AI Determines How the Brain Predicts and Processes Thoughts
September 10, 2024
Summary: A new study using artificial intelligence has provided novel insights into how the brain predicts future events and processes information. Researchers discovered that the brain’s spontaneous activity, even without external stimuli, plays a critical role in how we think and feel.
By analyzing local field potentials (LFPs), they uncovered how the brain remains active in anticipating possible scenarios, even in a resting state. These findings could lead to better diagnostic tools and treatments for neurological diseases.
Key Facts:
- AI helped uncover how the brain predicts events even without stimuli.
- Local field potentials (LFPs) are key in brain activity during resting states.
- This research may improve diagnosis and treatment of brain diseases.
Source: FAU
In a pioneering study, the two scientists Dr. Patrick Krauss and Dr. Achim Schilling from the Cognitive Computational Neuroscience Group at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) have now used artificial intelligence to gain major insights into how our brains work that may substantially change our understanding of human thought processes and emotions.
What comes next in a sentence? What will I see next? How does the environment change when I do this and what happens to my body when I do that?
The human brain is continuously occupied at all levels of complexity and abstraction with predicting what will happen next. Known as predictive coding, this is considered one of the main tasks of the human super organ, making adaptive behavior possible and allowing us to find our bearings in our surroundings.
Dr. Patrick Krauss and Dr. Achim Schilling from the Cognitive Computational Neuroscience Group at the Chair of Computer Science 5 Pattern Recognition at FAU have succeeded in underlining this widely held hypothesis and contributing new findings in their recent study.
Collaboration with Epilepsy Center at Uniklinikum Erlangen
The two physicists and neuroscientists analyzed the spontaneous activity of the human brain using auto-encoders, an advanced form of artificial intelligence that allows patterns and connections to be perceived in the complex quantities of data provided by our brain that would have been unachievable using more traditional methods.
This was made possible thanks to their collaboration with researchers from the Epilepsy Center at Uniklinikum Erlangen (speaker: Prof. Dr. med. Hajo Hamer). Epilepsy patients in the Center receive electrodes implanted into their brains before the surgical removal of epileptogenic foci.
Using the particularly rare and therefore especially valuable data received as a result, the researchers made a discovery that led to groundbreaking results: Certain spontaneous activities in our brain known as local field potential events (LFPs) were able to give decisive indicators regarding how our brains work.
These spontaneous signals seem to play an important role in how our brains process information even in the absence of external stimuli.
New avenues for research
“In our study, we realized that our brains are constantly progressing through active states defined by these LFPs. It is as if our brains are constantly playing through various options for what might happen next even if we are not doing or perceiving anything in particular and not receiving any external stimuli at that moment in time,” stresses Dr. Patrick Krauss.
“We have also discovered that the form of these LFPs can determine the direction of information flux within the brain. This could give us important insights into how thoughts and feelings are processed in our minds,” adds Dr. Achim Schilling.
Findings that not only open new avenues for research but may also lead to better methods for diagnosis and treatment for brain disease. These AI-based methods can also be used in conjunction with normal EEG or MEG measurements, where electrodes are attached to the surface of the skull to measure brain activity.
“Knowledge of what our brains usually do while we are at rest can be put to good use for diagnostic purposes. If we can gain an ever better understanding of how our brains work and process information, that will allow us to develop more specific methods of diagnosis and treatment for neurological diseases,” emphasizes Dr. Achim Schilling.
“If, for example, the brain enters a state that does not correlate with the external stimuli, that may be an indication of pathological changes.”
Increased reciprocity between technology and brain research
Whilst AI is being used as a tool, the results of the study from the two FAU researchers may also help to further develop AI. The long-term aim: AI inspired by neuroscience that is capable of continuously making predictions, even if it is not currently processing any input.
“This may be particularly useful in AI systems incorporated into vehicles, for example, especially when bearing safety in mind,” explains Dr. Achim Schilling.
Dr. Patrick Krauss continues, “Even if there is not much traffic and the car is only driving straight ahead on the highway, it would be beneficial for the AI to be considering in the background which traffic incidents could occur to which it may potentially have to react.”
The study from Dr. Patrick Krauss and Dr. Achim Schilling therefore shows that the synergetic connection between AI and brain research is capable of expanding the boundaries of what is known about cognitive processes and brain function, eventually leading to innovative new approaches in medical diagnosis and therapy.
The increasing fusion of technology and brain research also indicates how decisive interdisciplinary approaches are for decoding the complex systems found in nature.
With their discoveries, the FAU researchers are approaching nothing less than a better understanding of the perhaps most complex of all systems: the human brain.
About this AI and human cognition research news
Author: Sarah Armbrust
Source: FAU
Contact: Sarah Armbrust – FAU
Image: The image is credited to Neuroscience News
Original Research: Open access.
“Deep learning based decoding of single local field potential events” by Patrick Krauss et al. NeuroImage
Abstract
Deep learning based decoding of single local field potential events
How is information processed in the cerebral cortex? In most cases, recorded brain activity is averaged over many (stimulus) repetitions, which erases the fine-structure of the neural signal.
However, the brain is obviously a single-trial processor. Thus, we here demonstrate that an unsupervised machine learning approach can be used to extract meaningful information from electro-physiological recordings on a single-trial basis.
We use an auto-encoder network to reduce the dimensions of single local field potential (LFP) events to create interpretable clusters of different neural activity patterns.
Strikingly, certain LFP shapes correspond to latency differences in different recording channels. Hence, LFP shapes can be used to determine the direction of information flux in the cerebral cortex.
Furthermore, after clustering, we decoded the cluster centroids to reverse-engineer the underlying prototypical LFP event shapes. To evaluate our approach, we applied it to both extra-cellular neural recordings in rodents, and intra-cranial EEG recordings in humans.
Finally, we find that single channel LFP event shapes during spontaneous activity sample from the realm of possible stimulus evoked event shapes. A finding which so far has only been demonstrated for multi-channel population coding.
Join our Newsletter
I agree to have my personal information transferred to AWeber for Neuroscience Newsletter ( more information )
Sign up to receive our recent neuroscience headlines and summaries sent to your email once a day, totally free.
We hate spam and only use your email to contact you about newsletters. You can cancel your subscription any time.
Tags
AIArtificial Intelligencebrain researchcognitionDeep LearningFAUinformation processingMachine LearningneurobiologyNeuroscience
canisgallicus 