Lectures 16-18 Flashcards
What are the three core stages of human memory processing?
- Encoding – transforming sensory input into a memory trace
- Consolidation/Storage – stabilizing that trace (e.g., during sleep)
- Retrieval – accessing stored information when needed
What are the main ambitions of cognitive‑neuroscience research into human memory?
- Understand how healthy memory works
- Identify its neural basis
- Explore memory disorders (diagnosis, treatment, rehabilitation)
How do researchers manipulate and measure memory in human experiments?
- Manipulate encoding (e.g., by directing attention or perception during study)
- Manipulate retrieval (e.g., using recall vs. recognition tests)
Measure performance via:
- Hits (correctly endorsing old items)
- Misses (failing to recognize old items)
- False alarms (incorrectly endorsing new items)
- Correct rejections (correctly rejecting new items)
In a recognition test, how are “hits”, “misses”, “false alarms” and “correct rejections” defined?
- Hit: “Old” item presented and participant says “Old.”
- Miss: “Old” item presented but participant says “New.”
- False alarm: “New” item presented but participant says “Old.”
- Correct rejection: “New” item presented and participant says “New.”
What does the difference (Hit rate–False‑alarm rate) indicate in memory research?
It indexes memory sensitivity - how well a person can discriminate old from new items independent of response bias.
Why does directing a participant’s attention during encoding affect their later memory performance?
Because attention/perception determine which inputs are deeply processed - only attended information is likely to be encoded into a stable memory trace.
What is recall in human memory?
Recall is the ability to bring to mind contextual details of a past event when the original stimulus is not present; it can be free recall (no cues) or cued recall (with prompts).
What is recognition in human memory?
Recognition is the ability to identify a previously encountered stimulus when it is presented again, relying on a feeling of familiarity and/or retrieval of contextual details.
How do recollection and familiarity differ within recognition memory?
Recollection: retrieval of contextual details from encoding (a form of cued recall) that confirms recognition.
Familiarity: a sense that a stimulus has been seen before without recalling specific contextual details.
Why was the distinction between recollection and familiarity proposed?
Amnesic patients often show impaired recall but variable recognition - some retain familiarity despite losing recollection - indicating two separable recognition processes.
Draw and label the structure of the Medial Temporal Lobe (MTL).
https://imagekit.io/tools/asset-public-link?detail=%7B%22name%22%3A%22screenshot_1745332546246.png%22%2C%22type%22%3A%22image%2Fpng%22%2C%22signedurl_expire%22%3A%222028-04-21T14%3A35%3A46.937Z%22%2C%22signedUrl%22%3A%22https%3A%2F%2Fmedia-hosting.imagekit.io%2Fa0aca6156cbd4fab%2Fscreenshot_1745332546246.png%3FExpires%3D1839940547%26Key-Pair-Id%3DK2ZIVPTIP2VGHC%26Signature%3Didax9H2x-T0X4GCeytvOpg5uw1PaCai-LvIqSLHh-~7zXjlGuUc864uJV8hKY5oe33LUgvMYDOlLALGmoSXf1AXOD99LxLLiDJ4p~eCbKbEiqVLE3-mnZk~sNjjtWml7Eij~0IcRNxEJnOISluyC4m-oLPLnxGjz8h2TwWSPSvmEJgx9AA-TUfQnEMPAQ1pxojFpzLU-vuSyjj6C57LX6nUJrs2vgHm2gxu6TekegFl54IWpAeIbAM3tjgzJUxhbhhNNAXKgsvsZGlstTPRz5poq-wRw9jKS5P2kNGldM8BL2JdQli1QE7LLrlYcrSFwdu1AhhJkKCvLLc5fv-OkbA__%22%7D
Who was “HM” and what did his case reveal about the medial temporal lobes (MTL)?
“HM” was a patient who underwent bilateral removal of his MTL (including hippocampi) to treat epilepsy. He developed profound anterograde amnesia (couldn’t form new episodic memories) while retaining short‑term memory and procedural skills, demonstrating the MTL’s essential role in forming new long‑term declarative memories.
Which additional brain region - beyond the MTL - are critical for normal memory function?
Memory relies on a distributed network including:
- Midline diencephalon (mamillary bodies, anterior thalamus)
- Basal forebrain (cholinergic input to cortex/hippocampus)
- Prefrontal cortex (strategic encoding/retrieval)
- Parieto‑temporal cortex (storage of semantic/contextual details)
- Retrosplenial cortex (contextual/spatial integration)
- Ventral midbrain (dopaminergic modulation of plasticity)
What memory impairments arise from lesions of the midline diencephalon (e.g. mamillary bodies, thalamus)?
Lesions to the mamillary bodies or anterior thalamic nuclei cause anterograde amnesia very similar to MTL damage - patients cannot form new episodic memories - highlighting the diencephalon’s crucial role in the same memory circuit with the hippocampus.
How do MTL lesions compare to midline diencephalon lesions in terms of recall vs. recognition?
Both lesion types produce severe recall deficits, but subtle dissociations exist:
MTL damage often impairs recollection (recall of contextual details) more than familiarity.
Diencephalic damage (e.g. Korsakoff’s syndrome) can produce broader recognition deficits, suggesting slightly different contributions to recollection vs. familiarity processes.
What are the main anatomical subregions of the medial temporal lobe (MTL) and their general roles in memory?
Hippocampus (CA fields, dentate gyrus, subiculum): Binds and consolidates episodic and spatial memories; supports sequence encoding and pattern completion/separation.
Entorhinal cortex: Major cortical gateway to the hippocampus; integrates multisensory inputs and contains grid cells for spatial mapping.
Perirhinal cortex: Encodes and stores item- and object-specific information; supports familiarity-based recognition.
Parahippocampal (postrhinal) cortex: Processes contextual and scene information; contributes to “where” and “which” aspects of episodic memories.
Draw the connections between the PRC, PHC, ERC, Amygdala and Hippocampus (Bonus = show the sensory inputs to each of them).
https://imagekit.io/tools/asset-public-link?detail=%7B%22name%22%3A%22screenshot_1745332894053.png%22%2C%22type%22%3A%22image%2Fpng%22%2C%22signedurl_expire%22%3A%222028-04-21T14%3A41%3A34.405Z%22%2C%22signedUrl%22%3A%22https%3A%2F%2Fmedia-hosting.imagekit.io%2Fe28f67a99bbc4b54%2Fscreenshot_1745332894053.png%3FExpires%3D1839940894%26Key-Pair-Id%3DK2ZIVPTIP2VGHC%26Signature%3Dq82-znBGx46jPvuG75GdkddTkLeStbFY1i-h4UlgWKv0Ir5NMqQUAH9ZgDzX4rkYc2y0sXHpCGts17wcYgwq1BlLQqef1xaPHU0QZqG1RETXvkQRW8PC~H9wSeSqxh~qSRafoZaXA3KWh7~Z5VThm6b925B9lY1iu6nlDSBzsGaLTSXWLPC4GwGVhEfdiwshLBsQdk6GL3sZEcJf5LSHwn9zf1eGJkF~iNv-LGTe8BXHrc7vMlOvMM9FperOONOxt2Cw9tJBitNKnTWLA-7k4D7k25dUh8CkQNGC7dfPfG4zcfQx0RM3yh3XsBKwXAWBBwjhRj96bP28YJ0U65TElQ__%22%7D
How is information hierarchically routed through the medial temporal lobe (MTL) to the hippocampus and what does this imply about the hippocampus’s unique role?
Unimodal and polymodal neocortical areas send sensory‑specific inputs (visual, auditory, somatosensory, visuospatial) into two MTL “what/which” streams:
- Perirhinal cortex (PRC): converges mostly object and item information.
