Disorders of memory Flashcards
What are the strengths and weaknesses of using patient studies to inform theories of memory
strengths
patients who present with different characteristic memory impairments arising from different lesion locations provide insight into the fractions within the domain of memory and their respective neural correlates
the systematic errors patients make provide insight into the mechanisms that underpin these processes
particularly important for declarative forms of memory that are challenging to study using animal models
weaknesses
studies are often underpowered and there is an overreliance on case studies - obvious limitations to basing a theory of human memory on a single subject and the nuances of their brain damage
naturally occurring lesions do not respect functional boundaries
historically inaccurate reports of lesion location has driven research to focus on the wrong areas
lack of baseline measurements
spontaneous recovery e.g after stroke mean deficits change with time
overall, while patient observation has laid the foundation for our understanding of memory, convergent evidence from studies of memory in healthy people coupled with animal models is key to achieving a coherent theory of memory systems and their underlying cognitive mechanisms
can we fractionate memory within STM
left hemisphere - verbal WM
Shallice & Warrington 1970 - patient KF
left parietal lesion
tested serial recall and probe recall
reduced STM capacity - only recency effect for very last item
suggesting impaired STM but only seen for verbal tests
other tests support functioning LTM
10 word learning task
right hemisphere - spatial WM
Hanley et al., 1990 - patient ELD
right MCA
impaired learning of new faces despite intact perception of faces (Benton facial recognition test) - able to recognise which face she had seen recently if it was familiar but not unfamiliar
impaired on corsi block tapping
but normal recognition memory for words
different networks for verbal vs visual material Wagner 2009
neuroanatomic overlap for WM and spatial attention networks (LaBar et al., 1999)
- frontal and posterior parietal activations
- fMRI WM task (maintain on-line representation of foveally displayed letters against background distractors); spatial attention (shift visual attention in response to centrally presented directional cue)
- subtraction analysis shows overlap but some distinct areas
delay cells in PFC (Funahashi et al., 1989)
- spatial memory cells that persist to fire during delay
- neural activity for holding information
- presentation of stimulus to specific location requiring orientation later
what is the evidence for dissociable STM and LTM stores?
HM had severe long-term declarative memory impairments but had intact STM (Milner, 1966)
important aspects of STM and WM normal
verbal digit span up to 7 - beyond this where LTM is required v impaired (Drachman & Arbit, 1966)
MTL amnesic patients (Olson et al., 2006)
had to remember three objects, locations or object-location conjunctions
WM for objects and locations normal but memory for conjunctions impaired at 8s delay
what is working memory
Baddeley & Hitch’s working memory model
dissociation in verbal and spatial memory indicate multiple routes into LTM
phonological loop - verbal store
limited capacity store for speech and acoustic information subject to rapid decay - maintained through subvocal articulation
supported by phonological similarity effect (Conrad 1964) word length effect (Baddeley 1975) and developmental correlation between articulation and span (Nicolson, 1981)
visual spatial sketchpad - visuospatial store
limited-capacity storage for visual, spatial and kinaesthetic information. More limited decay
central executive - control mechanisms
proposed to contain a limited pool of general processing capacity - similar to supervisory attention system with purely control function
episodic buffer
interface for integrating perceptual WM subsystems and LTM into a small number of episodes
what have we learnt from patient studies with anterograde memory impairments
Patient HM - Scoville & Milner, 1957
bilateral MTL resection to relieve intractable epilepsy
anterograde amnesia
could not learn new episodic or semantic information regardless of modality
could not learn new words introduced into the lexicon following his surgery (jacuzzi)
BUT Warrington & McCarthy 1988
RFR encephalitic amnesia
tested patient’s knowledge for words that came into use during period of dense retrograde amnesia
able to define AIDS, telecom and provisionals
AND Vargha-Khadem et al., 1997
global anterograde amnesia in 3 patients with brain injuries that occurred in development (birth, age 4, age 9).
