Memory and Learning II

MEMORY & LEARNING PART 2: NEUROANATOMY OF MEMORY(General)

  • Hippocampi – specifically involved in registration and storage of new memories. Bilateral removal of hippocampus leads to severe anterograde amnesia similar to that seen in Korsakoff’s (but they do not have the other cognitive deficits associated w/Korsakoff’s- believed to be due to diffuse cortical atrophy). Common cause: inflammation, anoxia, temporal lobectomy, stroke (supplied by PCA)
    • Right hippocampal removal is associated w/ deficits in spatial learning and memory
    • Left hippocampal removal associated w/recall of verbal material
  • Amygdala – believed to be involved in associating a stimulus with a reward (necessary for learning)
  • Diencephalon (particularly the mammilary bodies) – the precise role is unclear, but the mammilary bodies consist of nerve fibers which connect the hypothalamus and the thalamus.
    • Bilateral damage results in global anterograde amnesia
      • Patients w/Korsakoff’s have been shown to have hemorrhagic lesions in the region around the diencephalon. (amnesia has also been shown in patients following trauma, stroke or tumors in this area)
  • Basal forebrain (nucleus accumbens, septal nuclei, anterior hypothalamus, nucleus basalis of Meynert, prefrontal cortex) – the nucleus basalis of Meynert is found in this area and also sends fibers to various regions including the hippocampus and cerebral cortex. It is a major source for Ach and is considered important for normal memory fxns.
    • Functions are disrupted with ACoA strokes and Alzheimer’s disease. Common cause of damage producing memory deficit: AD
  • Thalamus (dorsomedial nuclei and anterior nuclei) – associate with encoding and integration of new information. Common cause of damage: stroke
  • Cerebral cortex – e.g., in patients w/Korsakoff’s damage is not restricted to the brain regions typically associated w/short-term memory. Rather, the damage is widespread and thought to involve large regions of the cerebral cortex as well (patients w/Korsakoff’s tend to show cortical atrophy – seen as enlarged ventricles and widening of the sulci).
    • Specific cortical damage is seen in:
      • Prefrontal cortex – associated w/ emotional apathy and perseveration; also, attention, organization (e.g., self-generated encoding strategies), and time-tagging. Common cause of damage producing memory deficit: strokes, TBI
      • Bilateral temporal cortical atrophy- difficulty forming visual associations
      • Temporal cortex – lesions of anterior temporal lobe (which spare the hippocampus) result in a variety of impairments (but not global amnesia as seen in H.M.) (hemispheric specialization is same as seen in hippocampus lateralization) – damage is similar to unilateral removal of hippocampus. Lateral regions associated with immediate and short-term recall. Ultimately the “temporal lobe” (which includes anterior temporal neocortex, amygdala, hippocampus, entorhinal cortex) is responsible for amnesia, plus the affective component of memory
  • Temporal Lobe Functions:
  1. auditory sensation and auditory/visual perception
  2. long-term storage of sensory input
  3. affective tone to sensory input
  • Symptoms of Lesions:
  1. impaired auditory sensation & perception
  2. decreased selective attention of auditory & visual input
  3. decreased visual perception
  4. decreased organization and categorization of verbal material
  5. decreased language comprehension
  6. decreased LTM
  7. personality and affective behavioral changes
  8. sexual behavior
  • Neurotransmitters – both the cholinergic system (Ach) and the catecholamines (DA, NE, and epinephrine) have been associated w/memory
    • Ach has a major source in the nucleus basalis of Meynert which communicates directly w/the hippocampus and the cerebral cortex; deficiencies in Ach have been associated with Alz. disease. Korsakoff’s patients also have a reduction of Ach and cell loss in the basal forebrain (particularly the nucleus basalis of Meynert)** catecholamines – damage in Korsakoff’s patients frequently involve the origin of the NE pathway; deficits have been associated w/reduced activity of DA and NE.

THE ROLE OF THE HIPPOCAMPUS (Scoville & Milner)

  • Bilateral Hippocampal Damage – results in a total lack of recall for day-to-day events; i.e., anterograde amnesia. It can also lead to partial retrograde amnesia, but overall, early memories are usually intact. There is no deterioration in intellect or personality
  • Day to day fxning and basic reasoning skills may appear normal until signs of memory impairment are noticed
  • These patients can retain a 3-figure number or a pair of unrelated words for several minutes if they are not distracted during the interval – but they will forget as soon as they are distracted
  • When smaller portions of the hippocampus is damaged (leaving some of it intact) memory impairments are not as severe; thus there appears to be a positive relationship b/t the extent of damage to the anterior hippocampal complex and the degree of memory impairment.
  • Damage to mammillary bodies – have also been related to amnestic states (in Korsakoff’s syndrome). Thus, there appears to be two interrelated structures, which can lead to amnesia if damaged.

