Neural population geometry: An approach for understanding biological and artificial neural networks. (October 2021)

• Record Type: Journal Article
• Authors: Chung, SueYeon; Abbott, L.F.
• Is Part Of: Current opinion in neurobiology. Volume 70(2021)
• Page Start: 137
• Page End: 144
• Publication Date: 2021-10
• Subjects: Neurobiology -- Periodicals; 573.8
• DOI: 10.1016/j.conb.2021.10.010
• Languages: English
• ISSNs: 0959-4388

Abstract

Abstract: Advances in experimental neuroscience have transformed our ability to explore the structure and function of neural circuits. At the same time, advances in machine learning have unleashed the remarkable computational power of artificial neural networks (ANNs). While these two fields have different tools and applications, they present a similar challenge: namely, understanding how information is embedded and processed through high-dimensional representations to solve complex tasks. One approach to addressing this challenge is to utilize mathematical and computational tools to analyze the geometry of these high-dimensional representations, i.e., neural population geometry. We review examples of geometrical approaches providing insight into the function of biological and artificial neural networks: representation untangling in perception, a geometric theory of classification capacity, disentanglement, and abstraction in cognitive systems, topological representations underlying cognitive maps, dynamic untangling in motor systems, and a dynamical approach to cognition. Together, these findings illustrate an exciting trend at the intersection of machine learning, neuroscience, and geometry, in which neural population geometry provides a useful population-level mechanistic descriptor underlying task implementation. Importantly, geometric descriptions are applicable across sensory modalities, brain regions, network architectures, and timescales. Thus, neural population geometry has the potential to unify our understanding of structure and function in biological and artificial neural networks, bridging the gap between single neurons, population activities, and behavior. Highlights: Manifold-like representations arise when a set of neurons in a biological or artificial neural network exhibits variability in response to stimuli or through internal recurrent dynamics. Approaches focused on analyzing geometric properties of neural populations, i.e. neural population geometry, have emerged as a promising population-level analysis technique connecting neural responses and task implementation. We highlight recent studies of neural population geometry: untangling in perception, classification theory of manifolds, abstraction in cognitive systems, topology underlying cognitive maps, dynamic untangling in motor systems, and a dynamic approach to cognition. Future directions include developing geometric measures as a population-level hypothesis, connecting representational geometry to biophysical properties of neurons, developing theories of neural population geometry for a larger array of tasks.