Deep learning applications in fMRI - a ReviewWork

In modern neuroscience and clinical research, functional magnetic resonance imaging (fMRI) is a non-invasive imaging technique that uses magnetic resonance imaging to measure hemodynamic changes caused by neuronal activity. This technique is used to study human brain function and cognition in healthy individuals and groups with abnormal brain states. And it is one of the most commonly used imaging modalities. Because of its characteristics of containing temporal information, it is widely used in research in cognitive neuroscience, clinical psychiatry/psychology, and preoperative planning. Advances in artificial intelligence, especially the advent of deep learning techniques have shown promising results for better interpretation of fMRI data. This paper focuses on fMRI data and summarizes the current state of application of deep learning methods and models on resting-state and Task-evoked data. In the future, deep learning combined with advanced feature selection methods or task-state fMRI data has the potential to become a powerful tool for exploring the state and function of the human brain.