MRI - bilateral HPC pathology in all three cases
attend mainstream schools where they attained average level of speech and language competence, literacy and factual knowledge
only episodic fully dependent on HPC
knew who Martin Luther King was, define boast and comprehension questions such as why it is important for the government to ensure meat is inspect before it is sold
although greater plasticity
what is the evidence for dissociable declarative and non-declarative memory systems
preserved procedural memory
Milner & Corkin 1962
able to learn motor-skills such as mirror drawing where the patient looks at a reflection of their hand in a mirror
improved over 10 trials on consecutive days
also, classical conditioning and repetition priming
preserved priming
Warrington & Weiskrantz 1968
amnesic patients given a series of visual images to look at
presented with increasing details of an object - asked to report when they can recognise the object
presentation of image before enables patients to recognise the object faster
Can anterograde amnesia be explained as a deficit in retrieval like retrograde amnesia
Differences in severity of anterograde and retrograde amnesia ini HM and other amnesic patients suggest they are not necessarily linked
May be entirely dissociable (De Renzi & Lucchelli, 1993)
PI head trauma resulting in retrograde amnesia for autobiographic events and those learnt from books/media
intact intelligence, language, arithmtetic and perceptual motor skills
able to learn new information and performs above mean of controls on tests of LTM
pathological forgetting rate over time for information that is not rehearsed
variable evidence
may be that upper limit of encoding deficits reached sooner in anterograde amnesia with no further deficits seen with increasing lesion size (Nadel & Moscovitch, 1997)
What is Ribot’s law and its implications for theory?
Temporal gradient in retrograde amnesia - first in last out such that the earliest memories survive the best specifically following HPC damage
Based on the finding that HM had severe but graded retrograde amnesia
could not recall previously required memory close to the onset of amnesia
but lesion not restricted to HPC and some is spared
cases of flat gradient amnesia challenge this law and the theories based on it
what is the evidence for a role of HPC in episodic memory
initially HM - but post mortem shows much of the HPC intact whereas entorhinal cortex completely removed
AD - tau pathology starts in entorhinal cortex and progresses to HPC (Dickerson & Eichenbaum, 2010)
Vargha-Kadehm 1997
developmental bilateral HPC lesion impairs episodic memory but not semantic
Patient KC (Tulving, 2002)
damage to HPC resulting in loss of episodic but not semantic memory
e.g not able to recall train crash in his village where town had been evacuated but did know the difference between stalagmites and stalactites
with many repetitions patients was able to learn definitions e.g toothbrush performs a daily massage
what is the consolidation hypothesis
Squire’s (1992) Standard consolidation model
HPC-cortical interactions are required for consolidation
memories are eventually transferred to the neocortex when HPC is no longer required for retrieval
thus lesions of HPC do not erase old memories consolidated and robustly represented in cortex
BUT
cannot account for flat gradient
complete lesions of HPC produce flat gradient
HM’s lesions not restricted to the HPC - the medial HPC, posterior body and tail spared (Augustinacket et al., 2014)
animal models: rodent HPC lesions result in flat gradient for contextual memory extending over 100 days pre surgery (Barry & Maguire, 2019)
- value of animal models for focal lesions and mechanism
patients with large HPC lesions display no gradient for the most recent 3 decades with little to no autobiographical memory (Nadel & Moscovitch, 1997)
implausible that there would be an adaptive mechanism for memory consolidation extending beyond the average lifespand for humans throughout history
what are the strengths and weaknesses of alternative accounts to the standard consolidation hypothesis
MTT (Nadel & Moscovitch, 1997)
Memories are encoded in HPC-cortical networks
reactivation leads to multiple traces in HPC linked to cortical networks - each reactivation of a memory occurs ini a different context resulting in expanded and strengthened traces such that memory can be updated with experience
traces in the HPC provide spatial and temporal context while traces in cortex are context-free
retrieval of contextual rich episodes depends on the HPC which provides the spatial and temporal context for a memory
retrieval of remote semantic memories is possible in the absence of the HPC
thus complete HPC lesions will result in a flat gradient for episodic but not semantic memories (and any gradient reflects spared HPC)
also able to account for dissociation in episodic and semantic memory in anterograde and retrograde amnesia
