Task representations in neural networks trained to perform many cognitive tasks

• Published in: Nature neuroscience Vol. 22; no. 2; pp. 297 - 306
• Main Authors: Yang, Guangyu Robert, Joglekar, Madhura R., Song, H. Francis, Newsome, William T., Wang, Xiao-Jing
• Format: Journal Article
• Language: English
• Published: New York Nature Publishing Group US 01.02.2019; Nature Publishing Group
• Subjects: 631/378/116/1925; 631/378/2649; Analysis; Animal Genetics and Genomics; Artificial neural networks; Behavioral Sciences; Biological Techniques; Biomedical and Life Sciences; Biomedicine; Brain; Brain - physiology; Cognition; Cognition - physiology; Cognitive ability; Cognitive tasks; Computational neuroscience; Computer Simulation; Decision making; Decision Making - physiology; Electrophysiology; Humans; Learning; Learning - physiology; Learning theory; Memory, Short-Term - physiology; Models, Neurological; Neural networks; Neural Networks, Computer; Neurobiology; Neurons; Neurons - physiology; Neurophysiology; Neurosciences; Representations; Selectivity; Short term memory
• ISSN: 1097-6256; 1546-1726; 1546-1726
• DOI: 10.1038/s41593-018-0310-2

The brain has the ability to flexibly perform many tasks, but the underlying mechanism cannot be elucidated in traditional experimental and modeling studies designed for one task at a time. Here, we trained single network models to perform 20 cognitive tasks that depend on working memory, decision making, categorization, and inhibitory control. We found that after training, recurrent units can develop into clusters that are functionally specialized for different cognitive processes, and we introduce a simple yet effective measure to quantify relationships between single-unit neural representations of tasks. Learning often gives rise to compositionality of task representations, a critical feature for cognitive flexibility, whereby one task can be performed by recombining instructions for other tasks. Finally, networks developed mixed task selectivity similar to recorded prefrontal neurons after learning multiple tasks sequentially with a continual-learning technique. This work provides a computational platform to investigate neural representations of many cognitive tasks. Prefrontal cortex can be flexibly engaged in many different tasks. Yang et al. trained an artificial neural network to solve 20 cognitive tasks. Functionally specialized modules and compositional representations emerged in the network after training.