1) HIPPOCAMPUS AND THE MTL

Anatomy and functions What structures make up the medial temporal lobe (MTL)?

• hippocampus

• entorhinal cortex

• perirhinal cortex

• parahippocampal cortex

• amygdala

Anatomy and functions What are place cells?

Neurons that fire when an individual is in a particular location.

Anatomy and functions What are grid cells?

Neurons that fire at multiple locations, forming a repeating hexagonal grid pattern

Anatomy and functions What are head-direction cells?

Neurons that fire when the head faces a particular direction.

Anatomy and functions What are boundary cells?

Neurons that respond when an individual is near environmental boundaries.

Anatomy and functions What did the case of H.M. demonstrate about the MTL?

Following MTL surgery, H.M. experienced profound anterograde amnesia and some retrograde amnesia, but relatively preserved implicit and skill learning.

• → demonstrated that the MTL is important for forming new declarative memories.

Anatomy and functions What are the major functions associated with the hippocampus?

• Episodic memory

• relational binding

• spatial navigation

• retrieval of event details

• imagining future events.

Anatomy and functions How might spatial navigation relate to memory?

• Navigation requires remembering locations and relationships between places.

• Spatial-processing mechanisms may also help organise memories and construct imagined events.

Relationship between localisation and explanation What is localisation in neuroscience?

Identifying which brain region is associated with a particular cognitive function

Relationship between localisation and explanation Why is localisation not a complete explanation of a cognitive function?

Knowing where a function occurs does not explain the computations, representations or neural mechanisms that produce it.

Relationship between localisation and explanation Why is saying that a brain region is “involved in” a task insufficient?

It does not specify what the region contributes or how its activity produces the behaviour.

Relationship between localisation and explanation Why does brain activation not prove that a region stores a memory?

Activation may reflect attention, retrieval, processing or support functions rather than storage itself

Marr's three levels of analysis

• Computational: what is the problem being solved, and why

• Algorithmic/representational: what information is represented, and what operations are performed

• Implementational: how is the process physically realised in neural tissue

Marr's three levels of analysis How can Marr's levels be applied to memory?

• Computational: why must the brain distinguish and retrieve experiences

• Algorithmic/representational: how are experiences represented, distinguished, and associated?

• Implementation: how do neural circuits carry out these operations

Marr's three levels of analysis Why are all three levels needed to explain cognition?

Each level addresses a different part of the explanation.

• Identifying a brain region alone cannot explain the problem being solved, the representations used or the underlying neural implementation.

Marr's three levels of analysis What did Aly and Turk-Browne (2018) suggest about the hippocampus?

The hippocampus receives diverse inputs, and its core computations may help explain its involvement in multiple cognitive functions.

Pattern separation, pattern completion, relational processing What is pattern separation

The process of encoding similar experiences as distinct neural representations, reducing confusion between overlapping memories.

Pattern separation, pattern completion, relational processing What is pattern completion

The process whereby a partial cue activates a part of a memory, allowing activation to spread and retrieve remaining features of the experience

Pattern separation, pattern completion, relational processing What is relational processing

Binding and representing relationships between elements, such as objects and locations or different features of an event.

Pattern separation, pattern completion, relational processing What is item memory

Remembering an individual element, such as an object.

Pattern separation, pattern completion, relational processing What is relational memory

Remembering the association between elements

• which location an object occupied.

Pattern separation, pattern completion, relational processing How do pattern separation and pattern completion complement each other?

Pattern separation helps keep similar experiences distinct, while pattern completion helps retrieve a particular experience from incomplete information.

Pattern separation, pattern completion, relational processing How could relational processing explain the hippocampus's involvement in multiple cognitive functions?

The hippocampus may support a common operation—representing relationships—that is useful for spatial navigation, episodic memory, conceptual knowledge and imagination.

Evidence for pattern separation What did Berron et al. (2016) investigate?

Whether hippocampal subfields distinguish repeated images from highly similar images, particularly through pattern separation.

Evidence for pattern separation How did Berron et al. (2016) manipulate image similarity?

Participants viewed images, including exact repeats and highly similar lure images, such as similar chair arrangements.

Evidence for pattern separation What was the task in Berron et al. (2016)?

• Participants responded when the first image in a sequence appeared for the third time.

• Researchers compared brain responses to first presentations, exact repeats and similar lures.

Evidence for pattern separation Why did Berron et al. use high-resolution fMRI?

Smaller voxels helped researchers examine functional differences between hippocampal subfields, including CA1 and the dentate gyrus (DG).

Evidence for pattern separation What is fMRI adaptation?

A reduced BOLD response when a stimulus or representation is repeated.

• → Similar responses to different stimuli may indicate reduced sensitivity to their differences.

Evidence for pattern separation What adaptation pattern occurred in CA1 in Berron et al. (2016)?

CA1 adapted to both exact repeats and similar lures, responding to them relatively similarly.

Evidence for pattern separation What adaptation pattern occurred in the dentate gyrus (DG)?