THE ROLE OF THE FRONTAL LOBES (Shimamura et al.)

The role of frontal lobes in memory has been a controversial issue

  • some believe that memory functions per se are not impaired, but any memory difficulties are secondary to on-line processes (e.g., encoding, attention, and disinhibition)
  • others think frontal lobes are involved in (1) conditioned associative learning and (2) memory for temporal order
  • damage can result in impairments of STM for tasks w/c require temporal memory for stimuli (e.g., delayed matching to stimuli)

Cognitive Results (Fluency and Problems Solving)

Initiation and Perseveration Index – From Mattis’ Dementia Rating Scale This score reflects difficulties in organizing and searching information in semantic memory (similar to word fluency)

  • frontal patients impaired on Initiation and Perseveration Index of Dementia Rating Scale – if damage to left or bilateral damage (right frontal lobes unimpaired; but right frontal lobes have been implicated in nonverbal fluency tests)
  • amnesics impaired on Memory Scale of DRS
  • Korsakoff’s impaired on both (i.e., they have cognitive deficits and memory impairments) – thus they have frontal and diencephalic damage (confirmed by a radiological study)

Short-term Memory (e.g., Digit Span)

  • frontal patients – impaired
  • “amnesic patients” – intact

New Learning Ability

People with just frontal impairment typically do not display significant impairment in learning, i.e. no anterograde amnesia

Semantic Encoding and Release from Proactive Interference Theory – that one aspect of frontal lobe dysfunction is a failure to encode information semantically (developed due to responses on a test of “the release from proactive interference.”

  • release from proactive interference – subjects are presented 3 words and then asked to recall them following a 15 – 20 second distraction task. Trials 1 – 4: words from same semantic category are presented. (normals show a decline in performance across the four trials due to proactive interference). Trial 5, words are from a different category (normals perform as well as they did on trial 1)
  • Korsakoff patients – do not show this release (do poor on trial 5) presumably due to inadequate encoding in a semantic manner (so don’t benefit from a shift in semantic categories); attributed to frontal lobe damage (possibly secondary to degeneration of connections from the dorsomedial nucleus)
  • frontal patients – performance similar to normals. Thus does not support the theory, rather it suggests that both amnesia and frontal lobe dysfxn is necessary for the deficits seen in Korsakoff patients

Metamemory Metamemory is defined as knowledge about one’s own memory capabilities and knowledge about strategies that can aid memory by helping to plan, monitor and organize appropriate memory strategies.

  • components of metamemory – (include some fxns of frontal lobes) – planning, monitoring and organizing motor and cognitive fxns. Some believe that it is instead directly related to memory ability. e.g., the feeling of knowing (the subjective belief that the patient will know if they recognize an answer)
    • Korsakoff and frontal patients had deficits in feeling-of-knowing accuracy – BUT amnestics did not
    • thus amnestic patients can accurately express knowledge about what they do and do not know, despite having severe impairment in the ability to learn new information; and it is not a fxn of memory per se
    • the deficits seen in Korsakoff and frontal patients were not identical. For example, frontal patients did not have problems when performance was assessed after a short (5 minute) delay. Perhaps Korsakoff’s is worse b/c of deficits in memory and metamemory deficits.
  • thus, frontal lobes are critically involved in manipulation and organization of information (including information already in storage), despite not being critically involved in establishing new information in memory

Memory for the Temporal Order of Events

  • frontal patients – are impaired on recency memory (i.e., which item was presented most recently), but are not impaired on tests of item recognition memory — thus temporal order can be impaired even if their memory for the items is good
  • Korsakoff patients – are even more impaired on recency memory

Disorders of Source Memory

Source Memory refers to the process of remembering where the information was learned (contextual memory)

  • frontal patients – source memory was significantly impaired despite good recall for the facts themselves (suggests that frontal lobes contribute specifically to contextual memory)
  • amnestics can also be severely impaired in this (but not necessarily a requirement)
  • can also occur with normal aging; re: neuronal cell loss associated with aging occurs prominently in the frontal lobes, therefore source memory in normal aging may also be related to subtle frontal lobe dysfunction