- Parahippocampal (postrhinal) cortex (PHC): converges contextual and spatial information.
- Entorhinal cortex (ERC) receives and integrates PRC + PHC outputs (plus direct amygdala/emotional inputs).
- Hippocampus (DG → CA3 → CA1 → subiculum) sits at the apex, receiving these highly convergent ERC projections.
Because the hippocampus is the final convergence zone for all sensory and contextual streams, its unique role is to bind together disparate elements (objects, places, time, emotion) into unified episodic memories.
What are the four main medial temporal‑lobe subregions and their core functions?
Perirhinal cortex (PRC): encodes “what” (objects, items)
Parahippocampal cortex (PHC): encodes “where/which” (spatial context)
Entorhinal cortex (ERC): gateway that integrates PRC+PHC inputs
Hippocampus: binds integrated inputs into unified episodic memories
Draw the MTL surrounded by it’s two pathways (recollection and familiarity)
https://imagekit.io/tools/asset-public-link?detail=%7B%22name%22%3A%22screenshot_1745333300694.png%22%2C%22type%22%3A%22image%2Fpng%22%2C%22signedurl_expire%22%3A%222028-04-21T14%3A48%3A20.913Z%22%2C%22signedUrl%22%3A%22https%3A%2F%2Fmedia-hosting.imagekit.io%2F303fed99b1e24df6%2Fscreenshot_1745333300694.png%3FExpires%3D1839941301%26Key-Pair-Id%3DK2ZIVPTIP2VGHC%26Signature%3DvTyqX9TvVAbH1FvYKFg0A7-LQPO6rONM28hiSaAQdXfcFBEN6vz5-vVkrpguL5xyyUf~X9u8GrEYFNwBgnD-9KUH5OOtcKmwa4vP1sNTmrhleMgeHcpxeliBh7jLkrbnI0KM2TlI3q9qsVq2mzlK-mZxLco-iGKTBl38PH9gHmXPw3e1iPb8HUSHqa7-~h~buT9BTKvsdPd04G58XRdI~qvdQLEZGxWfyxg681EZndCoP0sgXn7dBVlPS5IhjDOmYDC0yXnLxrYOxliLBNdQ4WGeC4Sfk~umNNNrDjCO9mqQKC-~0lTpPaoRUNuhfwEPMgh~MVY~v8JVrHCIhqvACw__%22%7D
What is the hippocampal–diencephalic–retrosplenial “recall/recollection” pathway?
- Hippocampus
- ↓ via the fornix
- Mammillary bodies
- ↓ via the mammillothalamic tract
- Anterior thalamic nuclei
- ↓ to the retrosplenial (posterior cingulate) cortex
- ↺ back to the hippocampus (via cingulum)
This loop supports episodic recall and recollection by circulating memory‑related information through medial temporal and midline diencephalic structures.
Which brain network underlies the familiarity component of recognition memory?
A non‑hippocampal loop comprising:
- Perirhinal (PR) cortex
- ↔ Dorsomedial thalamic (DM‑Th) nucleus
- ↔ Prefrontal cortex (PFC)
This circuit supports the feeling of prior encounter (“familiarity”) independent of hippocampal recollection.
Draw a familiarity/recollection heterogeneity model of the MTL.
https://imagekit.io/tools/asset-public-link?detail=%7B%22name%22%3A%22screenshot_1745333755474.png%22%2C%22type%22%3A%22image%2Fpng%22%2C%22signedurl_expire%22%3A%222028-04-21T14%3A55%3A55.723Z%22%2C%22signedUrl%22%3A%22https%3A%2F%2Fmedia-hosting.imagekit.io%2F32b8a75b1572469b%2Fscreenshot_1745333755474.png%3FExpires%3D1839941756%26Key-Pair-Id%3DK2ZIVPTIP2VGHC%26Signature%3DQIrTzTuzz7pJLxWQO27gslth5-wyHrP69zd8ft1z-eir5iw~6afrHCYcZvNj-iXzU4LTe65nfy3meTRdWIRnL1KxtkClZdM0aIgE41MWJsWx1VHpJNd1O9VmknrBlXwG6B4YxrasG7vki164QZHgRK-IYQ4E5LNPLKaZgydSQaMnJpBEHnYJfZjJzozXe1GFqeZ~kknkt-gwihlndrrz5Oomldlx8jS~Q18PuPNAySVXMt2Sf9BpGK2WDN5DvjxIW38Pke~f014FZGZMfe0hw5kp1RyM-2iBV2AUxzePXp~kaDlfFOWJRk83LgB4~jjG461syoTl9jBDYejuZ5YL5A__%22%7D
What evidence supports functional heterogeneity between the hippocampus and neighboring medial‑temporal cortices?
Convergent vs. divergent inputs
-Perirhinal/parahippocampal cortices each receive distinct modality‑specific streams (visual, auditory, spatial), whereas the hippocampus sits at their apex, integrating all inputs into unified representations.
Cytoarchitectonic differences
-PRC/PHC/ERC are three‑layer allocortex with local feedforward–feedback loops;
-Hippocampus (DG→CA3→CA1→subiculum) is four‑layer archicortex with specialized laminar wiring for pattern separation/completion.
Distinct memory representations
-Cortices generate graded “familiarity” signals for single items;
-Hippocampus binds “what–where–when” into sparse, orthogonalized episodic engrams capable of pattern completion.
Differing computational algorithms
-MTL cortices implement fast, feedforward matching to support item recognition;
-Hippocampus runs auto‑associative recall and sequence encoding operations (e.g. LTP‑driven pattern completion, theta‑gamma paced ordering).
What do heterogeneity models propose about medial temporal lobe (MTL) components?
Different memory processes (i.e. algorithms) distinguish MTL components
According to heterogeneity models, how should lesions within the MTL affect memory?
Lesions should impair recollection and familiarity somewhat selectively, depending on which MTL component is damaged
What do heterogeneity models predict regarding fMRI activation patterns in recognition memory tasks?
fMRI activation patterns should differentiate recollection and familiarity processes
What does the unitary model argue about recognition memory processes and medial temporal lobe (MTL) components?
It is not recognition memory processes that distinguish MTL components
Differences between MTL structures are not due to separate memory algorithms
According to the unitary model, how should hippocampal lesions affect recollection and familiarity?
Hippocampal lesions should impair recollection and familiarity somewhat equivalently
Following MTL damage, no selective deficit between the two processes
What does the unitary model predict about fMRI activation patterns during recollection versus familiarity?
Activation patterns should show similar MTL engagement across both recollection and familiarity
Why is it important to measure recollection and familiarity separately if they regularly co occur in human life?
In neuroscience we need to be able to tell them apart to understand the neural basis of each of them.
How does the Yes/No recognition test operate and what memory process does it rely on?
Presents one stimulus at a time, so direct comparison between items is impossible
Depends primarily on recollection of contextual or episodic details
Often impaired by hippocampal damage due to reliance on recollective processes
What characterises the forced-choice recognition test and which process can support performance without a hippocampus?
Presents multiple stimuli simultaneously, enabling direct comparison
Can be solved using familiarity judgements alone (“which one feels most known”)
Remains intact in individuals lacking a hippocampus because familiarity is preserved
How does the Remember/Know test dissociate recollection and familiarity?
Encoding: participants study a series of stimuli
Test: items presented one at a time; participant first judges whether they have seen each item before
If “yes,” they report “Remember” if they recall specific details (recollection) or “Know” if it merely feels familiar (familiarity)
Subjective, introspective method in use for 30–40 years
How does a source memory test operationalise recollection?
