Deep neuroevolution to predict astrocytoma grade from functional brain networks Conference Paper
| Authors: | Stember, J.; Jenabi, M.; Pasquini, L.; Peck, K.; Holodny, A.; Shalu, H. |
| Title: | Deep neuroevolution to predict astrocytoma grade from functional brain networks |
| Conference Title: | 5th International Conference on Intelligent Medicine and Image Processing (IMIP 2023) |
| Abstract: | Whereas MRI produces anatomic information about the brain, functional MRI (fMRI) tells us about neural activity within the brain, including how various regions communicate with each other. The full chorus of conversations within the brain is summarized elegantly in the adjacency matrix. Although information-rich, adjacency matrices typically provide little in the way of intuition. Whereas trained radiologists viewing anatomic MRI can readily distinguish between different kinds of brain cancer, a similar determination using adjacency matrices would exceed any expert's grasp. Artificial intelligence (AI) in radiology usually analyzes anatomic imaging, providing assistance to radiologists. For non-intuitive data types such as adjacency matrices, AI moves beyond the role of helpful assistant, emerging as indispensable. We sought here to show that AI can learn to discern between two important brain tumor types, high-grade glioma (HGG) and low-grade glioma (LGG), based on adjacency matrices. We trained a convolutional neural networks (CNN) with the method of deep neuroevolution (DNE), because of the latter's recent promising results; DNE has produced remarkably accurate CNNs even when relying on small and noisy training sets, or performing nuanced tasks. After training on just 30 adjacency matrices, our CNN could tell HGG apart from LGG with perfect testing set accuracy. Saliency maps revealed that the network learned highly sophisticated and complex features to achieve its success. Hence, we have shown that it is possible for AI to recognize brain tumor type from functional connectivity. In future work, we will apply DNE to other noisy and somewhat cryptic forms of medical data, including further explorations with fMRI. © 2023 IEEE. |
| Keywords: | magnetic resonance imaging; neurons; brain; medical imaging; tumors; evolutionary algorithms; low-grade gliomas; high-grade gliomas; high grade glioma; deep learning; convolution; convolutional neural network; resting state functional mri; convolutional neural networks; resting state; low grade glioma; matrix algebra; neuro evolutions; deep neuroevolution; convolutional neural networks (dnn); deep neuroevolution (dne); evolutionary strategies; high grade glioma (hgg); low grade glioma (lgg); resting state functional mri (rfmri); small data ai; convolutional neural network (dnn); small data; small data artificial intelligence |
| Journal Title | Proceedings of the 5th International Conference on Intelligent Medicine and Image Processing |
| Conference Dates: | 2023 Mar 17-20 |
| Conference Location: | Tianjin, China |
| ISBN: | 979-8-3503-3739-6 |
| Publisher: | Institute of Electrical and Electronics Engineers Inc. |
| Date Published: | 2023-01-01 |
| Start Page: | 1 |
| End Page: | 6 |
| Language: | English |
| DOI: | 10.1109/imip57114.2023.00008 |
| PROVIDER: | scopus |
| DOI/URL: | |
| Notes: | Conference paper -- Source: Scopus |
Altmetric
Citation Impact
BMJ Impact Analytics
Related MSK Work