Prospective Memory

  • Prospective memory refers to the strategies by which one remembers to perform and monitor future actions; could also include strategies involved in planning, monitoring and organizing memory, not just actions.
  • Theory – So, includes self-initiated searches and retrieval of information in memory. Re: in many ways memory searches are like problem solving tasks that require fluency, initiation and flexible thinking. Furthermore, impairment in inhibitory control could also reduce ability to discriminate b/t appropriate and useful strategies from inappropriate ones. This deficit would particularly impair free recall and temporal order tasks
  • Therefore, as expected frontal patients have difficult with this
  • Re: a deficit in prospective memory does not necessarily imply a deficit in declarative memory, and vice versa (e.g., amnesics do well on many tests of prospective memory)

Summary: The Memory and Cognitive Disorders in Frontal Patients

  • Memory Disorders
    • short-term memory (e.g., digit span)
    • free recall
    • metamemory
    • memory for temporal order
    • source memory
    • prospective memory
    • new learning not significantly impaired except on sensitive tests of free recall
  • Cognitive Disorders (Dysexecutive dysfunction)
    • planning
    • problem solving
    • initiation
    • perseveration
    • fluency
    • disinhibition

Notes from Kolb & Wishaw: Frontal patients show the following deficits:

  • failure to show release from proactive interference
  • errors of sequence when completing actions; poor memory for serial order
  • errors of intrusion and omission when asked to copy a series of facial movements
  • difficulty w/short-term memory

PATHWAYS OF SENSORY STIMULI AND MEMORIES (Mishkin & Appenzeller)

Theory: The hippocampal system interacts with perceptual pathways to transform sensory stimuli into memories

  1. Info is initially processed through the sensory cortex.
  2. From there, there are 2 distinct memory circuits involved in visual recognition:
    1. Hippocampus – specializes in memory for spatial location
    2. Amygdala – concerned with emotional significance of the material; is also responsible for combining memories obtained through different sensory modalities
      1. These systems work in a circuit (NOT totally independent) shown by the fact that cutting the connections between the two structures produced the same memory impairments seen when the structures themselves were damaged
        1. If either hippocampus or amygdala is destroyed, memories can still be stored because of parallel systems
  3. Diencephalon
  4. Prefrontal Cortex
    1. Hippocampus, amygdala, diencephalon (mammillary bodies), and prefrontal cortex all have connections with basal forebrain which is a cholinergic source
    2. Ach is hypothesized to be involved with the permanent structural (at the cellular level) changes that are responsible for long term storage
  5. Basal forebrain, via Ach, feeds back to the sensory cortex to complete the loop
  6. When same stimulus is again received by the sensory cortex, the “memory” is activated

Procedural Memory (i.e., “habits”)

  • Considered a separate system of learning, independent of the limbic system
  • it is noncognitive; NOT founded on knowledge or on memories, but on automatic conxns b/t a stimulus and a response
  • neural substrate – striatum (in the forebrain) – w/c receives extensive conxns from sensory systems and sends fibers to cerebral structures that communicate to the premotor cortex
    • The actual structural change responsible for LTM is described as, initially, a cell assembly that reverberates (STM). Structural changes and consolidation are a result of reverberation, which are then responsible for LTM.
  • damage to striatum has been show to impair habit formation

Criticisms of Mishkin’s Theory:

  • Does not address retrograde amnesia
  • Work was based on a very few simple experiments (i.e., delayed nonmatching to sample with monkeys)
  • Later work by Squire and Zola-Morgan suggest that it is the entorhinal cortex that is actually necessary for LTM, not the amygdala or hippocampus

Diencephalon (thalamus and hypothalamus)

  • damage can occur from Korsakoff’s syndrome, strokes, injuries, infections and tumors.
  • receive fibers running from the hippocampus and the amygdala (so directly connected to the memory systems)

The Amygdala

  • has connections with all sensory systems and communicates with the thalamus
  • also, the same parts of the amygdala on w/c sensory inputs converge to send fibers into the hypothalamus (the source of emotional responses) – thus, the amygdala links sensory experiences and emotions
    • animals w/ damage to amygdala lose fear of humans and even aversion to sensations such as pinching; they also have trouble remembering the positive associations of a familiar stimulus
  • memories are not sensory specific (e.g., sound of a voice will summon a memory of the person’s face; or a smell of a food will summon memories of its appearance, texture and taste); the amygdala mediates the association of memories formed through different senses

A COMPONENT PROCESS MODEL OF MEMORY (Moscovitch)

Memory is not unitary, but depends on the operation of potentially independent, but typically interactive, components; he argues that memory (like perception) consists of the operation of modules and central systems.