Encoding: each stimulus is presented with a distinct source cue (e.g. screen location or background colour)
Test: for recognised items, participants indicate the original source attribute
Correct source identification indexes recollection of contextual detail
Incorrect or guessed source responses may reflect reliance on familiarity
Why is subjectivity inherent in paradigms that separate recollection from familiarity?
Participants must introspectively distinguish detailed recall from mere familiarity
Both R/K and source memory tasks depend on subjective judgements
Despite this, these methods are essential to reveal distinct neural bases of each process
What is item recognition in memory paradigms?
Memory for individual stimuli
Typically tested with a yes/no recognition task, where each item is judged as “old” or “new”
What is associative recognition in memory paradigms?
Memory for the relationship between two or more items (e.g. arbitrary pairs)
Tested by presenting item pairs at test and asking whether the combination matches the originally studied pairing
According to the traditional medial temporal-lobe (MTL) model, which structures support item versus associative recognition?
MTL cortices (perirhinal, entorhinal, parahippocampal) support item familiarity
Hippocampus supports binding of distinct elements into associative memories
What is the hippocampal binding hypothesis?
Hippocampus is necessary for associating or “binding” separate elements into a unified memory trace
MTL cortical regions do not perform this binding but relay item information to the hippocampus
What behavioural prediction follows if familiarity alone cannot support associative memory?
Associative recognition should depend exclusively on recollection and thus on hippocampal integrity
Item recognition may be supported by familiarity and hence remain relatively preserved after hippocampal damage
How can one experimentally dissociate item and associative memory demands?
Use a yes/no recognition task for individual items (item recognition)
Use a paired-comparison or forced-choice associative test for studied item pairs (associative recognition)
Draw a diagram showing the difference of item representations in the hippocampus vs the MTL cortex
https://imagekit.io/tools/asset-public-link?detail=%7B%22name%22%3A%22screenshot_1745403551200.png%22%2C%22type%22%3A%22image%2Fpng%22%2C%22signedurl_expire%22%3A%222028-04-22T10%3A19%3A11.022Z%22%2C%22signedUrl%22%3A%22https%3A%2F%2Fmedia-hosting.imagekit.io%2F39217593f8ff4e42%2Fscreenshot_1745403551200.png%3FExpires%3D1840011551%26Key-Pair-Id%3DK2ZIVPTIP2VGHC%26Signature%3DK8xIT2ruIASYWh9ES-SziO96x-pbNiz44P5LNvwFDR~Y2~xu6pF-61MYxD60TsPiTM3FOiUAoaKIj1~DvgnB66z4llwwdQAF1977aDchJHWH4u59vP4Vl4WJ-BTJKvUOnkU-6fWTfbjImCtlzuJ~y2E7ic4A63uDB38tFG9YTvlq-7sHReof6nr22K1L5mBYhpsj10QaZmf~shoXedKUaJIRyQRCiH3ud-tYjnqeAgbKI9a0w~pt7AHQZjHzD1vonWHfZdWDBGSaLPIPt813YrY1Zr6sHassMzLj109lEwYM~UMvGFnAc3WNjOAzHqu2BUX~wRoVxqM9if~PRDpl1g__%22%7D
According to Norman & O’Reilly (2003), how do hippocampal and MTL cortical networks differ in their neural representations of similar events?
Hippocampus: sparse, non-overlapping ensembles that minimise interference
Perirhinal (PRC) and parahippocampal (PHC) cortices: dense, overlapping ensembles that capture similarity
What is pattern separation and which structure primarily implements it?
The process of forming distinct neural codes for similar inputs to reduce memory interference
Implemented by the hippocampus via sparse, orthogonalised representations
What is pattern completion and how does it support recollection?
Retrieving a complete memory trace from a partial or degraded cue
Supported by hippocampal auto-associative networks using the separated representations
How do overlapping cortical representations support familiarity and generalisation?
Shared neural units across inputs emphasise common features
Enable extraction of statistical regularities and rapid “sense” of prior encounter
Which MTL cortices are involved in overlapping representations and what functional role do they play?
Perirhinal cortex (PRC) and parahippocampal cortex (PHC)
Support item familiarity and generalisation across exemplars via overlapping codes
How does goal-directed encoding influence hippocampal versus cortical processing?
Task demands determine whether differences or commonalities between inputs are emphasised
Hippocampus emphasises distinctions (pattern separation) when precise recall is needed
Cortical regions emphasise overlap (generalisation) when rapid familiarity or categorisation suffices
What is a similarity‐insensitive encoding algorithm and how would its input–output transformation appear?
Input representations with any degree of overlap are mapped to identical outputs
No pattern separation or generalisation occurs
Neural similarity between outputs remains the same as between inputs
What is cued recall and how does it illustrate hippocampal pattern completion?
Partial cues prompt the hippocampus to reconstruct the full memory via its auto‐associative network
Behavioural analogue: in a cued‐recall task, a reminder fragment reliably elicits retrieval of the entire event
Everyday example: parking in the same car park each day but slightly different spot - returning to the lot cues recall of exactly where you left your car
Why are human lesion studies considered the gold standard for establishing causality in cognitive neuroscience?
Lesions provide direct evidence of necessity by showing which functions are lost when a region is damaged
Complement correlational data from fMRI with causal inference
What are the main limitations of human lesion studies?
Lesions are incidental and cannot be experimentally controlled
Damage rarely respects neuroanatomical boundaries, especially in densely packed regions like the MTL
Patient heterogeneity: lesion extent, location and aetiology vary across individuals
Why is high-resolution structural MRI critical in lesion research?
Ensures precise mapping of lesion location and extent
Clinical MRI scans alone are often insufficient for detailed neuroanatomical delineation
What neuropsychological and experimental controls are essential when studying lesioned patients?
Administer standardised tests to rule out non-mnemonic deficits (e.g. IQ, language)
Compare patient performance to a matched control group (typically ≥15 participants matched on age, IQ, etc.)
Who is Patient YR and what is the nature of her lesion?
Experienced a cerebrovascular incident in 1986
Sustained relatively focal, bilateral hippocampal damage along its full anterior–posterior extent
How was YR’s hippocampal damage quantified and what were the findings?
Employed high-resolution structural MRI to measure hippocampal cross-sectional areas at successive coronal slices
Compared each measurement to a matched healthy control sample
YR’s hippocampal volumes were consistently more than two standard deviations below the control mean
What methodological caveats apply when interpreting YR’s lesion data?
MRI scans were acquired circa 2000 with lower spatial resolution than current standards
Lesion delineation relied on manual segmentation protocols that may vary between operators
Despite these factors, the volume reductions are large and specific to the hippocampus, supporting a focal lesion interpretation
In Patient YR, what pattern of performance was observed on forced‐choice versus yes/no object recognition tasks using highly similar foils?
METHODS:
- Study phase: eight object silhouettes (e.g. anchor, frog, horse) presented once each
- Test phase:
- Forced‐choice: four-item arrays containing one target and three similar foils (familiarity sufficient)
- Yes/no: single-item probes judged as “old” or “new” (recollection required for high foil similarity)
RESULTS:
- Forced‐choice (familiarity): YR performed within ~1 SD of control mean, indicating preserved familiarity
- Yes/no (recollection): YR scored >3 SD below control mean, reflecting a profound recollection deficit
INTERPRETATION:
- Selective hippocampal damage abolished recollection‐based recognition while sparing familiarity‐based performance
In Bayley et al 2008, how did patients with selective hippocampal damage perform on forced‐choice versus yes/no recognition tests and what conclusion did the authors draw?
