non-surgical treatment among patients with moderate-to-severe traumatic brain injury (TBI).

METHOD

We conducted a retrospective review of patients hospitalized for moderate-to-severe TBI in our department from 2011 to 2020. Five machine learning (ML) algorithms and the conventional logistic regression (LR) model were employed to predict in-hospital mortality and the Glasgow Outcome Scale (GOS) functional outcomes. These models utilized clinical and routine blood data collected upon admission.

RESULTS

This study included a total of 196 patients who received only non-surgical treatment after moderate-to-severe TBI. When predicting mortality, ML models achieved area under the curve (AUC) values of 0.921 to 0.994 using clinical and routine blood data, and 0.877 to 0.982 using only clinical data. In comparison, LR models yielded AUCs of 0.762 and 0.730 respectively. When predicting the GOS outcome, ML models achieved AUCs of 0.870 to 0.915 using clinical and routine blood data, and 0.858 to 0.927 using only clinical data. In comparison, the LR model yielded AUCs of 0.798 and 0.787 respectively. Repeated internal validation showed that the contributions of routine blood data for prediction models may depend on different prediction algorithms and different outcome measurements.

CONCLUSION

The study reported ML-based prediction models that provided rapid and accurate predictions on short-term outcomes after non-surgical treatment among patients with moderate-to-severe TBI. The study also highlighted the superiority of ML models over conventional LR models and proposed the complex contributions of routine blood data in such predictions.

Integrating EEG and Machine Learning to Analyze Brain Changes during the Rehabilitation of Broca's Aphasia

The fusion of electroencephalography (EEG) with machine learning is transforming rehabilitation. Our study introduces a neural network model proficient in distinguishing pre- and post-rehabilitation states in patients with Broca's aphasia, based on brain connectivity metrics derived from EEG recordings during verbal and spatial working memory tasks. The Granger causality (GC), phase-locking value (PLV), weighted phase-lag index (wPLI), mutual information (MI), and complex Pearson correlation coefficient (CPCC) across the delta, theta, and low- and high-gamma bands were used (excluding GC, which spanned the entire frequency spectrum). Across eight participants, employing leave-one-out validation for each, we evaluated the intersubject prediction accuracy across all connectivity methods and frequency bands. GC, MI theta, and PLV low-gamma emerged as the top performers, achieving 89.4%, 85.8%, and 82.7% accuracy in classifying verbal working memory task data. Intriguingly, measures designed to eliminate volume conduction exhibited the poorest performance in predicting rehabilitation-induced brain changes. This observation, coupled with variations in model performance across frequency bands, implies that different connectivity measures capture distinct brain processes involved in rehabilitation. The results of this paper contribute to current knowledge by presenting a clear strategy of utilizing limited data to achieve valid and meaningful results of machine learning on post-stroke rehabilitation EEG data, and they show that the differences in classification accuracy likely reflect distinct brain processes underlying rehabilitation after stroke.

Machine Learning-Based Asthma Attack Prediction Models From Routinely Collected Electronic Health Records: Systematic Scoping Review

BACKGROUND

An early warning tool to predict attacks could enhance asthma management and reduce the likelihood of serious consequences. Electronic health records (EHRs) providing access to historical data about patients with asthma coupled with machine learning (ML) provide an opportunity to develop such a tool. Several studies have developed ML-based tools to predict asthma attacks.

OBJECTIVE

This study aims to critically evaluate ML-based models derived using EHRs for the prediction of asthma attacks.

METHODS

We systematically searched PubMed and Scopus (the search period was between January 1, 2012, and January 31, 2023) for papers meeting the following inclusion criteria: (1) used EHR data as the main data source, (2) used asthma attack as the outcome, and (3) compared ML-based prediction models' performance. We excluded non-English papers and nonresearch papers, such as commentary and systematic review papers. In addition, we also excluded papers that did not provide any details about the respective ML approach and its result, including protocol papers. The selected studies were then summarized across multiple dimensions including data preprocessing methods, ML algorithms, model validation, model explainability, and model implementation.

RESULTS

Overall, 17 papers were included at the end of the selection process. There was considerable heterogeneity in how asthma attacks were defined. Of the 17 studies, 8 (47%) studies used routinely collected data both from primary care and secondary care practices together. Extreme imbalanced data was a notable issue in most studies (13/17, 76%), but only 38% (5/13) of them explicitly dealt with it in their data preprocessing pipeline. The gradient boosting-based method was the best ML method in 59% (10/17) of the studies. Of the 17 studies, 14 (82%) studies used a model explanation method to identify the most important predictors. None of the studies followed the standard reporting guidelines, and none were prospectively validated.

