NUS Scientists Harness Machine Learning to Uncover New Insights into the Human Brain


Assistant Teacher Thomas Yeo led an inter-disciplinary research study group to uncover new insights into the cellular architecture of the human brain.

An inter-disciplinary research study group led by scientists from the National University of Singapore (NUS) has actually effectively utilized artificial intelligence to uncover new insights into the cellular architecture of the human brain.

The group showed a technique that instantly approximates specifications of the brain utilizing information gathered from practical magnetic resonance imaging (fMRI), making it possible for neuroscientists to presume the cellular homes of various brain areas without penetrating the brain utilizing surgical ways. This method might possibly be utilized to evaluate treatment of neurological conditions, and to establish new treatments.

“The underlying pathways of many diseases occur at the cellular level, and many pharmaceuticals operate at the microscale level. To know what really happens at the innermost levels of the human brain, it is crucial for us to develop methods that can delve into the depths of the brain non-invasively,” stated group leader Assistant Teacher Thomas Yeo, who is from the Singapore Institute for Neurotechnology (SINAPSE) at NUS, and the A * STAR-NUS Medical Imaging Research Study Centre (CIRC).

The new research study, carried out in cooperation with scientists from the Netherlands and Spain was initially reported online in clinical journal Science Advances on 9 January 2019.

Unravelling the intricacy of the human brain

The brain is the most detailed organ of the body, and it is comprised of 100 billion afferent neuron that remain in turn linked to around 1,000 others. Any damage or illness impacting even the tiniest part of the brain might lead to extreme problems.

Presently, most human brain research studies are restricted to non-invasive methods, such as magnetic resonance imaging (MRI). This limitations the evaluation of the human brain at the cellular level, which might provide unique insights into the advancement, and prospective treatment, of numerous neurological illness.

Various research study groups around the world have actually utilized biophysical modelling to bridge this space in between non-invasive imaging and cellular understanding of the human brain. The biophysical brain designs might be utilized to imitate brain activity, making it possible for neuroscientists to gain insights into the brain. Nevertheless, a number of these designs count on extremely simple presumptions, such as, all brain areas have the very same cellular homes, which scientists have actually understood to be incorrect for more than 100 years.

Building virtual brain designs

Asst Prof Yeo and his group dealt with scientists from Universitat Pompeu Fabra, Universitat Barcelona and University Medical Center Utrecht to evaluate imaging information from 452 individuals of the Human Connectome Task. Leaving from previous modelling work, they permitted each brain area to have unique cellular homes and made use of machine learning algorithms to instantly quote the design specifications.

“Our approach achieves a much better fit with real data. Furthermore, we discovered that the micro-scale model parameters estimated by the machine learning algorithm reflect how the brain processes information,” stated Dr Peng Wang, who is the very first author of the paper, and had actually carried out the research study when he was a postdoctoral scientist in Asst Prof Yeo’s group.

The research study group discovered that brain areas associated with sensory understanding, such as vision, hearing and touch, display cellular homes opposite from brain areas associated with internal idea and memories. The spatial pattern of the human brain’s cellular architecture carefully shows how the brain hierarchically processes details from the environments. This type of hierarchical processing is a crucial function of both the human brain and current advances in expert system.

“Our study suggests that the processing hierarchy of the brain is supported by micro-scale differentiation among its regions, which may provide further clues for breakthroughs in artificial intelligence,” stated Asst Prof Yeo, who is likewise with the Department of Electrical and Computer System Engineering at the NUS Professors of Engineering.

Next actions

Progressing, the NUS group strategies to use their method to analyze the brain information of specific participants, to much better comprehend how specific variation in the brain’s cellular architecture might relate to distinctions in cognitive capabilities. The group hopes that these most current outcomes can be an action towards the advancement of individualised treatment strategies with particular drugs or brain stimulation techniques.

Recommended For You

About the Author: livescience

Leave a Reply

Your email address will not be published. Required fields are marked *