METHODS:
- N = 5 patients with confirmed bilateral hippocampal lesions
- Forced‐choice: 12 studied items, 12 test trials each with 4 alternatives (target + 3 similar foils)
- Yes/No: 12 studied items, 60 single‐item test trials (target and foil trials intermixed)
RESULTS:
- Patients were impaired on both tests but showed disproportionately poorer performance on the yes/no test
INTERPRETATION:
- Argued that the larger number of yes/no trials (and greater time/difficulty) accounts for the performance gap
- Concluded that task difficulty, not selective hippocampal involvement in recollection, explains the differential deficit
What did Manns et al (2003) report in N = 7 patients with selective hippocampal damage regarding recognition versus recall?
Both recognition and recall were similarly impaired, suggesting no dissociation between familiarity and recollection in this sample (Manns et al 2003)
What did Jeneson et al (2010) find using a similar-foil recognition paradigm in N = 5 hippocampal patients?
Observed the same pattern of deficits across both forced-choice and yes/no recognition tasks
Provided converging evidence that task difficulty did not uniquely account for performance differences (Jeneson et al 2010)
What is the implication of repeatedly testing the same small patient cohorts across multiple paradigms?
Confirms that different recognition tasks reliably evoke the same pattern of impairment
Does not increase the independence of data points, since patient samples substantially overlap
What are the three types of associative recognition tasks used to probe different binding demands?
- Intra-item associative recognition – recognising a modified version of the same item (e.g. “sandstorm” ↔ “snowstorm”), minimal cross-item binding
- Within-domain inter-item associative recognition – recognising pairs drawn from the same stimulus class (e.g. two faces), moderate binding demands within a single domain
- Between-domain inter-item associative recognition – recognising pairs drawn from different domains (e.g. face + object), high binding demands across distinct representational systems
How did Patient YR perform on intra-item, within-domain and between-domain associative recognition tasks?
METHODS:
- Study phase: sets of four stimuli or stimulus pairs (intra-item, within-domain, between-domain)
- Test phase: forced-choice arrays requiring identification of intact versus recombined or modified pairs
RESULTS:
- Intra-item: performance within normal range relative to controls
- Within-domain inter-item: performance within normal range, indicating preserved same-domain binding
- Between-domain inter-item: profound impairment (well below control mean), indicating a selective deficit in cross-domain binding
INTERPRETATION:
- Selective hippocampal damage spares simple and same-domain associative judgements but abolishes the ability to bind across distinct stimulus domains
What is the domain dichotomy view of medial temporal-lobe memory processing?
Proposes that perirhinal cortex (PRC) and hippocampus both contribute to item memory and within-domain associations but via different algorithms
Suggests that only the hippocampus supports binding across distinct domains
Which MTL region supports item familiarity and within-domain associative familiarity, and by what mechanism?
Perirhinal cortex (PRC)
Uses overlapping, dense neural representations to generalise across similar stimuli, yielding a familiarity signal for single items and same-domain pairs
How does the hippocampus support recollection for items and within-domain associations?
Forms sparse, orthogonalised codes that enable pattern separation at encoding
Utilises auto-associative pattern completion at retrieval to reinstate detailed memories (recollection) for items and same-domain pairs
According to the domain dichotomy view, what kind of associative memory depends exclusively on the hippocampus?
Between-domain associations (e.g. face–object pairs)
Requires hippocampal binding because PRC representations of distinct domains do not overlap sufficiently to support familiarity‐based association
What functional implication arises from convergence of cortical inputs in PRC versus hippocampus?
Similar stimuli converge earlier into PRC, permitting some degree of associative binding for within-domain pairs
Different-domain inputs converge later in hippocampus, making cross-domain binding reliant on hippocampal circuitry
Which white-matter tract carries hippocampal outputs to the mammillary bodies and anterior thalamus?
The fornix
Why are patients who undergo colloid-cyst removal considered an ideal group for studying fornix lesions?
Surgery often produces a focal disconnection of the fornix
Provides a larger, more homogenous patient cohort than typical incidental hippocampal lesions
Allows direct testing of recollection deficits without neocortical damage
What role do mammillary bodies and anterior thalamus lesions play in human amnesia?
Lesions to these diencephalic structures yield profound recall/recollection impairments
Highlight necessity of the subcortical loop beyond the hippocampus alone
How do lesions to the fornix, such as from colloid-cyst removal surgery, affect memory?
Surgical sectioning or thinning of the fornix produces profound anterograde amnesia characterised by impaired recall and recollection
Memory deficits arise despite intact neocortex and MTL cortical regions
What neuroimaging findings confirm fornix damage after colloid-cyst surgery?
Coronal MRI slices show comparison between healthy controls (intact fornix) and patients with bilaterally sectioned or unilaterally thinned fornix
Focal disruption of the C-shaped fornix bundle evident beneath the corpus callosum
Why do neurosurgeons take care to spare the fornix during colloid-cyst excision?
Even slight damage to the fornix fibres leads to severe memory impairment
Preserving the fornix minimises postoperative amnesia while allowing cyst removal
What is the role of the fornix in episodic memory circuits?
Major white-matter tract conveying hippocampal outputs to mammillary bodies and anterior thalamus
Integral link in the hippocampus → fornix → mammillary body → mammillothalamic tract → anterior thalamus → retrosplenial cortex loop supporting detailed recollection
How does fornix damage affect mammillary body integrity in colloid-cyst surgery patients?
The fornix fibres project directly into the mammillary bodies
Surgical disruption of the fornix leads to secondary shrinkage of the mammillary bodies
Why are mammillary body volumes useful as a proxy measure of fornix integrity?
Mammillary bodies have a distinctive shape that allows reliable volumetric measurement on MRI
Fornix damage and mammillary body atrophy correlate strongly across patients
What neuroimaging methods are used to quantify fornix and mammillary body damage?
Coronal structural MRI slices to visualise the C-shaped fornix and the pill-shaped mammillary bodies
Manual or semi-automated segmentation of each structure to derive normalised volume estimates
Box-plot comparisons between patients and matched controls to identify significant atrophy
What is the functional significance of combined fornix and mammillary body atrophy?
Disruption of the hippocampus→fornix→mammillary bodies→anterior thalamus loop abolishes detailed recollection
Severity of memory impairment scales with the degree of mammillary body shrinkage, reflecting fornix disconnection
How does mammillary body volume relate to recall performance in patients with fornix disconnection?
Patients with smaller normalised mammillary body volumes exhibit poorer recall scores
Linear regression shows R² ≈ 0.41, indicating that ~41 % of recall variance is explained by mammillary body atrophy
Demonstrates that the integrity of the fornix→mammillary body pathway is critical for detailed recollection
Which analytic methods were employed to dissociate recollection and familiarity in the colloid-cyst patient cohort?
Remember / Know probability estimation
Structured equation modelling (SEM) of latent recollection and familiarity constructs
Receiver operating characteristic (ROC) curve modelling
What did structured equation modelling reveal about memory processes in patients with severe versus mild mammillary body atrophy?
Severe atrophy group (SMB) showed a marked reduction in recollection estimates (*** p < 0.001)
Familiarity estimates were comparable or modestly elevated in the severe atrophy group
How did ROC curve analysis differentiate recollection and familiarity across patients with differing mammillary body volumes?
ROC-derived recollection parameter was significantly reduced in the severe atrophy group (** p < 0.01)
ROC-derived familiarity parameter did not differ between groups, indicating preserved familiarity despite mammillary body shrinkage
What characterises a block‐design fMRI experiment and why is it suboptimal for detecting medial temporal‐lobe (MTL) activations in memory studies?