CONCLUSIONS

Our review indicates that this research field is still underdeveloped, given the limited body of evidence, heterogeneity of methods, lack of external validation, and suboptimally reported models. We highlighted several technical challenges (class imbalance, external validation, model explanation, and adherence to reporting guidelines to aid reproducibility) that need to be addressed to make progress toward clinical adoption.

Predicting return to work after traumatic brain injury using machine learning and administrative data

BACKGROUND

Accurate patient-specific predictions on return-to-work after traumatic brain injury (TBI) can support both clinical practice and policymaking. The use of machine learning on large administrative data provides interesting opportunities to create such prognostic models.

AIM

The current study assesses whether return-to-work one year after TBI can be predicted accurately from administrative data. Additionally, this study explores how model performance and feature importance change depending on whether a distinction is made between mild and moderate-to-severe TBI.

METHODS

This study used a population-based dataset that combined discharge, claims and social security data of patients hospitalized with a TBI in Belgium during the year 2016. The prediction of TBI was attempted with three algorithms, elastic net logistic regression, random forest and gradient boosting and compared in their performance by their accuracy, sensitivity, specificity and area under the receiver operator curve (ROC AUC).

RESULTS

The distinct modelling algorithms resulted in similar results, with 83% accuracy (ROC AUC 85%) for a binary classification of employed vs. not employed and up to 76% (ROC AUC 82%) for a multiclass operationalization of employment outcome. Modelling mild and moderate-to-severe TBI separately did not result in considerable differences in model performance and feature importance. The features of main importance for return-to-work prediction were related to pre-injury employment.

DISCUSSION

While clearly offering some information beneficial for predicting return-to-work, administrative data needs to be supplemented with additional information to allow further improvement of patient-specific prognose.

State-of-the-Art on Brain-Computer Interface Technology

This paper provides a comprehensive overview of the state-of-the-art in brain-computer interfaces (BCI). It begins by providing an introduction to BCIs, describing their main operation principles and most widely used platforms. The paper then examines the various components of a BCI system, such as hardware, software, and signal processing algorithms. Finally, it looks at current trends in research related to BCI use for medical, educational, and other purposes, as well as potential future applications of this technology. The paper concludes by highlighting some key challenges that still need to be addressed before widespread adoption can occur. By presenting an up-to-date assessment of the state-of-the-art in BCI technology, this paper will provide valuable insight into where this field is heading in terms of progress and innovation.

Modeling outcome trajectories in patients with acquired brain injury using a non-linear dynamic evolution approach

This study describes a dynamic non-linear mathematical approach for modeling the course of disease in acquired brain injury (ABI) patients. Data from a multicentric study were used to evaluate the reliability of the Michaelis-Menten (MM) model applied to well-known clinical variables that assess the outcome of ABI patients. The sample consisted of 156 ABI patients admitted to eight neurorehabilitation subacute units and evaluated at baseline (T0), 4 months after the event (T1) and at discharge (T2). The MM model was used to characterize the trend of the first Principal Component Analysis (PCA) dimension (represented by the variables: feeding modality, RLAS, ERBI-A, Tracheostomy, CRS-r and ERBI-B) in order to predict the most plausible outcome, in terms of positive or negative Glasgow outcome score (GOS) at discharge. Exploring the evolution of the PCA dimension 1 over time, after day 86 the MM model better differentiated between the time course for individuals with a positive and negative GOS (accuracy: 85%; sensitivity: 90.6%; specificity: 62.5%). The non-linear dynamic mathematical model can be used to provide more comprehensive trajectories of the clinical evolution of ABI patients during the rehabilitation period. Our model can be used to address patients for interventions designed for a specific outcome trajectory.

Application of an Artificial Neural Network to Identify the Factors Influencing Neurorehabilitation Outcomes of Patients with Ischemic Stroke Treated with Thrombolysis

The administration of thrombolysis usually reduces the risk of death and the consequences of stroke in the acute phase. However, having received thrombolysis administration is not a prognostic factor for neurorehabilitation outcome in the subacute phase of stroke. It is conceivably due to the complex intertwining of many clinical factors. An artificial neural network (ANN) analysis could be helpful in identifying the prognostic factors of neurorehabilitation outcomes and assigning a weight to each of the factors considered. This study hypothesizes that the prognostic factors could be different between patients who received and those who did not receive thrombolytic treatment, even if thrombolysis is not a prognostic factor per se. In a sample of 862 patients with ischemic stroke, the tested ANN identified some common factors (such as disability at admission, age, unilateral spatial neglect), some factors with higher weight in patients who received thrombolysis (hypertension, epilepsy, aphasia, obesity), and some other factors with higher weight in the other patients (dysphagia, malnutrition, total arterial circulatory infarction). Despite the fact that thrombolysis is not an independent prognostic factor for neurorehabilitation, it seems to modify the relative importance of other clinical factors in predicting which patients will better respond to neurorehabilitation.