Learning an Efficient Hippocampal Place Map from Entorhinal Inputs Using Non-Negative Sparse Coding

Distinct spatial maps and multiple object codes in the lateral entorhinal cortex

The lateral entorhinal cortex (LEC) is a major cortical input area to the hippocampus, and it is crucial for associative object-place-context memories. An unresolved question is whether these associations are performed exclusively in the hippocampus or also upstream of it. Anatomical evidence suggests that the LEC processes both object and spatial information. We describe here a gradient of spatial selectivity along the antero-posterior axis of the LEC. We demonstrate that the LEC generates distinct spatial maps for different contexts that are independent of object coding and vice versa, thus providing evidence for pure spatial and pure object codes upstream of the hippocampus. While space and object coding occur by and large separately in the LEC, we identified neurons that encode for space and objects conjunctively. Together, these findings point to a scenario in which the LEC sustains both distinct space and object coding and associative space-object coding.

Learning the Vector Coding of Egocentric Boundary Cells from Visual Data

The use of spatial maps to navigate through the world requires a complex ongoing transformation of egocentric views of the environment into position within the allocentric map. Recent research has discovered neurons in retrosplenial cortex and other structures that could mediate the transformation from egocentric views to allocentric views. These egocentric boundary cells respond to the egocentric direction and distance of barriers relative to an animal's point of view. This egocentric coding based on the visual features of barriers would seem to require complex dynamics of cortical interactions. However, computational models presented here show that egocentric boundary cells can be generated with a remarkably simple synaptic learning rule that forms a sparse representation of visual input as an animal explores the environment. Simulation of this simple sparse synaptic modification generates a population of egocentric boundary cells with distributions of direction and distance coding that strikingly resemble those observed within the retrosplenial cortex. Furthermore, some egocentric boundary cells learnt by the model can still function in new environments without retraining. This provides a framework for understanding the properties of neuronal populations in the retrosplenial cortex that may be essential for interfacing egocentric sensory information with allocentric spatial maps of the world formed by neurons in downstream areas, including the grid cells in entorhinal cortex and place cells in the hippocampus.SIGNIFICANCE STATEMENT The computational model presented here demonstrates that the recently discovered egocentric boundary cells in retrosplenial cortex can be generated with a remarkably simple synaptic learning rule that forms a sparse representation of visual input as an animal explores the environment. Additionally, our model generates a population of egocentric boundary cells with distributions of direction and distance coding that strikingly resemble those observed within the retrosplenial cortex. This transformation between sensory input and egocentric representation in the navigational system could have implications for the way in which egocentric and allocentric representations interface in other brain areas.

Gated transformations from egocentric to allocentric reference frames involving retrosplenial cortex, entorhinal cortex, and hippocampus

This paper reviews the recent experimental finding that neurons in behaving rodents show egocentric coding of the environment in a number of structures associated with the hippocampus. Many animals generating behavior on the basis of sensory input must deal with the transformation of coordinates from the egocentric position of sensory input relative to the animal, into an allocentric framework concerning the position of multiple goals and objects relative to each other in the environment. Neurons in retrosplenial cortex show egocentric coding of the position of boundaries in relation to an animal. These neuronal responses are discussed in relation to existing models of the transformation from egocentric to allocentric coordinates using gain fields and a new model proposing transformations of phase coding that differ from current models. The same type of transformations could allow hierarchical representations of complex scenes. The responses in rodents are also discussed in comparison to work on coordinate transformations in humans and non-human primates.

