Benchmarking explanation methods for mental state decoding with deep learning models

Comparative evaluation of interpretation methods in surface-based age prediction for neonates

Significant changes in brain morphology occur during the third trimester of gestation. The capability of deep learning in leveraging these morphological features has enhanced the accuracy of brain age predictions for this critical period. Yet, the opaque nature of deep learning techniques, often described as "black box" approaches, limits their interpretability, posing challenges in clinical applications. Traditional interpretable methods developed for computer vision and natural language processing may not directly translate to the distinct demands of neuroimaging. In response, our research evaluates the effectiveness and adaptability of two interpretative methods-regional age prediction and the perturbation-based saliency map approach-for predicting the brain age of neonates. Analyzing 664 T1 MRI scans with the NEOCIVET pipeline to extract brain surface and cortical features, we assess how these methods illuminate key brain regions for age prediction, focusing on technical analysis with clinical insight. Through a comparative analysis of the saliency index (SI) with relative brain age (RBA) and the examination of structural covariance networks, we uncover the saliency index's enhanced ability to pinpoint regions vital for accurate indication of clinical factors. Our results highlight the advantages of perturbation techniques in addressing the complexities of medical data, steering clinical interventions for premature neonates towards more personalized and interpretable approaches. This study not only reveals the promise of these methods in complex medical scenarios but also offers a blueprint for implementing more interpretable and clinically relevant deep learning models in healthcare settings.

An Adaptively Weighted Averaging Method for Regional Time Series Extraction of fMRI-Based Brain Decoding

Brain decoding that classifies cognitive states using the functional fluctuations of the brain can provide insightful information for understanding the brain mechanisms of cognitive functions. Among the common procedures of decoding the brain cognitive states with functional magnetic resonance imaging (fMRI), extracting the time series of each brain region after brain parcellation traditionally averages across the voxels within a brain region. This neglects the spatial information among the voxels and the requirement of extracting information for the downstream tasks. In this study, we propose to use a fully connected neural network that is jointly trained with the brain decoder to perform an adaptively weighted average across the voxels within each brain region. We perform extensive evaluations by cognitive state decoding, manifold learning, and interpretability analysis on the Human Connectome Project (HCP) dataset. The performance comparison of the cognitive state decoding presents an accuracy increase of up to 5% and stable accuracy improvement under different time window sizes, resampling sizes, and training data sizes. The results of manifold learning show that our method presents a considerable separability among cognitive states and basically excludes subject-specific information. The interpretability analysis shows that our method can identify reasonable brain regions corresponding to each cognitive state. Our study would aid the improvement of the basic pipeline of fMRI processing.

Prediction of brain sex from EEG: using large-scale heterogeneous dataset for developing a highly accurate and interpretable ML model

This study presents a comprehensive examination of sex-related differences in resting-state electroencephalogram (EEG) data, leveraging two different types of machine learning models to predict an individual's sex. We utilized data from the Two Decades-Brainclinics Research Archive for Insights in Neurophysiology (TDBRAIN) EEG study, affirming that gender prediction can be attained with noteworthy accuracy. The best performing model achieved an accuracy of 85% and an ROC AUC of 89%, surpassing all prior benchmarks set using EEG data and rivaling the top-tier results derived from fMRI studies. A comparative analysis of LightGBM and Deep Convolutional Neural Network (DCNN) models revealed DCNN's superior performance, attributed to its ability to learn complex spatial-temporal patterns in the EEG data and handle large volumes of data effectively. Despite this, interpretability remained a challenge for the DCNN model. The LightGBM interpretability analysis revealed that the most important EEG features for accurate sex prediction were related to left fronto-central and parietal EEG connectivity. We also showed the role of both low (delta and theta) and high (beta and gamma) activity in the accurate sex prediction. These results, however, have to be approached with caution, because it was obtained from a dataset comprised largely of participants with various mental health conditions, which limits the generalizability of the results and necessitates further validation in future studies. . Overall, the study illuminates the potential of interpretable machine learning for sex prediction, alongside highlighting the importance of considering individual differences in prediction sex from brain activity.