Hippocampal Representations
egocentric? allocentric?

The term “egocentric” has nothing to do with Freud or selfishness. It’s a geometric term meaning that part of the self is the center of the spatial coordinate frame (“ego” = self). The contrasting term, “allocentric” means something other than the self is the center of the coordinate frame (“allo”= other). A comparison has helped me: think of geocentric and heliocentric models of the solar system.

The question I’m going to address is whether our brains, our perception of the world, our behavior, and our consciousness operate in egocentric or allocentric coordinate frames.

As a starting point, it is generally accepted that immediate sensory and motor interactions with the world work in egocentric frames. For example, visual information, photons, enter the eye and interact with photoreceptors on the retina. This first stage or processing is centered on the retina’s center, the fovea. If the fovea is the center of a “retinotopic map” the representation is egocentric since the fovea is part of the body. Primary visual cortex in the occipital lobe has a retinotopic organization and is, therefore, egocentric. In a similar sense, the map of the skin on the post central gyrus and the motor map on the precentral gyrus are egocentric.

At the final stage of motor output commands are egocentric. The action of a muscle is in reference to a joint, which is part of the body. At a high level of spatial perception, in parietal cortex, space is egocentric: the left side of the world, in reference to the body is processed in right parietal cortex and the right side of the world is processed in left parietal cortex.

In philosophy it is common to call all experience and all consciousness as “subjective” which can be interpreted as ‘egocentric’. What is roughly meant is that all experience comes from the point-of-view of an individual. There is no experiential state without an observer. Think of a photographic image. Photographic images correspond to a retinal image. That is, all photographs have a point of view and a center. The center is (was) a location in the camera.

We appear to be moving towards the conclusion that all brain processing, all of cognition and all of consciousness is egocentric. But this is not the case. Egocentric processing appears to embedded within an allocentric frame in two separate brain systems: the cerebellum and hippocampus.

Sensory Subtraction and the Cerebellum
The first example of allocentric representation we will call “sensory subtraction”. When the sensory apparatus itself moves through space, the change in sensory processing based on this movement is, generally, subtracted from perception and behavioral interaction. Some simple examples: move your gaze from left  to right. Does the world appear to move? No. Close one eye and gently push the lateral side of the open eye with a finger. Does the world move (answer should be ‘yes’). Tilt of your head back and forth, from left to right. Does the world move? Walk around your house or apartment, from one room the next. Are you moving through the world or is the world moving around your? If the world is stationary, the reference frame is allocentric. How is this accomplished?

The general conclusion is that when we move our sensory receptors (mostly in the head) through the world, the sensory change produced from the movement is ‘subtracted’ from expectations resulting in a stable world image. We have, effectively, changed the position of our viewpoint in an world-based (inertial) reference frame. If you plot your updated location in the world, you will use an plot without you at the center of the coordinate frame. The plot will be allocentric. The plot could have its center anywhere on earth (best if its close to you), but it be an earth-based, not a you-based frame.

We are moving towards the following conclusion: when are bodies are still (especially our sensory receptors) we interact with the world in an egocentric manner. When are bodies are in motion (especially our sensory receptors) we subtract movement expectations, before interacting with the world in an egocentric manner.

What information is subtracted? Three types of signals cooperate to maintain a stable world frame. First, the vestibular system of the ear detects position and movement of the head in space. When the head accelerates, the vestibular system sends a subtraction signal. These are typically redundant to visual signals of “optic flow” — the overall shift of data of the visual fields. Finally, during voluntary movements, these two are further aided by a copy of the motor command (“efference copy”) which updates sensory expectations. This triply redundant signal indicates the importance of getting the update with great accuracy. Processing is assumed to occur in the cerebellum. This suggests a fundamental role of the cerebellum is to transfer world interactions from an egocentric frame to a world-centered (inertial) frame. Cerebellar processing is thought to be critical for a center fielder catching a fly ball or a basketball player, taking a jump shot, subtracting self-motion and accurately calculating the forces necessary to accurately take a jump shot.

Path Integration and the Hippocampus.

Hippocampus (curved structure on left) adjacent to the lateral ventricle.

In the 1940s and 50s two famous American psychologists, Tolman and Hull, engaged in debate: when rats solve maze problems, are they using a map-like representation (Tolman) or do they memorize a series of turns (Hull)? Tolman’s maps would be allocentric, while Hull’s turns egocentric.

In the early 1970s John O’Keefe discovered place cells in the rat’s hippocampus. Place cells are neurons that fire when a rat crosses a region of space — the cell’s firing field. O’Keefe argued that the set of place cells covered accessible space and are the neural basis for Tolman’s cognitive map.

On the right is a “map” of the firing rate of a single place cell averaged over a 15 minute recording session. The view is from an overhead camera which tracks the location of a rat in a cylindrical chamber. Average firing rate in each pixel is coded by color, with yellow being a zero rate, and darker colors coding higher rates (highest rate is 15 spikes/sec). Clearly, the overhead camera does an excellent job of organizing the firing of the cell. The rate map produced is allocentric, in that the center of the coordinate frame is earth based, not rat based. The fact that external observer can make a good map suggests, but does not prove, that this allocentric representation is the brain’s representation. But the evidence is strong. Additional support comes from behavioral experiments which show that normal rats solve maze problems as if they have a map, while rats without a functioning hippocampus have great difficulties with complex navigation. Finally, three other cell types in the region of the hippocampus — grid cells, head-direction cells and conjunctive cells — appear to have allocentric representations of space. (The image at the top of this post has a grid cell and another place cell map.)

In brief, a strong argument can be made that the hippocampal system computes allocentric representations of space and uses these map-like representations to solve complex navigational problems. The mechanism for egocentric-to-allocentric transform is not well worked out, but appears to use, in part a process called “path integration”. Path integration works something like this: If you know a previous location on a map, and you know your vector of motion since the last location, you can update your location estimate by adding the movement vector to your prior location. The important signals are an integration of locomotor speed and heading direction (from head-direction cells). Solid behavioral evidence that animals from desert ants, to rats to humans use path integration to as an important component of navigation. Although path integration can short-cut a complete mapping solution, it depends, in part, on allocentric spatial representations.

Conclusions: The ego-to-allo transform has different functions in the cerebellar system and the hippocampal system. The cerebellum appears to be used to solve motor problems with objects that are stable in the inertial frame, while the hippocampal system appears to be used for finding optimal direct paths between locations in earth-based maps. In both cases direct sensory and motor interaction with the world is egocentric, but, in both cases, the egocentric reference frame is embedded within an allocentric map. The short answer to the question at the top of the post is that our brains, our perceptions and our consciousness use both egocentric and allocentric frames.

: