vol I chap 3 apendix

Previous: 3.3. Cognitive procedures for creating scientific knowledge.

Appendix 3A. Summary of John O’Keefe Nobel Lecture.¶

Spatial Cells in the Hippocampal Formation.¶

Accepted knowledge or questions under discussion in O´Keefe´s times.

• EC Tolman wrote in 1948 that: “We believe that in the course of learning, something like a field map of the environment gets established in the rat’s brain.”

• Works by Brenda Milner in 1957 on a patient who had “extensive removal of both medial temporal lobes including the hippocampus as well as surrounding structures for the relief of severe temporal lobe epilepsy…. He not only failed to lay down new memories of his day-to-day experiences but was unable to recall ones previously stored. He suffered a profound episodic memory déficit”.

• In order to test predictions of the cognitive map theory, in 1981 Richard Morris devised ”a simple but powerful spatial navigation task which required the animal to approach a location defined by distal environmental cues when started from any one of several different start positions thus forcing it to approach the goal from different directions on different trials.”

• In 2000 Eleanor Maguire and collaborators wrote a paper on Navigation-related structural change in the hippocampi of taxi drivers. “For having some clues about what kind of strategy London taxists used to apply, she explored if they were using either a cognitive map strategy or a route-finding one in which they followed a series of marked paths, a non-hippocampal strategy.”

• In 2013 Yartsev and Ulanovsky reported on Representation of three-dimensional space in the hippocampus of flying bats.

O´Keefe´s contributions or explanations.

• “In the late 1960s, I had been recording somatosensory cells in the thalamus of the rat using our newly developed techniques for recording from single units in freely moving animals and on one occasion had inadvertently positioned electrodes more laterally in the hippocampus. There I found a cell… which clearly was activated by some higher-order aspect of the animal’s movement such as speed of movement of the head… I moved more to the study of the hippocampus in an attempt to see what memories looked like at the single cell level.”

• In a paper published with Jonathan Dostrovsky in 1971 we wrote that “These findings suggest that the hippocampus provides the rest of the brain with a spatial reference map.”

• “In a subsequent paper I wrote in 1971: Each place cell receives two different inputs, one conveying information about a large number of environmental stimuli or events, and the other from a navigational system which calculates where an animal is in an environment independently of the stimuli impinging on it at that moment”.

• “In the book “Hippocampus as a Cognitive Map”, published by me and Lynn Nadel in 1978, we said that: The distance and direction vectors which connect the places in the map of an environment are derived from the animal’s movements in that environment”. … “We predicted the existence of hippocampal signals coding for direction, distance and speed of movement and showed how the known effects of hippocampal lesions could be explained by impaired place learning, navigation, and exploration.”

• In a paper published in 1996 O´Keefe reported about a type of vectorial representation created by the animal to indicate distances in certain directions as well as “to generate predictions about what will be experienced at particular locations forms the basis of learning in the cognitive map”

• In a paper published in 1998 O’Keefe and collaborators reported that “place cell firing rate is modulated by the animal’s speed of running…. Subsequently a small number of pure speed cells have been found in the hippocampus itself.”

• “In papers published in 2009, 2012 and 2015 we described the spatial cells included in the hippocampal formation: place cells tell the animal where it is in a familiar environment, head direction cells tell which direction it is pointing in, boundary cells tell how close it is to a boundary in a particular direction, and grid cells provide the metric for measuring distances between points in the map.”

Appendix 3B. Excerpts from May-Britt Moser Nobel Lecture.¶

Grid Cells, Place Cells and Memory.¶

• “The long-term vision of my lab is to understand how higher cognitive functions are generated by neural activity.”

… • “In this lecture, I will show how the discovery of place cells and grid cells has opened our eyes to some of the secrets of the brain, and how work on these cells has put us on the track of the neural computations responsible for perception of space as well as cognitive brain functions in general.”

…. • “We have not yet deciphered the neural codes for the grid pattern or the localized firing of the place cells but the presence of experimentally controllable firing correlates, combined with the access to activity patterns of multiple discrete cell types, provides us with a powerful model system. This model system can be used to determine not only how specific activity is generated but also how it gets transformed from one cell type to another. By extracting the principles for formation and transformation of firing patterns in the hippocampal-entorhinal space circuit, we can learn a lot about how the cortex operates in general.”

