In a new development, researchers at Carnegie Mellon University have introduced a novel AI-driven dynamic brain imaging which could rapidly map changing electrical activity in the brain with high resolution, high speed, and low cost. The finding is a leap from magnetoencephalography, MRI, and electroencephalography that have been used to study brain activity.
The development is on the back of more than three decades of research of the lead researcher who has focused to develop improved non-invasive dynamic brain imaging technology.
Physiologically, brain activity is distributed over the 3D brain and changes rapidly over time. Imaging of brain function and dysfunction has been explained via several methods, with pros and cons of each method. For instance, MRI is commonly used to study brain activity, but is not fast enough to seize brain dynamics. While EEG is a favorable in comparison to MRI, the less-than optimal spatial resolution of the technology is a major hindrance for its wide utility for imaging.
“Meanwhile, as part of efforts of several decades to develop non-invasive, innovative, functional neuroimaging solutions, work on fluid brain imaging technology that can provide accuracy, is effective, and easy to use to better serve researchers and clinicians is underway,” stated the lead researcher.
The research group is the first to reach the goal with the introduction of AI and multi-scale brain prototypes. With the use of biophysically inspired neural networks, it implies innovation of the deep learning approach to train a neural network that can translate scalp EEG signals precisely back to neural circuit activity in the brain without human intervention.
The study published in Proceedings of the National Academy of Sciences evaluates the performance of the new approach by cognitive brain responses and imaging sensory in 20 healthy human participants.