Stimuli of the same condition (A or B) are presented continuously in long blocks (e.g. 30 s per condition)
BOLD responses are averaged over each block, limiting temporal resolution
Poor sensitivity to isolate activity evoked by individual memory events, especially in deeper MTL structures
How does an event‐related fMRI design improve sensitivity for memory research?
Presents stimuli in a randomised, time-jittered sequence so each trial’s BOLD response can be modelled separately
Enables linking of specific behavioural outcomes (e.g. remembered vs forgotten trials) to their neural signature
Allows deconvolution of overlapping haemodynamic responses, revealing MTL activations associated with recollection and familiarity
What analytical advances made event-related fMRI feasible for studying recognition memory?
- Development of statistical deconvolution methods to estimate trial-specific haemodynamic responses
- Introduction of general linear model frameworks accommodating rapid event sampling
- Optimisation of experimental timing (jittered inter-trial intervals) to enhance design efficiency and detection power
What are the two principal fMRI paradigms for investigating memory processes?
Scan‐Encoding
- Participants encode stimuli while inside the scanner
- Memory is tested behaviourally outside the scanner afterwards
- Used to identify neural bases of successful encoding
Scan‐Retrieval
- Participants study stimuli before scanning
- Retrieval is tested inside the scanner
- Used to identify neural bases of successful retrieval
What is a logistical challenge of combining scan‐encoding and scan‐retrieval within a single fMRI session?
- Necessitates a very long total scan time
- Increases participant fatigue and movement artefacts
- May require splitting into multiple sessions to maintain data quality
In scan-encoding fMRI studies, how can recollection at test be behaviourally confirmed?
Have participants perform a free-recall test after they exit the scanner
Only feasible when scanning during encoding, since recollection must be probed later
What type of encoding task minimises opportunities for recollection and thus accentuates familiarity in an fMRI study?
A low-level encoding task that restricts conceptual or semantic processing (e.g. simple perceptual judgements)
Reduces the encoding of rich contextual details, suppressing recollection and highlighting familiarity signals
How can a retrieval task be modified to bias decisions toward familiarity during fMRI scanning?
Use a forced-choice or relative-judgement paradigm that requires only a sense of prior encounter
Avoid tasks demanding explicit recall of context, thereby limiting engagement of recollective mechanisms
In Eldridge et al. (2000), how did hippocampal BOLD responses differ for “Remember” versus “Know” judgments during a word‐recognition task?
METHODS:
- Event‐related fMRI with very slow trial timing (one word every 20 s) to isolate individual haemodynamic responses
- Subjects made Remember/Know judgments for each old/new word probe while in the scanner
- Regions of interest defined in left and right hippocampus
RESULTS:
- “Remember” (recollection) trials elicited a clear hippocampal BOLD increase (~0.15–0.25 % signal change peaking ~5 s post-cue) in both hemispheres
- “Know” (familiarity) trials produced little or no hippocampal BOLD signal above baseline
- Correct rejections and misses showed distinct hippocampal signal profiles different from “Remember” trials
INTERPRETATION:
- First fMRI evidence that hippocampal activation selectively tracks recollection rather than mere familiarity Eldridge et al. 2000
During encoding, which medial temporal‐lobe regions predict later recognition with accurate source memory (recollection)?
Bilateral hippocampus
Left parahippocampal cortex
These regions show higher BOLD signal for items subsequently recognised with correct source details Davachi et al. 2003
During encoding, which region predicts later recognition without source memory (familiarity)?
Perirhinal cortex
Exhibits increased activation for items later recognised as “old” but lacking source‐detail recollection Davachi et al. 2003
In an event‐related fMRI study of word recognition using the Remember / Know paradigm, how were recollection and familiarity dissociated and what did hippocampal activity reveal?
METHODS:
- Healthy adult volunteers studied a list of words outside the scanner
- During fMRI, single‐word probes appeared every 20 s; participants made old/new judgments and then classified each “old” item as “Remember” (recollection) or “Know” (familiarity)
- Regions of interest manually delineated in left and right hippocampus
RESULTS:
- “Remember” trials evoked a clear hippocampal BOLD increase (~0.15–0.25 % signal change peaking ~5 s post‐cue) in both hemispheres
- “Know” trials produced no significant hippocampal response above baseline
- Correct rejections and misses showed distinct, lower‐amplitude hippocampal signal profiles
INTERPRETATION:
- Hippocampal activation selectively tracks recollection rather than a general familiarity signal
In an fMRI encoding experiment using a source memory paradigm, which medial temporal‐lobe regions predicted later recollection (source‐correct recognition) versus familiarity (item recognition only)?
METHODS:
- Participants studied words under two encoding tasks: generate a mental image versus silently pronounce each word backward
- Memory test after scanning required old/new judgments plus identification of the encoding task (source) if “old”
- Encoding‐phase BOLD signal was modelled for trials later (a) recognised with correct source (recollection), (b) recognised without correct source (familiarity), (c) forgotten
RESULTS:
- Trials later recognised with correct source elicited greater encoding‐phase activation in bilateral hippocampus and left parahippocampal cortex
- Trials later recognised without correct source (item only) showed increased activation selectively in perirhinal cortex
- Forgotten items showed minimal MTL activation
INTERPRETATION:
- Distinct MTL subregions support different memory processes: hippocampus/parahippocampal cortex for recollection of contextual details, perirhinal cortex for familiarity‐based item recognition
What novel contributions did Ranganath et al. (2004) make to fMRI studies dissociating recollection and familiarity during encoding?
Parametric confidence and source modelling
- Used a 6-point old/new confidence scale plus separate colour-source judgments
- Treated recollection strength as a continuous variable rather than a binary outcome
Finer anatomical precision along the MTL axis
- Demonstrated that posterior hippocampus and parahippocampal/fusiform cortex predict later source-correct recognition (recollection)
- Showed perirhinal cortex activation selectively predicts later item-only recognition (familiarity)
Whole-brain network mapping
- Identified lateral parietal and prefrontal regions whose encoding-phase activity covaries with subsequent recollection versus familiarity Ranganath et al. 2004
In Henson et al. (2003), what evidence supported a special role for the perirhinal cortex in familiarity memory?
Meta-analysis of multiple event-related fMRI studies contrasting New versus Old item recognition
Showed consistent perirhinal cortex clusters exhibiting relative deactivation to Old stimuli (New>Old BOLD response)
Demonstrates that perirhinal cortex reliably tracks familiarity signals across tasks and is distinct from recollection-related regions
What did Kirwin et al. (2008) propose about hippocampal activation and memory strength?
METHODS:
- Event-related fMRI during word encoding with two semantic tasks (size judgement vs animate/inanimate decision)
- Later recognition test with confidence ratings and source-colour judgments
- Analysed hippocampal BOLD as a function of item-confidence, collapsing across correct and incorrect source
RESULTS & CONCLUSION:
- When high-confidence “old” trials were equated for source correctness, hippocampal activation did not differ between source-correct and source-incorrect items
- Argued that hippocampal BOLD reflects memory strength (confidence) rather than the retrieval of specific contextual details
How did Kafkas & Migo (2009) challenge the memory-strength interpretation of Kirwin et al.?
Pointed out that high-confidence source-incorrect trials may still involve recollection of non-target details (e.g. semantic or perceptual aspects)
Argued that matching confidence does not guarantee absence of recollection, since confidence increases with the amount of retrieved detail, not solely target source
Concluded that hippocampal activation likely still indexes qualitative recollection processes rather than a generic strength signal
Why might high-confidence “familiarity” trials still evoke hippocampal activation in source-memory fMRI studies?
