Paving the way for fresh discoveries and technologies, as well as closer integration of the brain, an international team involving UCL designed artificial neural networks that transform raw data from brain activity.
Florida News Times reported new approaches may accelerate such a discovery of how the activity of the brain is related to behavior.
The study, Interpreting wide-band neural activity using convolutional neural networks, published in ELife, co-led by the Kavli Institute for Systems Neuroscience in Trondheim and the Max Planck Institute for Human Cognitive and Brain Sciences Leipzig, and financially backed by Welcome and European Research Council presents that "convolutional neural network" a certain deep learning algorithm type, is able to decipher a lot of different behaviors and stimuli from a great range of brain regions in various species which include humans.
DeepInsight Network Tested
According to Markus Frey, the study's lead researcher from Kavli Institute for Systems Neuroscience, neuroscientists have been able to record bigger and bigger datasets from the brain although deciphering the information contained in that particular data, specifically reading the neural code, remains a difficult problem to solve.
In most circumstances, it remains unknown, what messages are being conveyed. Frey explained they wanted to develop an automatic approach to assess raw neural data for various different types, evading the need to decipher them manually.
The team tried the network known as "DeepInsight," on neural signs from rats, discovering an opera arena and found it was able to accurately foresee the animals' position, head direction, and running speed.
Even with the lack of manual processing, the outcomes were more precise compared to those obtained with conventional assessments.
New Aspects of Neural Code Found
UCL Cell & Development Biology's Professor Caswell Barry, the study's senior author explained, the present methods are missing a lot of potential information in neural recordings since they can only decode the elements they presently understand.
Barry added, their network is able to access much more of the neural code. In addition, by doing so, it teaches the researchers to read several of those other elements.
He also said, they are able to decode neural data more precisely compared to before, although the actual advance is that the network is not limited by the present knowledge.
In a similar report, Medical Xpress specified that the researchers discovered that their model was able to detect new aspects of the neural code, which they exhibited by detecting a formerly unidentified representation of head direction that interneurons encoded in an area of the hippocampus that's among the first to exhibit functional deficiencies in people suffering from Alzheimer's disease.
Through artificial intelligence, the researchers showed in their study that the same network can predict behaviors from various types of recording across areas of the brain. More so, they can be used as well, to conclude hand movements in human volunteers, which they identified by having their network tested on brain activity's preexisting dataset, recorded in people.
Co-author of the study, Professor Christian Doeller of Kavli Institute for Systems Neuroscience and Max Planck Institute for Human Cognitive and Brain Sciences said, this method could enable them in the future to forecast more precisely higher-level cognitive processes in humans, like problem-solving and reasoning.
Commenting on their work, Frey said their framework allows researchers to get a fast automated evaluation of their unprocessed neural data, which saves the time that can be spent on just the most promising suppositions, through the use of more conservative approaches.
Related information about AI and the brain is shown on Bold Business's YouTube video below:
Check out more news and information on Artificial Intelligence in Science Times.