Lecture 9: Navigation II

What this lecture covers

Continuing from the previous lecture, this session moves from the cortical scene-processing regions to the specific populations of neurons that implement navigation: place cells, head direction cells, border cells, and grid cells. It then focuses on reorientation, the problem of regaining your sense of direction after becoming disoriented, using a classic line of experiments in rats, infants, and language-impaired adults. It closes by describing recent evidence that this same spatial system is reused for memory, abstract conceptual spaces, social relationships, and decision-making.

After watching, you can name and describe the four main classes of navigationally relevant neurons, explain the informational-encapsulation finding from reorientation experiments, and describe at least two examples of the navigation system being repurposed for non-spatial cognition.

Key ideas

• Place cells: hippocampal neurons that fire when an animal or person is in one specific location, regardless of which direction they're facing; collectively, a population of place cells forms a "you are here" map.
• Head direction cells: neurons (found in the subiculum and elsewhere) that fire only when the animal is facing a particular direction, forming a neural compass that draws on visual, vestibular, and other cues.
• Grid cells: entorhinal cortex neurons whose firing locations form a repeating hexagonal grid across an environment; they are thought to support path integration, tracking how far an animal has traveled, and were the basis of the Mosers' and O'Keefe's 2014 Nobel Prize.
• Border cells: neurons that fire along the boundaries of an environment, such as walls or a cliff edge, signaling awareness of navigational barriers.
• Reorientation and geometric dominance: when disoriented, rats, toddlers, and even language-impaired adults (with language "shadowing" tying up verbal resources) rely overwhelmingly on the shape of the room to reorient, ignoring salient color or texture landmarks that could resolve the ambiguity outright.
• Informational encapsulation: the idea that a specialized system, such as the reorientation system, may only have access to a restricted set of inputs (here, geometric shape) even when other available information could solve the task.
• Place-cell and behavior link: in a mouse study, when place cells were ambiguous about orientation, the animal's search behavior matched exactly which corner the place cells "thought" it was in, showing place cells directly track (or determine) the animal's belief about its location.
• Repurposing the navigation system: hippocampal pattern similarity tracks not only spatial but temporal distance between remembered events, and grid-like neural signatures appear when people learn abstract two-dimensional concept spaces (such as birds varying in neck and leg length) or navigate social relationships.

Walkthrough

Recap and the structure of navigation (0:10)

The lecture restates the two core navigation questions (where am I, how do I get from here to there) and their components, then reviews retrosplenial cortex's apparent role in linking recognized places to a broader cognitive map, illustrated by patients who can recognize landmarks but cannot use them to find their way or learn new routes.

Place cells (6:13)

Using an auditory demonstration and rodent recordings, the lecture introduces place cells: hippocampal neurons that fire when an animal occupies one specific location, regardless of heading. It covers evidence that this works even without vision, given familiar auditory or tactile cues, that bats (which navigate in 3D) have three-dimensional place fields, and that human place cells have been recorded via clinical electrodes during a virtual-navigation video game task.

Head direction cells and grid cells (20:19)

Head direction cells are introduced as a separate population that fires only for a specific facing direction, drawing on vision, vestibular sensation, and other cues. The lecture then presents grid cells in entorhinal cortex, whose firing fields form a hexagonal grid across an environment, and explains the current hypothesis that they support path integration (tracking distance traveled) alongside head direction cells (tracking orientation). Border cells are introduced as a fourth cell type, firing near environmental boundaries such as walls or cliff edges.

The reorientation problem (30:25)

The lecture defines reorientation: figuring out which way you're facing within an already-recognized environment, distinct from knowing where you are. Gallistel's classic rat experiments show that disoriented rats searching a rectangular room use only the room's shape (long vs. short walls) to guess a hidden food location, splitting their search 50/50 between the two geometrically equivalent corners even when a highly salient colored wall could disambiguate the correct corner. The same pattern appears in toddlers and in language-shadowing adults, suggesting a dedicated, evolutionarily old reorientation system that is informationally encapsulated to geometric cues.

Linking place cells to reorientation behavior (44:34)

A mouse experiment (by former lab member Josh Julian) shows that mice can use wall-stripe orientation to identify which of two rooms they are in, but still fail to use those same stripes to disambiguate their orientation within a room. Simultaneous place-cell recordings show that on trials where the animal searches the "wrong" corner, its place cells are correspondingly flipped 180 degrees, along with head direction and grid cells, demonstrating a tight link between the spatial cell code and the animal's behavioral choice.

Repurposing the navigation system (54:39)

The lecture surveys recent findings that extend the spatial system beyond literal navigation: hippocampal activity patterns track how close in space and time two remembered life events were, grid-cell-like signatures appear when people learn an abstract two-dimensional space of artificial bird features, similar signatures appear in a social-relationship "space," and social place cells in bats represent the location of another observed bat, not just the self. A final example shows place-cell activity in a decision-making rat appearing to simulate upcoming path options at a maze intersection before it commits to a direction.

Before you watch

• Review the previous lecture's coverage of the parahippocampal place area, occipital place area, and retrosplenial cortex, since this lecture builds directly on those regions.
• Recall the concept of a cognitive map and Tolman's rat experiment from the prior lecture.
• Familiarity with double dissociation and informational specificity arguments from earlier in the course will help with the reorientation discussion.

Check your understanding

1. What is the functional difference between a place cell and a head direction cell?
2. In Gallistel's rat reorientation experiments, why did rats search two geometrically equivalent corners equally often even when a salient colored wall could have told them exactly which corner was correct?
3. What does "informational encapsulation" mean, and what evidence from the reorientation studies supports it?
4. In the mouse experiment linking place cells to behavior, what showed that place-cell activity was directly tied to the animal's choice of search corner rather than just reflecting its true location?
5. Give one example from the lecture of the spatial navigation system being used for a non-spatial cognitive task, and explain what evidence supports that reuse.