The relationship between grid cells and place cells.¶

• “The hippocampal-entorhinal space circuit consists of several functionally distinct cell types. The first one to be discovered was the place cell, …they fire specifically when animals are in certain locations in the environment. Each cell has its own set of preferred firing locations—or place fields. … The unique firing patterns associated with every single position led to propose that the hippocampus was the cognitive map of the brain.”

… • “The discovery of place cells raised questions about the origin of the place signal… an obvious place to look for counterparts of the place cells was the entorhinal cortex, since nearly all cortical input to the hippocampus is mediated via this region…. Our exploration of this area a decade ago led to the discovery of grid cells.” ….. “Each grid cell fires at multiple locations and these multiple locations form a periodic hexagonal lattice that covers the entire surface of the local environment. …. place fields emerge by linear summation of output from grid cells over a range of spatial scales.”

… • “Grid cells were not the only entorhinal cell type with a spatially modulated firing pattern. In 2006, one year after the discovery of grid cells, we found that grid cells co-localized with head direction cells. … These cells fire if and only if the rat faces a certain direction in the environment.”

…. • “Two year later, we found a third cell type in the same circuit—the border cell. …. Border cells fired if and only if the animal was near one or several of the borders of the local environment, for example a wall of a recording enclosure or an edge of a table.”

…. • “Grid cells, head direction cells and border cells maintained their firing properties across environments.” …. “The existence of multiple functional cell types in the medial entorhinal cortex led us to ask if place cells receive spatial information not only from grid cells but also from head direction cells and border cells.” …. “Experimental results suggest that the hippocampus receives input from a broad spectrum of entorhinal cell types, including grid cells and border cells, but also a large fraction of cells with no detectable spatial firing pattern.” ….

Entorhinal speed cells.¶

• “I have told you that the entorhinal cortex contains spatially modulated cell types such as grid cells and border cells. These cell types provide accurate information about the animal’s current location, but how is this information updated in accordance with the animal’s movements in the environment? .... To estimate the relationship between firing rates and speed, we recorded neuronal activity under strict control of the animal’s running speed. …. The faster the speed, the higher the firing rate. The relationship was linear. The observations in the Flintstone car* suggested that we had come across a new functional cell type—speed cells.” [* Reference to a cartoon cavemobile working by foot power.]

…. • “But if speed cells represent speed, and nothing else, they should also do so in the open field environments where we identified both grid cells and border cells. …. Having established that a considerable number of entorhinal cells respond to instantaneous running speed, we asked whether these cells form a population of their own, or if they overlap with other cell types…. Taken together, our findings show that the medial entorhinal cortex has a large, dedicated population of speed cells, characterized by a linear response to firing rate.”

The hippocampus-memory or space?¶

• “Until now the focus of my talk has been on hippocampal and entorhinal circuits for spatial representation. However, are these circuits only important for space?... The link between space and memory has been known since ancient times…. Place cells may also reflect the memory of a location, expressed as a position-correlated firing pattern in the absence of the sensory inputs that originally elicited the firing.”

… • “Space is used as a framework for storing memories… There is an association between position and memory cues. … place cells are likely to receive information about discrete objects and events in addition to the spatial information they receive from grid cells and border cells in the medial entorhinal cortex.”

Mechanisms for associating events with place-odours as a gateway.¶

• “I would now like to present some recent data that illustrate potential mechanisms for interaction between the lateral entorhinal cortex and the hippocampus. …. When we experience a situation and there are strong odours associated with this situation, the odours are likely encoded in the hippocampus along with the place and other aspects of the experience. When we re-encounter the odour, it may function as a retrieval cue, and the entire situation may be re-experienced in memory.”

…. • “Odour selectivity in the hippocampus and the lateral entorhinal cortex increased strongly during the cue period when the rat was sniffing in the odour port…. Taken together, these findings demonstrate a form of hippocampal learning where improvement in associative performance coincides with increased coupling of 20–40 Hz oscillations in connected cell populations.”

Remapping keeps memories apart.¶

• “One of the greatest challenges for a memory system is to keep memories apart. Avoiding memory interference is particularly a challenge for the hippocampal system, given its involvement in episodic memory…. The overlap of hippocampal representations has until now been tested by recording place cells in pairs of environments—in two differently shaped recording enclosures, or in boxes located in two different rooms…. The replacement of activity patterns between environments was referred to as ‘remapping’”

… • “In sum, these findings suggest that hippocampal place cells have the capacity to form large numbers of independent representations. Spatial patterns never carried over between environments. This independence among representations is exactly what is needed for a memory system to minimize interference.”

