Brain-wide representations of prior information in mouse decision-making

Charles Findling,Felix Hubert,International Brain Laboratory,Luigi Acerbi,Brandon Benson,Julius Benson,Daniel Birman,Niccolò Bonacchi,Sebastian Bruijns,Matteo Carandini,Joana A Catarino,Gaelle A Chapuis,Anne K Churchland,Yang Dan,Felicia Davatolhagh,Eric EJ DeWitt,Tatiana A Engel,Michele Fabbri,Mayo Faulkner,Ila Rani Fiete,Laura Freitas-Silva,Berk Gerçek,Kenneth D Harris,Michael Häusser,Sonja B Hofer,Fei Hu,Julia M Huntenburg,Anup Khanal,Chris Krasniak,Christopher Langdon,Peter E Latham,Petrina Y P Lau,Zach Mainen,Guido T Meijer,Nathaniel J Miska,Thomas D Mrsic-Flogel,Jean-Paul Noel,Kai Nylund,Alejandro Pan-Vazquez,Liam Paninski,Jonathan Pillow,Cyrille Rossant,Noam Roth,Rylan Schaeffer,Michael Schartner,Yanliang Shi,Karolina Z Socha,Nicholas A Steinmetz,Karel Svoboda,Charline Tessereau,Anne E Urai,Miles J Wells,Steven Jon West,Matthew R Whiteway,Olivier Winter,Ilana B Witten,Anthony Zador,Yizi Zhang,Peter Dayan,Alexandre Pouget

DOI: https://doi.org/10.1101/2023.07.04.547684

2024-11-18

Abstract:The neural representations of prior information about the state of the world are poorly understood. To investigate them, we examined brain-wide Neuropixels recordings and widefield calcium imaging collected by the International Brain Laboratory. Mice were trained to indicate the location of a visual grating stimulus, which appeared on the left or right with prior probability alternating between 0.2 and 0.8 in blocks of variable length. We found that mice estimate this prior probability and thereby improve their decision accuracy. Furthermore, we report that this subjective prior is encoded in at least 20% to 30% of brain regions which, remarkably, span all levels of processing, from early sensory areas (LGd, VISp) to motor regions (MOs, MOp, GRN) and high level cortical regions (ACAd, ORBvl). This widespread representation of the prior is consistent with a neural model of Bayesian inference involving loops between areas, as opposed to a model in which the prior is incorporated only in decision-making areas. This study offers the first brain-wide perspective on prior encoding at cellular resolution, underscoring the importance of using large scale recordings on a single standardized task.

Neuroscience

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to understand how the brain represents prior information about the state of the world. Specifically, by analyzing the whole - brain neural activity of mice in decision - making tasks, the researchers explored how prior probabilities are encoded in different regions of the brain. The core of this problem lies in understanding whether and how the brain uses prior knowledge to optimize the decision - making process, especially in the case of probability changes in the location of visual stimuli.

The research background indicates that although a great deal of research has been carried out theoretically on the neural representation of Bayesian inference, the question of how and where prior knowledge is specifically represented in the brain remains controversial. Some studies suggest that prior information may be processed only in high - level decision - making regions, while other models believe that the entire brain may be involved in the processing of such prior information, that is, the brain may operate like a large - scale Bayesian network, performing probability inference in all directions.

To explore this issue, the authors used data provided by the International Brain Laboratory, which included electrophysiological recordings from 242 brain regions and wide - field calcium imaging data from mice performing the same decision - making task. By analyzing these data, the researchers found that mice were able to estimate the current prior probability and thereby improve the accuracy of their decisions. Further research shows that this subjective prior is encoded by at least 20% to 30% of the brain regions, covering all processing levels from the primary sensory area to the motor area and then to the high - level cortical area. This is consistent with a neural model involving circulation between various regions, rather than a model that combines prior information only in the decision - making region. This finding provides a new perspective for understanding how the brain processes prior information and emphasizes the importance of using large - scale recordings and a single standardized task.