Understanding Auditory Evoked Brain Signal via Physics-informed Embedding Network with Multi-Task Transformer

TL;DR

This work tackles decoding auditory signals from task-based fMRI by introducing PEMT-Net, a physics-informed embedding network that combines neural diffusion, a physically informed location encoding, adaptive embedding fusion, and a multi-task Transformer with soft parameter sharing. The method learns rich brain-region embeddings that capture both local and non-local interactions and spatial relationships, leading to superior performance on an eight-category auditory decoding task compared with several baselines and ablations. Key contributions include a diffusion-based neural embedding pipeline, a Fruchterman-Reingold–inspired physical position encoding, multi-round adaptive embedding fusion, and a parameter-sharing Transformer architecture that improves generalization across tasks. The results demonstrate enhanced decoding accuracy and offer insights into how physical constraints and diffusion dynamics can improve brain signal representation, with potential applicability to other neuroimaging decoding problems.

Abstract

In the fields of brain-computer interaction and cognitive neuroscience, effective decoding of auditory signals from task-based functional magnetic resonance imaging (fMRI) is key to understanding how the brain processes complex auditory information. Although existing methods have enhanced decoding capabilities, limitations remain in information utilization and model representation. To overcome these challenges, we propose an innovative multi-task learning model, Physics-informed Embedding Network with Multi-Task Transformer (PEMT-Net), which enhances decoding performance through physics-informed embedding and deep learning techniques. PEMT-Net consists of two principal components: feature augmentation and classification. For feature augmentation, we propose a novel approach by creating neural embedding graphs via node embedding, utilizing random walks to simulate the physical diffusion of neural information. This method captures both local and non-local information overflow and proposes a position encoding based on relative physical coordinates. In the classification segment, we propose adaptive embedding fusion to maximally capture linear and non-linear characteristics. Furthermore, we propose an innovative parameter-sharing mechanism to optimize the retention and learning of extracted features. Experiments on a specific dataset demonstrate PEMT-Net's significant performance in multi-task auditory signal decoding, surpassing existing methods and offering new insights into the brain's mechanisms for processing complex auditory information.

Figures (2)

• Figure 1: Flowchart of our PEMT-Net Model
• Figure 2: A represents the distribution of the original features after dimensionality reduction by the t-SNE method; B represents the distribution of the high-dimensional embeddings encoded with physical locations after dimensionality reduction by the t-SNE method; C represents the error bar graphs of each metric for each method.