• “In the final part of my talk I would like to ask how the brain uses hippocampal and entorhinal maps for navigation. For animals to get from one place to another, there must be mechanisms for reading out the information expressed in place cells and grid cells.”

Appendix 3C. Excerpts from Edvard I. Moser Nobel Lecture.¶

Grid Cells and the Enthorinal Map of Space.¶

1) From psychology to neurophysiology—and back.

• “Yet, while the scientific rigour of behaviourism attracted us, we were at the same time disappointed by the fact that physiology was deliberately left out from most behaviourist theories, due to lack of both methods and concepts….. The potentials of this development—from the laws of learning to its detailed synaptic implementation—really caught May-Britt and me. … Based on experiments on rats running in various types of mazes, Tolman suggested from the 1930s to the 1950s that animals form internal maps of the external environment.”

…. • “In 1971, John O’Keefe and John Dostrovsky discovered that hippocampal cells tend to fire specifically when animals are at certain locations of the environment…. In this overview, I will first review the events that led up to the discovery of grid cells and the organization of a grid cell-based map of space in the medial entorhinal cortex. Then, in the second part, I will present recent work on the interactions between grid cells and the geometry of the external environment, the topography of the grid-cell map, and the mechanisms underlying the hexagonal symmetry of the grid cells.”

2) Moving into unknown territory—the entorhinal cortex.

• “Cells in this area of the entorhinal cortex had clearly defined firing fields. …First of all each cell had many firing fields. Second the fields were not at random positions in the environment, but neighboring fields rather seemed to be separated by a constant distance. There was a peculiar regularity about this pattern, but we did not understand the underlying algorithm.”

3) Grid cells and their functional organization.

• “The multiple firing fields of the cell formed a hexagonal grid that tiled the entire surface space available to the animal, much like the holes in a beehive or a Chinese checkerboard. Many entorhinal cells fired like this, and we named them grid cells.”

… • “There are at least three parameters of variation among grid cells: grid phase, grid scale and grid orientation. …. Following the discovery of the new cell type, we set out to determine how the grid map was organised according to parameters such as phase and scale…. The recordings showed indeed that grid spacing increases in discrete steps.”

4) A universal map.

• …. “we found that simultaneously recorded grid cells maintained scale, phase and orientation relationships across environments and experiences. … Cross-correlation maps of cells from different environments had vertices that were spaced in a hexagonal grid pattern reminiscent of the grid pattern in individual cells.”

… • “Taken together, these observations point to a key difference between grid cells and place cells. Grid modules are universal and rigid—as would be expected by a metric for space—whereas place cells take on a variety of patterns, as would be expected if they also participate in memory for events associated with the locations stored in the place-cell maps. … Thus, grid cells, like the other spatial cell types, exist across many mammalian orders and probably originated early in mammalian evolution, or before.”

5) Grid cells and the geometry of the environment.

• “Taken together, these analyses clearly point to shearing as the mechanism for grid deformation and grid rotation. The findings imply that local boundaries exert distance-dependent effects on the grid pattern. Whether these forces are mediated through the existence of border cells, which intermingle with the grid cells remains to be determined—but it is certainly a possibility.”

6) Fine-scale topography of the grid-cell network.

• “In the next state-of-the-art part of my lecture, I wish to return to the anatomical organisation of the grid network. I showed you that within the resolution constraints of the tetrode technique, there seems to be no topography of grid phase, i.e. grid cells with different grid phases seem to be scrambled, at least at a macroscopic scale.”

7) How is the grid pattern generated?

• “In the final part of my lecture, I would like to address one of the most important questions raised by the discovery of grid cells, which is how these hexagonal patterns are generated. There is a variety of computational models for the formation of grid patterns, but most researchers now believe that attractor mechanisms are somehow involved. In a continuous attractor network, localized firing can be generated by mutual excitation between cells with similar firing locations.”

8) General principles of network function.

• “In some sense the network of grid cells, border cells, head direction cells and place cells provide the components of the neural computations envisaged at the behavior level by Tolman (1948) and at the cell assembly level by Hebb (1949).”

Next: 4.1 On Euclidean geometry and non-Euclidean geometries.