Abstract:Decades of theoretical and empirical work have suggested the hippocampus instantiates some form of a cognitive map. Yet, tests of competing theories have been limited in scope and largely qualitative in nature. Here, we develop a novel framework to benchmark model predictions against observed neuronal population dynamics as animals navigate a series of geometrically distinct environments. In this task space, we show a representational structure in the dynamics of hippocampal remapping that generalizes across brains, discriminates between competing theoretical models, and effectively constrains biologically viable model parameters. With this approach, we find that accurate models capture the correspondence in spatial coding of a changing environment. The present dataset and framework thus serve to empirically evaluate and advance theories of cognitive mapping in the brain.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: How to quantify and compare the effectiveness of different theoretical models in explaining the construction process of cognitive maps in the hippocampus (especially in the CA1 region). Specifically, the researchers developed a new framework for evaluating the predictive ability of different theoretical models for the observed dynamic changes in neuronal populations, especially when animals navigate a series of geometrically different environments. Through this method, the researchers hope to identify a representational structure that is ubiquitous in different brains, which can distinguish different theoretical models and effectively limit the range of biologically feasible model parameters. ### Research Background and Problems The hippocampus and its related neocortical areas are considered to support various learning and memory processes through cognitive maps, which are formed by the activities of principal neurons. There is increasing evidence that such maps are formed in metric space to support flexible goal - directed navigation and long - term memory functions. However, there is currently no consensus on how to compare the predictions of different theoretical models and the relationship between these predictions and empirical observations. Therefore, this study aims to collect a large number of data sets through experiments, which can be quantitatively compared in a common task space, so as to gain a deeper understanding of the organization of neural coding in the brain. ### Specific Problems 1. **Identifying a Ubiquitous Representational Structure**: The researchers attempted to identify a representational structure that is ubiquitous in different brains during the remapping process in the CA1 region. 2. **Comparing Theoretical Models**: Evaluate the effectiveness of these models by directly comparing the observed neuronal population dynamics with the predictions of different theoretical models. 3. **Constraining Model Parameters**: Determine which model parameters can biologically reasonably explain the changes in spatial coding. ### Methods The researchers recorded the large - scale neuronal population activities in the CA1 region of the hippocampus, using a geometric deformation paradigm with rich conditions (including 10 different geometric environments). Through the representational similarity analysis (RSA) framework, they constructed representational similarity matrices (RSM) and used non - metric multi - dimensional scaling analysis (nMDS) and cluster analysis to visualize and quantify the representational similarity in different environments. ### Results The study found that the remapping patterns in the CA1 region showed a high degree of consistency across different brains, and this consistency increased with experience. In addition, the model based on local boundary distance and direction can better predict the representational structure in the CA1 region, which provides an important empirical basis for further research. ### Significance This study provides a new quantitative benchmark for evaluating and promoting theoretical research on the construction of cognitive maps in the brain. Through this method, the researchers can not only better understand the role of the hippocampus in spatial navigation, but also provide a more stringent test standard for future theoretical models.

Identifying representational structure in CA1 to benchmark theoretical models of cognitive mapping

How to build a cognitive map: insights from models of the hippocampal formation

An Entorhinal-Hippocampal Model for Simultaneous Cognitive Map Building.

Cognitive Memory and Mapping in a Brain-Like System for Robotic Navigation

Parallel cognitive maps for multiple knowledge structures in the hippocampal formation

Idiosyncratic navigation determines mouse CA1 representational structure in a multicompartment environment

A Novel Mathematic Entorhinal-Hippocampal System Building Cognitive Map

Multi-Scale Extension in an Entorhinal-Hippocampal Model for Cognitive Map Building.

Entorhinal-hippocampal Interactions Lead to Globally Coherent Representations of Space

The mosaic structure of the mammalian cognitive map

Memory and Space: Towards an Understanding of the Cognitive Map

Latent representations in hippocampal network model co-evolve with behavioral exploration of task structure

Vision Enhanced Neuro-Cognitive Structure for Robotic Spatial Cognition

Remapping revisited: how the hippocampus represents different spaces

Topological schemas of cognitive maps and spatial learning in the hippocampus

Decoding the Cognitive map: Learning place cells and remapping

Forming cognitive maps for abstract spaces: the roles of the human hippocampus and orbitofrontal cortex

Recurrent Connectivity Shapes Spatial Coding in Hippocampal CA3 Subregions

A model of how hierarchical representations constructed in the hippocampus are used to navigate through space

A Biological Model for Robotic Cognitive Mapping and Global Planning

A hierarchy of associations in hippocampo-cortical systems: cognitive maps and navigation strategies