Source-correctness only indexes recall of the target context (e.g. colour), not all retrieved details
High confidence increases as more contextual information—target or non-target—is recollected
Participants may recollect alternative features (semantic, perceptual) on “source-incorrect” trials
Thus, hippocampal BOLD on confident non-source trials could reflect residual recollection rather than pure familiarity
How did Experiment 1’s perceptual matching task bias participants towards familiarity rather than recollection?
Participants viewed one “target” picture above and two “choice” pictures below, and had 4 s to select which bottom image was identical to the top
Stimuli were everyday scenes differing only in fine perceptual details (e.g. chimney size)
Short decision window minimised semantic or contextual processing, preventing formation of rich episodic traces.
Why does a fast, perceptual-matching task selectively tap familiarity mechanisms?
Familiarity arises from generalised perceptual overlap—participants need only detect coarse pattern similarity
Recollection requires hippocampal binding of contextual details, which time-limited shallow encoding disrupts
By forcing rapid, detail-focused judgements, the task ensures memory decisions rely on perirhinal‐mediated familiarity signals rather than hippocampal recollection
In the Montaldi et al. (2006) fMRI study designed to isolate familiarity memory, what was the overall experimental workflow including encoding task, retention interval, training and in-scanner test instructions?
Encoding task: Perceptual matching of natural scene triplets (choose which of two bottom images matched the top image on subtle visual features)
Retention interval: 48-hour delay between encoding and scanning
Training: Participants received explicit recollection-vs-familiarity instruction and practice before scanning
In-scanner test:
- Rate familiarity of each scene on a six-point scale (CR, M/F0, F1, F2, F3, R)
- Do not attempt deliberate recollection
- Report any spontaneous recollection that occurs
How were recognition events classified in the Montaldi et al. (2006) fMRI analysis to parametrically model familiarity strength?
CR (Correct Rejection): “Definitely new” images correctly identified as new
M (Miss, F0): Old images incorrectly judged new
F1 (Weak Familiarity): Old images judged “weakly old”
F2 (Medium Familiarity): Old images judged “moderately old”
F3 (Strong Familiarity): Old images judged “strongly old”
R (Recollection): Old images for which spontaneous recollection was reported
What pattern of perirhinal cortex (PRc) BOLD response was observed across increasing familiarity levels (F1-F3) and recollection (R) in Montaldi et al. (2006)?
Parametric deactivation: Left and right PRc signal decreased progressively from F1 through F3 (significant parametric modulation of the old>new contrast)
Recollection (R): Produced PRc activity equivalent to the highest familiarity level (F3), not an additional increase
Implication: PRc is sensitive to the strength of familiarity but does not differentiate recollection from very strong familiarity
How did hippocampal activation distinguish recollection (R) from strong familiarity (F3) in Montaldi et al. (2006), and what does this reveal about hippocampal function?
Selective activation: Both left and right hippocampi showed significantly greater BOLD response for R than for F3 (p < 0.05 for R > F3 contrast)
No graded familiarity effect: Hippocampal activity did not increase across F1–F3 levels
Conclusion: The hippocampus supports recollection-specific retrieval processes, not graded familiarity memory
In Montaldi and Mayes (2010), how were “strong familiarity” (F3) and “recollection” (R) responses matched for memory strength and what were the observed behavioural accuracy and reaction-time (RT) outcomes for these two categories?
Response definitions:
- F3 (strong familiarity): Old items judged “definitely old” without spontaneous recollection
- R (recollection): Old items with reported episodic recall
Accuracy results:
- F1 (weak familiarity): ~0.54
- F2 (medium familiarity): ~0.65
- F3: ~0.88
- R: ~0.89
Reaction-time results (ms):
- F1: ~2100 ms
- F2: ~1940 ms
- F3: ~1850 ms
- R: ~1850 ms
Key finding: No significant difference in accuracy or RT between F3 and R, demonstrating that strong familiarity and recollection can be equated on behavioural strength metrics.
In an fMRI word-recognition task where healthy adults rated each test item on a 6-point confidence scale (1 = “definitely new” to 6 = “definitely old”), how did hippocampal and perirhinal cortex BOLD signals vary across confidence levels, and why does this pattern challenge attributing hippocampal activation solely to recollection?
Hippocampus
- Confidence 1–4: flat, minimal BOLD change
- Confidence 5–6: sharp, non-linear surge in activity
Perirhinal cortex
- Highest activity at lowest confidence (novelty/low-strength detection)
- Monotonic decline in BOLD as confidence increases from 1 → 6
Confound
- High-confidence (5–6) items also have greatest memory strength and accuracy demands, so the hippocampal “spike” may reflect strength effects rather than discrete recollection processes.
In Montaldi & Mayes (2010), how did the authors behaviourally match strong familiarity (F3) and recollection (R), and what did hippocampal fMRI reveal under these conditions?
METHODS:
- Adult human volunteers encoded scene photographs using a perceptual‐matching task (minimising semantic elaboration)
- Retention interval of 48 hours before memory test
- In‐scanner test:
- Rate item familiarity on a three-point scale (F1–F3)
- Report any spontaneous recollection (R) without being prompted to recollect
- Behavioural matching: select F3 and R trials such that mean recognition accuracy was equated (F3 ≈ 0.88; R ≈ 0.89) and mean reaction time was equated (both ≈ 1 850 ms)
RESULTS:
- Hippocampal BOLD signal was significantly greater for R than for F3—even though F3 and R were identical in accuracy and RT
CONCLUSION:
- Hippocampal engagement reflects a specific role in episodic recollection rather than a general response to high memory strength.
In event-related fMRI experiments where recognition accuracy (memory strength) is matched across conditions, what pattern of hippocampal activation is observed for familiarity-based versus recollection-based memory retrieval?
When participants recognise items based solely on a feeling of familiarity (without recalling contextual details), the hippocampus shows no significant activation.
In contrast, when participants report recollection (retrieval of contextual/source details), hippocampal activity is robustly increased.
How does varying the strength of familiarity responses (e.g. parametric modulation from weak to strong familiarity judgments, labelled F1–F3) influence hippocampal activation during retrieval?
Parametric increases in familiarity strength (F1 → F3) produce no systematic change in hippocampal BOLD signal.
This absence of modulation indicates that the hippocampus is not engaged by familiarity, regardless of confidence level, when accuracy is held constant.
Which medial temporal lobe structure selectively supports familiarity-based recognition memory and which supports recollection-based recognition memory when memory strength is equated?
Perirhinal cortex supports familiarity-based recognition (sensitivity to old/new contrasts and graded familiarity).
Hippocampus supports recollection-based recognition (retrieval of contextual/source details), with no engagement by familiarity memory.
In Kafkas et al. (2017), which MTL subregions showed material‐specific linear modulation with rising familiarity for scenes, faces and objects?
Methods (brief):
- Adult humans first encoded 90 scenes, 90 faces and 90 objects using a matching‐to‐sample task.
- During fMRI retrieval, 135 items of each category were presented (fixation 800 ms → stimulus 3 000 ms → implicit baseline 3 000 ms).
- On each trial participants rated familiarity (New, F1–F3) and reported any spontaneous recollection (R).
- Behavioural F3 and R trials were equated for accuracy and reaction time.
Results (familiarity slope):
- Scenes: PRC/entorhinal cortex and parahippocampal cortex (PHC) activity rose linearly from F1 → F3.
- Faces: Amygdala and fusiform gyrus activity rose linearly with increasing familiarity.
