AUTHOR=Ardalan Zaniar , Subbian Vignesh 

TITLE=Transfer Learning Approaches for Neuroimaging Analysis: A Scoping Review

JOURNAL=Frontiers in Artificial Intelligence

VOLUME=Volume 5 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2022.780405

DOI=10.3389/frai.2022.780405

ISSN=2624-8212

ABSTRACT=Deep learning algorithms have been moderately successful in diagnoses of diseases by analyzing medical images especially neuroimaging. However, in the neuroimaging area, there are a number of annotated data due to issues including privacy regulations, and high cost of data acquisition and distribution. Transfer learning methods demonstrated astonishing performance in tackling annotated data shortage. It utilizes and transfers knowledge learned from a source domain to target domain even with small numbers of data. There are multiple approaches in transfer learning that would result in different performance in diagnosis, detection, classification of medical issues. Therefore, in this paper, we reviewed transfer learning approaches and their different behaviors in neuroimaging problems. We reviewed two main literature databases and included most relevant papers using some predefined criteria. Among 32 reviewed studies, more than half of them are on transfer learning for Alzheimer’s disease. Brain mapping and brain tumor detection are second and third most discussed research problems respectively. Most used source dataset utilized for transfer learning is ImageNet, which is not a neuroimaging dataset. This suggests that the majority of studies preferred pre-trained models instead of training their own model on a neuroimaging dataset. Although, about one third of studies designed their own architecture, most studies used existing Convolutional Neural Network architectures. MRI was the most common imaging modality. In almost all studies, transfer learning contributed to obtaining better performance, in diagnosis, classification, segmentation of different neuroimaging diseases and problems, than methods without transfer learning. Among all transfer learning approaches, fine-tuning all convolutional and fully-connected layers approach and freezing convolutional layers and fine-tuning fully-connected layers approach demonstrates superior performance in terms of accuracy. Recent approach not only shows great performance but also needs less computational resources and time.