FOUR ESSENTIAL COMPONENTS

Each component aligns with different types of memory tests

  1. Nonfrontal neocortical component – made up of various perceptual and “semantic” modules; w/c mediate performance on item-specific, implicit tests of memory
  2. Basal-ganglia component – mediates performance on sensorimotor procedural tests of memory
  3. Medial-temporal/hippocampal component – mediates encoding, storage, and retrieval on explicit episodic memory tests (that are associative/cue dependent)
  4. Central-system frontal-lobe component – “works” with memory and mediates performance on strategic explicit and rule-based explicit tests

MODULES share several features:

  • Domain specificity (so damage to a specific area leads to impairment in purported domain)
  • information encapsulation – processes subserved by one module are not affected by deficits in other areas.
    • these two aspects support the criteria for “double dissociation”; i.e., each module is unique for a particular task
  • shallow output – output from a module is specific to the domain it covers; it has no meaning beyond the value assigned to it by the module. e.g., patients with associative agnosia can process faces visually but they cannot assign any meaning to the visual information computed

Thus, a module is essentially a stupid, closed computational device that delivers its shallow output to the interpretive central systems where meaning and relevance are assigned

CENTRAL SYSTEMS – unlike modules in that they:

  • integrate information from superficially dissimilar domains
  • are open to top-down influence
  • output is deep or meaningful
  • interlevel representations may be available to consciousness

CLASSIFICATION OF MEMORY TESTS

According to Moscovitch, there are 2 classes of memory:

  • explicit (associative and strategic types)
  • implicit (procedural and item-specific types)
  1. Associative tests – test in which the cue is sufficient for retrieval (e.g., have you read the bible)?
  2. Strategic Tests – the cue does not automatically elicit the target memory, but only provides the starting point of a memory search (e.g., what did you do 2 weekends ago?)
  3. Procedural Tests – assess learning and retention of sensory-motor skills, procedures or rules.
  4. Item-specific tests – assess memory for a particular item, such as a certain word, face or object, by looking at the accuracy or speed of identification of the item when it is repeated

PROCEDURAL IMPLICIT TESTS

  • sensorimotor skills – should be retained in normal amnesic and demented patients with intact sensorimotor structures, but will be impaired if they have degenerative disorders associated with damage to the basal ganglia – part of extrapyramidal motor system (e.g., Huntington’s or Parkinson’s patients). But these same patients would perform normally on perceptual, item-specific implicit tests

ITEM-SPECIFIC IMPLICIT TESTS

Registration/recording (in this model) refers to the neocortical process involved in the forming of perceptual records or engrams necessary for implicit tests. (Thus, this differs from consolidation w/c involves long-term memory traces)

Perceptual Input Modules and Perceptual Repetition Priming (a module)

  • consistent with properties of modules discussed above, perceptual input modules restrict their information to a specific domain, are not affected by higher-order semantic information and the information is restricted to presemantic structural descriptions. (e.g., face-recognition system)
  • anatomical localization – posterior neocortex are sources for mediation of repetition-priming effects (there is both negative and evidence implicating these structures)
    • negative evidence – repetition-priming effects are well-preserved in amnestics (w/damage to medial temporal lobes/limbic structures)
    • positive evidence – domain specific agnosias and PET scans in normal people have supported this premise
  • exception: damage to the critical region for face recognition leads to impairments on both explicit and implicit tests of knowledge (they show no repetition-priming effects for faces). On the other hand, some patients (whose damage spares the crucial region) can respond differently to familiar and unfamiliar faces on implicit tests and show normal repetition-priming effects

Conceptual Repetition Effects and Semantic Records (a central system)

  • tasks differ from perceptual tests, b/c the target is not repeated at test (even in degraded form), rather it is elicited by semantic cues (e.g., a related word or a question).
  • influenced by semantic variables (so unlikely to be mediated by presemantic input modules) and are likely mediated by central systems which interpret the shallow output of perceptual modules and store a semantic record of their activity or representations
  • b/c they are mediated by interpretive central-system structures, conceptual, repetition-priming effects are reduced or absent in Alz. patients (despite intact perceptual priming). This pattern has also been seen in patients w/unilateral temporal lobectomies (BUT – damage to hippocampus leaves this process intact – so must be due to the cortex itself, not underlying subcortical structures)

The Hippocampal Component (a module for episodic, associative memory)