- Objects: Perirhinal cortex (PRC) and PHC activity declined linearly as familiarity increased.
Across scenes, faces and objects, what hippocampal activation pattern did Kafkas et al. (2017) observe when comparing recollection (R) to the strongest familiarity level (F3) under matched performance?
Design control: F3 and R trials matched for recognition accuracy (~0.80–0.90) and reaction time (~1 800 ms).
Hippocampal result: Robust BOLD increases for R but no activation for F3 in all three stimulus categories.
Inference: Hippocampus is selectively engaged by episodic recollection, regardless of stimulus type or overall memory strength.
In Kafkas et al. (2020), how did distinct thalamic subnuclei respond to recollection (R) versus graded familiarity (F1–F3) for objects, faces and scenes when behavioural performance was matched?
METHODS:
- Healthy adult volunteers encoded 90 objects, 90 faces and 90 scenes in a matching-to-sample task before scanning.
- Retrieval session in the MRI: 135 trials per category (fixation 800 ms → stimulus 3 000 ms → implicit baseline 3 000 ms).
- On each trial participants rated familiarity (New, F1–F3) and reported spontaneous recollection (R).
- F3 and R trials were selected to equate mean recognition accuracy and reaction time.
RESULTS:
- Anterior thalamus (AT): significant activation for R > F3, but no parametric change across F1→F3.
- Mediodorsal thalamus (MD): activity increased linearly with familiarity strength (F1→F3) for all stimulus types.
- Ventrolateral thalamus (VL): parametric familiarity effect only for faces (F1→F3).
- Posterior thalamus (PT): parametric familiarity effect only for scenes (F1→F3).
CONCLUSION:
- AT supports episodic recollection specifically, whereas MD integrates familiarity signals across domains, with VL and PT showing material-specific familiarity responses.
In Kafkas et al. (2020), what did psychophysiological interaction (PPI) analyses reveal about mediodorsal thalamus connectivity with medial temporal lobe cortices as familiarity strength increased?
METHODS:
- Seed region: MD thalamus peak from the familiarity contrast (F1–F3).
- PPI analysis examined MD connectivity with perirhinal cortex (PRC) and parahippocampal cortex (PHC) during familiarity trials.
- Familiarity strength indexed by hits minus false alarms across F1–F3.
RESULTS:
- Objects: MD–PRC connectivity increased with familiarity strength (R² = 0.42, p < 0.01).
- Faces: MD–PRC connectivity increased with familiarity (R² = 0.30, p < 0.05).
- Scenes: MD–PHC connectivity increased with familiarity (R² = 0.32, p < 0.05).
CONCLUSION:
- The mediodorsal thalamus serves as a hub, functionally coupling with material-specific MTL cortices in proportion to familiarity strength, thus orchestrating the subjective experience of familiarity.
In Kafkas et al. (2020), using an fMRI remember-know paradigm with matched accuracy and reaction time, how did distinct thalamic subnuclei respond to recollection (R) versus graded familiarity (F1–F3) for objects, faces and scenes?
Anterior thalamus (AT):
- Significant activation for recollection (R > F3)
- No parametric modulation across familiarity levels (F1→F3)
Mediodorsal thalamus (MD):
- Domain-general linear increase in activity with rising familiarity (F1→F3) for all stimulus types
Ventrolateral thalamus (VL):
- Familiarity-strength modulation (F1→F3) only for faces
Posterior thalamus (PT):
- Familiarity-strength modulation (F1→F3) only for scenes
What familiarity-specific functional connectivity did Kafkas et al. (2020) find between the mediodorsal thalamus (MD) and medial temporal lobe cortices for different stimulus categories?
Objects: MD–perirhinal cortex (PRC) connectivity increased with object familiarity (R² = 0.42, p < 0.01)
Faces: MD–PRC connectivity increased with face familiarity (R² = 0.30, p < 0.05)
Scenes: MD–parahippocampal cortex (PHC) connectivity increased with scene familiarity (R² = 0.32, p < 0.05)
Based on Kafkas et al. (2020), what unifying role does the mediodorsal thalamus play in familiarity processing?
Acts as a hub that functionally couples with material-specific MTL cortices in direct proportion to familiarity strength
Orchestrates the subjective experience of familiarity across objects, faces and scenes
What is the flow of information through amygdala subnuclei during acquisition of a fear memory?
Lateral nucleus: receives CS (conditioned stimulus) and US (unconditioned stimulus) inputs
Basolateral & accessory basal nuclei: integrate and refine associative signals
Central nucleus: final common output specifying fear responses
Which three downstream targets of the central nucleus of the amygdala mediate behavioural, autonomic and hormonal components of fear?
Periaqueductal gray: orchestrates defensive behaviours (e.g. freezing)
Lateral hypothalamus: drives autonomic responses (heart rate, blood pressure)
Bed nucleus of the stria terminalis: regulates hormonal stress responses (e.g. HPA axis)
ow do cue-based (Pavlovian) and context-based fear conditioning differ in terms of neural substrates and behavioural measures?
Cue-based conditioning:
- Procedure: pair discrete CS (tone/light) with aversive US (shock)
- Neural substrate: amygdala circuits alone
- Measure: freezing or startle to CS
Contextual conditioning:
- Procedure: expose animal to context paired with US
- Neural substrates: hippocampal context representation plus amygdala
- Measure: freezing when returned to same context
Why is ‘freezing’ duration widely used as a proxy for fear in rodent conditioning studies?
Freezing (cessation of all movement except respiration) is a robust, quantifiable defensive behaviour
Longer freezing indicates stronger associative fear memory
Applicable across sensory modalities and experimental setups
Which neural pathways convey the CS (tone) and US (shock) respectively to the lateral nucleus of the amygdala during fear conditioning?
CS pathway: auditory thalamus → lateral amygdala (LA), with a parallel route via auditory cortex → LA
US pathway: somatosensory thalamus → LA, with additional input from somatosensory cortex → LA
What behavioural measures indicate successful Pavlovian versus contextual fear conditioning in rats?
Pavlovian (cue) conditioning: rat freezes when the tone CS is presented alone
Contextual conditioning: rat freezes when returned to the training context without any tone or shock
How enduring is the fear memory formed by a single CS–US pairing in rats?
Rats retain the conditioned fear response lifelong, demonstrating remarkably durable one‐trial learning
Why is single‐trial auditory fear conditioning considered an adaptive form of learning?
Enables rapid detection and recall of danger signals
Increases an animal’s chance of survival by promoting immediate defensive behaviours (e.g. freezing) on re‐encountering the cue or context
In awake rats, how does the firing rate of lateral amygdala (LA) neurons to a tone CS change following tone–shock pairings?
METHODS:
- Simultaneous extracellular recordings from 4 LA neurons during:
- Ten CS‐only trials (tone alone) before conditioning
- Ten CS–US pairings (tone co‐terminating with footshock)
- Ten CS‐only trials after conditioning
RESULTS:
- Before conditioning: tone evoked modest, sparse spiking across LA units
- After conditioning: tone evoked a robust, large increase in spike rate in all recorded LA neurons
In awake rats, what happens to lateral amygdala (LA) neuron responses to a tone CS if the tone–shock presentations are explicitly unpaired?
METHODS:
- Simultaneous extracellular recordings from four LA neurons
- Phase 1 (Pre-test): ten trials of tone CS alone (3 sec tone; inter-trial interval ~30 sec)
- Phase 2 (Unpaired “conditioning”): ten tone CS presentations and ten footshock US presentations, pseudorandomly interleaved so that CS and US never co-occur
- Phase 3 (Post-test): ten trials of tone CS alone, identical to Phase 1
RESULTS:
- No increase in tone‐evoked firing after unpaired presentations; some neurons even showed a decrease in spiking
- Indicates that LA potentiation requires temporal CS–US association, not mere exposure to both stimuli
In rats with differential auditory fear conditioning (CS–, no shock; CS+, shock), what happens to lateral amygdala (LA) neuron firing and freezing when the central nucleus (CE) is inactivated with lidocaine?
