Artificial Intelligence Tools Are Transforming Neuroscience Research
At the 2025 Society for Neuroscience meeting, researchers highlighted how artificial intelligence is reshaping both fundamental neuroscience research and clinical applications. Deep learning models are now capable of predicting the structure of neuronal ion channels, systems have been developed to anticipate “freezing of gait” episodes in Parkinson’s patients before they occur, and artificial neural networks are increasingly used to model how the brain processes sensory information.
Researchers at the Max Planck Institute for Biological Intelligence demonstrated that integrating biological features such as neuronal diversity and connectivity into artificial neural networks enables these models to outperform traditional systems while requiring less data and learning more rapidly. This approach aims to bring computational models closer to the actual mechanisms of brain function.
At the University of California, scientists developed an AI tool called NeuroInverter, which can infer the ion channel composition of more than 170 neuron types, making it possible to generate “digital twin” neurons. Researchers emphasize that this technology could significantly improve the modeling of neurological and psychiatric conditions such as epilepsy and schizophrenia.
In clinical settings, researchers at Emory University used machine learning algorithms applied to ordinary smartphone videos to classify gait disorders with over 85% accuracy. Meanwhile, a team at the Cleveland Clinic developed a model capable of detecting “freezing of gait” episodes in Parkinson’s patients before they occur, potentially paving the way for future treatments such as adaptive deep brain stimulation.
Finally, researchers at the University of Alabama developed an AI system that can decode the semantic categories of words a person is thinking about from brain signals with 77% accuracy. This advance is considered an important step toward next-generation, language-based brain–computer interfaces for patients who are unable to speak.