The DG adapted to exact repeats but not to similar lures

• suggesting that it distinguished highly similar images.

Evidence for pattern separation What did multivoxel pattern analysis (MVPA) reveal in the DG?

Distinct multivoxel activity patterns for highly similar images, indicating sensitivity to image identity.

Evidence for pattern separation What did Berron et al. suggest about CA1 versus the DG?

• DG = sensitive to difference between similar inputs

• CA1 = showed less differentiation between repeats and lures in this experiment.

Evidence for pattern separation What is a limitation of interpreting fMRI adaptation?

Reduced BOLD responses can indicate repetition-related changes, but they do not uniquely establish the exact neural computation responsible.

Evidence for pattern separation What did Chanales et al. (2017) investigate?

How the hippocampus represents overlapping routes as people learn to distinguish them.

Evidence for pattern separation How were routes structured in Chanales et al. (2017)?

Participants viewed image sequences from routes that shared segments before diverging, creating overlapping experiences that could be confused.

Evidence for pattern separation How did participants' route-learning errors change?

Initially, they often selected a competing overlapping route. With learning, they became better at identifying the correct destination.

Evidence for pattern separation How did Chanales et al. analyse neural representations?

They used pattern-similarity analysis to compare representations in the hippocampus and parahippocampal place area (PPA).

Evidence for pattern separation What pattern did the PPA show during route learning?

graded similarity pattern:

• representations of the same route were most similar

• overlapping and non-overlapping routes were more dissimilar.

Evidence for pattern separation What pattern emerged in the hippocampus as participants learned the routes?

Overlapping routes became represented by more dissimilar neural patterns, particularly later in learning.

Evidence for pattern separation What is representational repulsion?

The process whereby similar experiences become represented by more dissimilar neural patterns, helping distinguish overlapping memories.

Evidence for pattern separation What do Berron et al. (2016) and Chanales et al. (2017) jointly support?

The hippocampus can distinguish highly similar inputs and overlapping experiences by forming differentiated neural representations

Evidence for pattern completion What did Hebscher et al. (2021) investigate?

Whether modulating the hippocampal network changes cortical reinstatement during episodic memory retrieval.

Evidence for pattern completion What is cortical reinstatement?

The reactivation during retrieval of cortical activity patterns that were present when a memory was originally encoded.

Evidence for pattern completion What is functional connectivity?

The degree to which activity in two brain regions correlates over time.

Evidence for pattern completion What is resting-state functional connectivity?

Correlated activity between brain regions measured while participants are not performing an externally directed task.

Evidence for pattern completion Why did Hebscher et al. target the angular gyrus with transcranial magnetic stimulation (TMS)?

The hippocampus is too deep for conventional TMS, so researchers stimulated an accessible lateral parietal region connected to the hippocampal network.

Evidence for pattern completion How was the TMS target selected (Hebscher et al., 2021)

Connectivity analysis identified a lateral parietal target connected to the hippocampal network. A control site was also stimulated for comparison.

Evidence for pattern completion What happened during the encoding phase in Hebscher et al. (2021)?

Participants watched short videos and performed an odd-even task as a baseline activity.

Evidence for pattern completion What happened during retrieval?

Participants received verbal memory cues, vividly recalled the events, rated their vividness and answered specific memory questions.

Evidence for pattern completion What did hippocampal-network-targeted TMS dohat happened during retrieval?

It increased reinstatement of encoding-related cortical activity patterns, particularly in the occipital cortex.

Evidence for pattern completion How did targeted stimulation compare with control stimulation?

• Control stimulation produced some reinstatement, including in fusiform and middle cortical regions,

• Targeted stimulation produced more extensive occipital reinstatement.

Evidence for pattern completion What does Hebscher et al. suggest about hippocampal function?

The hippocampal network helps coordinate the reinstatement of cortical representations during episodic retrieval.

Evidence for pattern completion Why is Hebscher et al. relevant to pattern completion?

It supports the idea that hippocampal-network activity helps reactivate distributed cortical features of an event from a partial retrieval cue.

Evidence for pattern completion What is a limitation of network-targeted TMS evidence?

TMS acts on an accessible cortical target and its connected network, so the findings do not establish that the hippocampus alone caused the effect.

Evidence for relational processing What did Tambini et al. (2018) investigate?

Whether TMS targeting a cortical region connected to the hippocampus affects associative memory differently from item memory.

Evidence for relational processing How did Tambini et al. distinguish item memory from associative memory?

Participants remembered objects and their grid locations.

• Item recognition tested memory for the object

• Location placement tested memory for the object-location association.

Evidence for relational processing What were “remember,” “know” and “new” responses? Tambini et al.

• Remember: Recollection of the specific encounter.

• Know: Familiarity without detailed recollection.

• New: Judged not to have been encountered previously.

Evidence for relational processing What did Tambini et al. (2018) find?

Hippocampal-network-targeted stimulation improved associative memory but did not improve item memory.