Coordinated drift of receptive fields in Hebbian/anti-Hebbian network models during noisy representation learning

Recent experiments have revealed that neural population codes in many brain areas continuously change even when animals have fully learned and stably perform their tasks. This representational 'drift' naturally leads to questions about its causes, dynamics and functions. Here we explore the hypothesis that neural representations optimize a representational objective with a degenerate solution space, and noisy synaptic updates drive the network to explore this (near-)optimal space causing representational drift. We illustrate this idea and explore its consequences in simple, biologically plausible Hebbian/anti-Hebbian network models of representation learning. We find that the drifting receptive fields of individual neurons can be characterized by a coordinated random walk, with effective diffusion constants depending on various parameters such as learning rate, noise amplitude and input statistics. Despite such drift, the representational similarity of population codes is stable over time. Our model recapitulates experimental observations in the hippocampus and posterior parietal cortex and makes testable predictions that can be probed in future experiments.

Place cells dynamically refine grid cell activities to reduce error accumulation during path integration in a continuous attractor model

Navigation is one of the most fundamental skills of animals. During spatial navigation, grid cells in the medial entorhinal cortex process speed and direction of the animal to map the environment. Hippocampal place cells, in turn, encode place using sensory signals and reduce the accumulated error of grid cells for path integration. Although both cell types are part of the path integration system, the dynamic relationship between place and grid cells and the error reduction mechanism is yet to be understood. We implemented a realistic model of grid cells based on a continuous attractor model. The grid cell model was coupled to a place cell model to address their dynamic relationship during a simulated animal's exploration of a square arena. The grid cell model processed the animal's velocity and place field information from place cells. Place cells incorporated salient visual features and proximity information with input from grid cells to define their place fields. Grid cells had similar spatial phases but a diversity of spacings and orientations. To determine the role of place cells in error reduction for path integration, the animal's position estimates were decoded from grid cell activities with and without the place field input. We found that the accumulated error was reduced as place fields emerged during the exploration. Place fields closer to the animal's current location contributed more to the error reduction than remote place fields. Place cells' fields encoding space could function as spatial anchoring signals for precise path integration by grid cells.

Artificial intelligence insights into hippocampal processing

Advances in artificial intelligence, machine learning, and deep neural networks have led to new discoveries in human and animal learning and intelligence. A recent artificial intelligence agent in the DeepMind family, muZero, can complete a variety of tasks with limited information about the world in which it is operating and with high uncertainty about features of current and future space. To perform, muZero uses only three functions that are general yet specific enough to allow learning across a variety of tasks without overgeneralization across different contexts. Similarly, humans and animals are able to learn and improve in complex environments while transferring learning from other contexts and without overgeneralizing. In particular, the mammalian extrahippocampal system (eHPCS) can guide spatial decision making while simultaneously encoding and processing spatial and contextual information. Like muZero, the eHPCS is also able to adjust contextual representations depending on the degree and significance of environmental changes and environmental cues. In this opinion, we will argue that the muZero functions parallel those of the hippocampal system. We will show that the different components of the muZero model provide a framework for thinking about generalizable learning in the eHPCS, and that the evaluation of how transitions in cell representations occur between similar and distinct contexts can be informed by advances in artificial intelligence agents such as muZero. We additionally explain how advances in AI agents will provide frameworks and predictions by which to investigate the expected link between state changes and neuronal firing. Specifically, we will discuss testable predictions about the eHPCS, including the functions of replay and remapping, informed by the mechanisms behind muZero learning. We conclude with additional ways in which agents such as muZero can aid in illuminating prospective questions about neural functioning, as well as how these agents may shed light on potential expected answers.

Place cells may simply be memory cells: Memory compression leads to spatial tuning and history dependence

The observation of place cells has suggested that the hippocampus plays a special role in encoding spatial information. However, place cell responses are modulated by several nonspatial variables and reported to be rather unstable. Here, we propose a memory model of the hippocampus that provides an interpretation of place cells consistent with these observations. We hypothesize that the hippocampus is a memory device that takes advantage of the correlations between sensory experiences to generate compressed representations of the episodes that are stored in memory. A simple neural network model that can efficiently compress information naturally produces place cells that are similar to those observed in experiments. It predicts that the activity of these cells is variable and that the fluctuations of the place fields encode information about the recent history of sensory experiences. Place cells may simply be a consequence of a memory compression process implemented in the hippocampus.