however this suggests that the hippocampus is necessary for encoding semantics - extracted from statistical regularities of events but Vargha-Kadhem demonstrates that it is possible to form semantic memories indepedently of HPC
trace transformation theory (Wincour et al., 2010; Gilboa & Moscovitch, 2021)
- events are retained in multiple forms in parallel with each trace dependent on distinct functional and neural substrates
- no consolidation without representation - change in neuronal representation should correspond to change in its psychological representation (structural-functional isomorphism Moscovitch et al., 2016)
- different representations can coexist
- different information processing/ synaptic consolidation
- episodic representations pHPC - link to episodic
- gist traces aHPC
-schematic traces vmPFC - posterior cortex interactions - link to confabulation
- semantic traces - anterior temporal lobe and
posterior cortex - link to semantic dementia
rather than different strength traces through repeated experience different kinds of memories are stored in different places of the brain
BUT lacks mechanistic detail, specification of temporal dynamics and so predictive power
what is an alternative non-consolidation approach
scene reconstruction theory (Hassabis & Maguire, 2007)
tested 5 patients with bilateral HPC damage (relatively restricted)
asked to imagine and describe ficticious scenarios and possible plausible future episodes
impaired relative to control participants
able to produce relevant details but descriptions lacked spatial coherence and were fragmented
HPC binds together disparate elements of an event/scene
Norman - both encoding and retrieval likely to be hippocampus dependent
but properties of HPC mean it is unlikely to be the basis for storage
Affordo turnover of all HPC dendrites within 4 weeks in rat brain
HPC also implicated in navigation and future planning
place cells (O’Keefe & Nadel 1978) - evidence for cognitive map from single-unit recording in rats
Measurement of place cell activity in the same environment at 5 and 30 days after initial exposure found 15% overlap in neural ensemble but no reduction in the number of cells or activity - reconstruction rather than reactivation of a trace that stabilises over time
receives convergent object and spatial input from entorhinal cortex and parahippocampal cortex
HPC reconstructs original trace in the absence of the original memory via HPC-neocortical interactions - integration of relevant semantic, contextual and sensory components
(could test the effect of intact HPC but deafferented - maybe a good account for HM)
need to think about functional circuits in terms of what information is being processed - if able to support episodic memory - representations in space in time - why not also present and future
what is the patient evidence for the hippocampus in
Graham et al., 2006
impaired perceptual learning in HPC patients for
scenes but not faces
discrimination - same/differnet judgements
categorisation - morphed image of face scene - had to indicate which of two images it was closer to
perceptual learning - performance on novel vs familiar discrimination
Pertzov et al., 2013
patients with MTL lesions (limbic encephalitis) show greater misbinding
correctly identify object but localise it to location of other object
both object and spatial representations intact in isolation
High level of overlap - HPC as a comparator for disambiguating overlapping representations i.e. pattern separation (Gray & McNaughton, 2000)
catastrophic hypermnesia
systematic errors in response - answer that would have been right to a different but relevant question (in comparison to confabulation - false memories)
is anterograde amnesia due to a deficit in encoding, storage or retrieval?
Norman et al., 2019
implanted electrodes in the hippocampus and cortex of epileptic patients brains and recorded activity during viewing and recall of famous faces and places (1 min interference task in between)
SWR rate during picture viewing predicted subsequent free-recall performance
transient increase in SWR rate preceded verbal report of recall by 1-2s and this increase was content-selective, recaptiuatling the sam picture preferences observed during viewing
during recollection high-order visual areas showed content-selective reactivation couples with SWR emission
implicates SWR in encoding new memories
and link to high-level representations (distributed)
BUT short duration
what is Korsakoff’s syndrome
thiamine deficiency (often in alcoholism) that damages the mammillary bodies and medial dorsal nucleus of the thalamus and limbic structures
e.g PZ (Butters & Cermak, 1986)
eminent scientist who had written an autobiography - able to quantify how much is remembered
gradient where older memories relatively spared but most recent memories lost
characterised by confabulation