Evidence for relational processing Why is associative memory evidence for relational processing?

It requires binding an object's identity to its spatial location rather than simply recognising the object.

Evidence for relational processing What predicted the associative-memory benefit in Tambini et al.?

The strength of intrinsic connectivity between the parietal stimulation target and the hippocampus.

Evidence for relational processing What did Finke et al. (2008) suggest about hippocampal amnesia and working memory?

Working-memory deficits can be especially evident when tasks require maintaining relationships between elements.

Evidence for relational processing How does a colour-location change-detection task test relational memory?

Participants remember colours, locations or colour-location pairings, then judge whether a probe matches the original display.

Evidence for relational processing What is a feature-swap error in relational working memory?

Remembering individual features but incorrectly binding them—for example, assigning a red item to the location originally occupied by a blue item.

Evidence for relational processing What is the key finding concerning relational working memory in hippocampal amnesia?

Individuals with hippocampal amnesia can show particular difficulty maintaining associations between elements, such as colour-location pairings.

Evidence for relational processing What is the constructive episodic simulation hypothesis?

Memory retrieves details from past experiences, and imagination recombines these details to construct possible future events.

Evidence for relational processing Why might imagining the future strongly recruit the hippocampus?

Constructing a novel scenario requires flexibly recombining details that may not have occurred together previously.

Evidence for relational processing How does relational processing explain the hippocampus's role in imagination?

The hippocampus may bind retrieved details into a coherent new scenario rather than simply replaying a stored memory.

Evidence for relational processing Why is spatial representation inherently relational?

It requires representing relationships between oneself and objects, as well as relationships among objects in an environment.

Evaluate whether grid cells encode spatial information only What is the characteristic grid-cell firing pattern?

Firing locations form a repeating hexagonal grid. Conjunctive responses can also vary with movement direction and speed.

Evaluate whether grid cells encode spatial information only What angular periodicity is characteristic of grid-like responses?

A 60° periodicity, reflecting the sixfold symmetry of a hexagonal grid.

Evaluate whether grid cells encode spatial information only How can grid-like responses be studied using fMRI?

Researchers compare BOLD responses when movement directions align versus misalign with 60° grid axes, while accounting for movement speed.

Evaluate whether grid cells encode spatial information only What did the virtual-navigation grid-like fMRI paradigm predict?

Entorhinal activity should vary with movement direction and speed, with greater responses at grid-aligned directions.

Evaluate whether grid cells encode spatial information only What did the virtual-navigation findings show?

Grid-like modulation was apparent during fast motion and did not occur at control periodicities such as 45° or 90°.

Evaluate whether grid cells encode spatial information only Why is speed important when studying grid-like responses?

Grid-like modulation can be stronger during faster movement, so speed must be considered when interpreting directional effects.

Evaluate whether grid cells encode spatial information only What evidence suggests grid-like coding extends beyond the entorhinal cortex?

Grid-like responses have been reported in DMN regions, including the ventromedial prefrontal cortex (vmPFC).

Evaluate whether grid cells encode spatial information only How can grid-like coding represent conceptual information?

Conceptual features can be treated as dimensions of an abstract space, with grid-like activity occurring as people move through that space.

Evaluate whether grid cells encode spatial information only How did the bird-concept experiment create a semantic space?

Birds varied in neck and leg length, forming two dimensions associated with different Christmas-related objects.

Evaluate whether grid cells encode spatial information only How was neck and leg length used to “find” christmas objects

certain lengths along the two dimensions were associated with particular christmas objects

Evaluate whether grid cells encode spatial information only What did participants do in the bird-concept experiment?

They learned feature-object associations, watched birds morph along neck and leg dimensions, and judged objects linked to imagined outcomes.

Evaluate whether grid cells encode spatial information only What was the key finding in the semantic grid-like study?

Several brain regions showed 60° periodicity during movement through conceptual space, including the vmPFC.

Evaluate whether grid cells encode spatial information only How did vmPFC grid-like modulation relate to task performance?

Greater modulation was associated with better cognitive performance.

Evaluate whether grid cells encode spatial information only What do grid-like responses in conceptual tasks suggest?

Hexagonal coding mechanisms may represent non-spatial conceptual relationships, not just physical locations.

Evaluate whether grid cells encode spatial information only How might grid-like coding apply to social knowledge?

People can represent others along dimensions such as competence and popularity.

• Moving through this social space can produce grid-like modulation.

Evaluate whether grid cells encode spatial information only Do grid cells encode spatial information only?

Evidence suggests grid-like coding may also structure conceptual and social information in abstract, multidimensional spaces.

Evaluate whether grid cells encode spatial information only What is the embodied-cognition implication of grid-like coding?

Evolutionarily older spatial-processing mechanisms may be repurposed to organise abstract, non-spatial knowledge.

Evaluate whether grid cells encode spatial information only What caution is needed when interpreting grid-like fMRI signals?

Grid-like BOLD modulation is indirect, population-level evidence. It does not prove that every active neuron is a grid cell.