METHODS:
- Adult rats underwent differential fear conditioning: CS– tone never paired; CS+ tone paired with footshock.
- During retrieval, simultaneous extracellular recordings from LA neurons.
- Temporary CE inactivation via lidocaine infusion before CS+ presentation.
- Measured CS-evoked LA spiking and freezing.
RESULTS:
- LA firing: CS+ still evoked a large increase in LA spike rate despite CE inactivation.
- Freezing: CE‐lidocaine abolished freezing to CS+ (no conditioned response).
CONCLUSION:
- LA activity reflects associative fear memory retrieval, whereas CE is required to transform that memory signal into overt defensive behaviour.
How do LA neuron firing and freezing behaviour dissociate when a novel (CS–) tone is presented in a context previously paired with shock?
METHODS:
- Rats conditioned to freeze in Context A (shock context) but never heard CS– tone there.
- During retrieval in Context A, presented CS– tone while recording LA neurons and measuring freezing.
RESULTS:
- Freezing: High levels of contextual freezing on CS– trials.
- LA firing: No increase in CS–‐evoked LA spike rate above baseline.
CONCLUSION:
- Contextual fear engages downstream fear circuits to drive freezing without LA spiking to the novel tone, confirming LA responses are CS associative, not behavioural per se.
What do combined LA recordings and CE inactivation reveal about the functional division of labour within the amygdala during fear memory retrieval?
LA: encodes the presence of learned CS+ associations (memory trace), showing robust spiking on CS+ retrieval regardless of freezing
CE: acts as the behavioural effector node; CE inactivation prevents freezing even though LA memory representations remain intact
OVERALL:
The amygdala’s memory‐behaviour circuit separates storage (LA) from expression (CE), ensuring that learned fear associations can be retrieved independently of immediate behavioural output.
In rats fear-conditioned to a tone CS, what happens to freezing behaviour and lateral amygdala (LA) neuron firing when the tone is repeatedly presented without shock (extinction) and what does this reveal about the underlying memory trace?
METHODS:
- Adult rats first underwent auditory fear conditioning: 3 blocks of 10 tone–footshock pairings (CS+ → US).
- Immediately afterwards, extinction comprised 3 blocks of 10 tone-only trials (CS alone; 3 sec tone; ~30 sec inter-trial interval).
- Simultaneous measurements of:
- Freezing behaviour (percentage time immobile per block)
- Extracellular spiking from LA neurons during each CS presentation
RESULTS:
- Freezing: peaked by the end of conditioning then declined progressively across extinction blocks.
- LA firing: CS-evoked spike rates likewise rose during conditioning but fell in parallel during extinction.
INTERPRETATION:
- Extinction does not erase the CS–US memory in LA; rather, it forms a new CS→no-US memory that inhibits expression of the original trace (via prefrontal control), causing both behavioural and neural responses to diminish.
In infralimbic (vmPFC) lesion studies, how does removal of infralimbic cortex affect extinction learning and recall in rats?
Methods:
- Adult rats received bilateral excitotoxic lesions of infralimbic cortex (vmPFC) or sham surgery.
- Day 1: all animals underwent auditory fear conditioning (tone CS + footshock US) followed immediately by extinction training (tone alone, 20 trials).
- Day 2: extinction recall test (tone alone, 20 trials) without any further extinction training.
Results:
- Extinction learning (Day 1): both lesion and sham groups showed equivalent decreases in freezing across extinction trials.
- Extinction recall (Day 2): sham rats exhibited low freezing to the tone, whereas vmPFC-lesioned rats continued to show high freezing.
Conclusion:
- Infralimbic/vmPFC is necessary for recall of the extinction memory but not for initial extinction learning of the CS–US association.
What firing patterns do infralimbic cortex neurons exhibit during extinction recall versus extinction learning?
Methods:
- Chronic single-unit recordings from infralimbic cortex in awake rats.
- Day 1: habituation, fear conditioning (tone + shock), then extinction training (tone alone).
- Day 2: extinction recall (tone alone) in the same context.
Results:
- Conditioning & extinction learning (Day 1): infralimbic neurons showed minimal tone-evoked changes.
- Extinction recall (Day 2): a subset of infralimbic neurons exhibited a marked increase in firing rate time-locked to tone onset.
Conclusion:
- Infralimbic cortex neurons selectively encode recall of the extinction (“safety”) memory.
How does extinction modify fear memory representations in the amygdala and PFC?
Conceptual basis: extinction does not erase the original CS→US trace in lateral amygdala (LA) but instead creates a new CS→no-US memory.
Evidence:
- LA neurons still retain potentiated responses to the CS after extinction, but behavioural freezing and LA spiking are suppressed by infralimbic inputs.
- Infralimbic lesion or inactivation unmasks the original fear response without altering LA plasticity.
Conclusion:
- Extinction establishes a parallel “safety” memory in infralimbic/vmPFC that inhibits retrieval of the original fear memory in the amygdala.
Which neural projection is thought to underlie top-down inhibition of fear expression during extinction recall?
Anatomy: infralimbic cortex sends glutamatergic projections to inhibitory interneurons in the lateral and basal nuclei of the amygdala.
Function: activation of this pathway during extinction recall suppresses output from the central nucleus, thereby preventing freezing.
Supportive data: optogenetic or chemogenetic stimulation of infralimbic→LA projections reduces conditioned freezing, while pathway blockade impairs extinction recall.
How are infralimbic/PFC-dependent extinction recall deficits linked to PTSD?
Clinical observations: patients with PTSD display impaired recall of extinction and persistent fear responses to trauma-related cues.
Neuroimaging: reduced ventromedial PFC (homologous to rodent infralimbic) activation during extinction recall in PTSD patients.
Interpretation: failure to engage vmPFC “safety” circuits may underlie pathological persistence of fear in PTSD.
In rodent models of fear conditioning, what distinct roles do prelimbic (PL) and infralimbic (IL) prefrontal cortices play in the expression and suppression of conditioned fear?
Prelimbic cortex (PL):
- Drives expression of conditioned fear by exciting the central nucleus of the amygdala (CE)
- Lesions or inactivation reduce freezing to the CS+ without affecting extinction learning
Infralimbic cortex (IL):
- Encodes and recalls extinction (“safety”) memories that suppress LA/CE output
- Lesions or inactivation impair extinction recall, resulting in persistent freezing
What is the anatomical loop underlying hippocampal‐dependent recollection as described by Aggleton & Brown (1999)?
Hippocampus → Fornix → Mammillary bodies → Anterior thalamus → Retrosplenial cortex → back to Hippocampus
How does hippocampal amnesic patient YR’s performance differ on forced‐choice versus yes/no recognition tests, and what does this reveal about familiarity and recollection?
4‐alternative forced‐choice recognition:
- YR performs at control levels, relying on preserved familiarity processes
Yes/no recognition with similar foils:
- YR is severely impaired (>3 SD below control mean), indicating recollection deficits
In “old versus new” fMRI recognition contrasts, how does perirhinal cortex BOLD activity typically respond to familiar items, and what does this signify?
Shows relative deactivation for previously seen (“old”) items compared to novel (“new”) items
Reflects sensitivity to graded familiarity strength rather than simple novelty detection