  • this component consists of a number of structures in the medial temporal lobes and diencephalon that form a circuit. These include (in addition to the hippocampus):
    • parahippocampal gyrus
    • entorhinal and perirhinal cortices
    • mammilary bodies
    • dorsomedial nucleus of the thalamus
    • cingulate cortex
    • fornix
    • amnesia is associated with bilateral damage to any one of these (except the fornix and cingulate where the evidence is unclear)
  • the domain of this module is consciously apprehended information (if an event does not receive full conscious attention, it is not processed by the hippocampal component)
  • b/c we don’t always know in advance what is worth committing to memory, the hippocampus is capable of encoding and storing information automatically as part of the process of consciously observing stimuli
  • the central idea of levels of processing theory (Craik & Lockhart) – that remembering is a natural by-product of cognition – follows directly from this view that the hippocampal component is modular. Thus what determines what will be remembered is not the intention to remember, but the extent to which events are attended and information is processed to a deep level and properly organized Paying close attention to the target and encoding it semantically makes it distinctive and makes its memory traces more easily retrievable.
  • the benefit of an automatic hippocampal component is that it does not draw cognitive resources away from other activities
  • disad of being modular – output is shallow; it is not interpreted properly in relation to other memories (i.e. spatial and temporal context or organizational memories – probably mediated by the frontal lobes)
  • this theory suggests, therefore, that a component of conscious recollection is no more intelligent or under voluntary control than perception.
  • damage to right hippocampal region (only) – associated with spatial memory loss
  • damage to left hippocampal region (only) – associated with loss of memory for verbal material

The Frontal Lobes (central systems and strategic explicit tests)

  • frontal lobes are necessary to convert remembering from a stupid reflexive act triggered by a cue (hippocampal module) to an intelligent, reflective, goal-directed activity under voluntary control
  • memory disorders of frontal lobes are not related to deficits in storage and retention (the hippocampal fxns), but with impaired organizational and strategic processes
  • deficits lead to problems with strategic explicit (and maybe some implicit) tasks. The tests are performed poorly not b/c the target event is forgotten, but b/c organization at encoding and strategic search and monitoring at retrieval id deficient

EXPLICIT & IMPLICIT MEMORY IN THE ELDERLY:

Evidence for double dissociation between medial temporal and frontal lobes (Winocur, Moscovitch, Stuss)

  • Not all aspects of memory fxn decline at the same rate during the aging process. This paper shows a fxnl double dissociation in old people b/t performance on implicit and explicit tests and performance on tests sensitive to different brain regions
  • explicit memories (which rely on recall or recognition techniques) – do not show a strong age difference (re: damage to hippocampus causes this kind of impairment)
  • implicit (or indirect) memory – is also generally preserved in the elderly, but there are some notable exceptions (e.g., word completion tasks after priming – as this paper shows)
  • implicit memory tests – measures performance that does not involve conscious awareness of any part of the prior experience (e.g., tests of general knowledge, skill learning and repetition priming in w/c exposure to a word biases subsequent identification in a word completion test)

Word Completion Tasks:

  • word-fragment completion – presents the complete word and then at test the subject is given a sample of the word’s letters (e.g., soldier; _o_ _i_ r)

– does not typically show an age difference

  • word-stem completion – still presented word at study period, but then at testing the first 3 letters of the word are presented (e.g., sol_ _ _ _)

– has been shown that elderly have less priming effects than younger subjects

  • since these two tests show different results in the elderly, it is believed that each task is mediated by different neural structures; that’s what this paper shows

Results:

  • age differences occur in word-stem completion but not in word-fragment completion tasks and the weakness in word-stem completion is correlated with tests of frontal lobe fxn
  • the dissociation of implicit memory tasks and explicit memory is not a new finding (previously seen in amnesics, and normals), but the finding that this correlates with frontal fxning is new; since frontal fxning was not involved in word-fragment completion, it suggests that 2 different processes underlie the 2 types of priming
  • therefore, they suggest that automatic and strategic processes can both contribute to conscious recollection and implicit memory
    • traditional belief is that automatic processes are associated with implicit memory and strategic processes with conscious recollection
    • if the cues on the implicit tests are inadequate and strategic search processes are necessary then the frontal lobes will b/c involved (and take over the automatic process led by the medial temporal lobes)
    • so, they suggests that word-completion tasks involve some strategic process (and, therefore, the frontal lobes), but that word-fragment tests do not because unlike initial stem letters, letters drawn randomly from the word are poor cues for locating words in the lexicon (believed to be mediated by posterior cortex)
  • this study also confirms previous studies that institutionalization exacerbates the process of cognitive decline associated with aging
  • the double dissociation of decline in different memory systems in the aging suggests that declines in performance with age can’t be attributed to a uniform effect across all fxns, nor to a selective effect on those most vulnerable in old age
    • in sum: age differences in word stem completion can be attributed to frontal lobes, but not medial temporal dysfxn, whereas the reverse is true of explicit tests of memory