Long-Term Memory

Majority notes taken from Cognitive Psychology: A Student’s Handbook by Eysenck & Keane (click image), which I recommend for its readability and theory evaluation.




Memory is the process in which information is encoded, stored, and retrieved

Memory Structure Tree Diagram
Memory Structure Tree Diagram



(this section is about disagreement between psychologists in grouping/dividing memory – it’s optional, but it’s the kind of discussion that will get you an uber-distinction!)


– Disagreement about subsystems of human memory

– e.g. Squire (2007) vs Schacter & Tulving’s (2000) memory models

– i.e. Schacter & Tulving split declarative into semantic & episodic but Squire has a single declarative system



processing approach: emphasises psychological processing

systems approach: emphasizes systems or structures, usu. with distinct brain locations


Roediger et al (1999) – criticised multiple systems theories for not specifying rules of how to distinguish a specific system

Schacter et al (2000) – proposed that distinct memory systems are defined by:

  1. class inclusion operations (e.g. general knowledge)
  2. properties and relations (e.g. how it works and what’s it for)
  3. convergent dissociations (e.g. amnesics can’t do X task but controls can)

Roediger et al (1999) – this is still inadequate


(in an essay, however, I would go on to say that despite all this, there is some overlap/agreement between theorists, so we can discuss non-declarative, declarative, episodic, semantic, procedural, implicit, explicit and evidence for each)


See my long-term memory essay plan: here




Retrograde amnesia: problems remembering events before the onset of amnesia

Anterograde amnesia: (e.g. HM) problems remembering new info learned after amnesia

Parkin (2003) defines amnesia as:

  • Unimpaired STM storage (e.g. digit span)
  • Severe and permanent anterograde amnesia
  • Intact semantic memory (but debatable)
  • Intact skill learning (procedural memory)
  • Retrograde amnesia which varies between patients depends on cause/damage

(from this you can see how amnesia patients help us divide up different components of memory)



– found that words at the beginning (primary) and end (recency) of lists were best recalled

– suggested words early in list went into LTM and end of list STM i.e. 2 diff stores



patient ‘HM’ (Wickelgren 1968) – removal of medial temporal lobe inc. hippocampus to cure epilepsy

– problems laying down certain types of LTM; good STM

KF (Shallice & Warrington, 1970)patient suffered head injury

– bad short-term memory but LTM was OK

These & other cases strongest argument for a separation at some level bet. STM &LTM




Ryle (1949) made a philosophical distinction adopted by psychologists:

Knowing “how” to do something, e.g. riding a bicycle (procedural skill-based memory)

Knowing “what” about something, e.g. what a bicycle is (declarative fact-based memory)

This has been adopted by psychologists



– memory for facts, operating if asked ‘what is the capital of England?

– memories that can be declared or described; needs conscious recall of previous exp.

– associated with dorsolateral prefrontal cortex



semantic memory: memory for facts

– general knowledge about the world, concepts, social norms, language, taste, colour


episodic memory: memory for events

– personal experiences happening at a given place & time inc. thoughts & feelings

– neuroimaging indicates parietal lobes (Simons et al 2008) BUT patients with damage here often don’t have great episodic problems



HM – bad declarative memory for events after trauma; fair non-declarative memory

e.g. Milner (1968) mirror drawing learning, with no declarative details of the test sessions

Schott et al (2006) – diff. parts of the brain activated for learning during dec & non-dec

Schott et al (2005) – diff areas associated with mem retrieval for dec & non-dec tasks

Aggleton & Brown (1999) – ‘extended hippocampal system’ for episodic memory



Reder et al (2009) – binding (p.279-81)

  • whether tasks are impaired depend on whether binding item & context is required
  • when taken into account amnesiacs can show failing non-decl and intact decl




Tulving (1972) made distinction within declarative memory episodic & semantic

– suggested amnesia was impairment in adding to episodic memory leaving semantic memory intact



Spiers et al (2001) – 147 cases anterograde amnesia: episodic impairment in all cases; modest semantic problems

Kapur (1999) – retrograde amnesia study review: some patients more episodic loss than semantic; others the opposite

Prince et al (2007) – interacting systems

  • neuroimaging: e.g. left hippocampus episodic encoding
  • some PFC regions associated with both episodic & semantic retrieval & encoding

i.e. systems combine e.g. episodic day out includes semantic food, bus etc



Eysenck & Keane(2010) – amnesic patients may do worse on episodic tasks because there’s less time for learning these tasks


Gabrieli, Cohen & Corkin (1988) – is amnesic semantic memory intact?

– If amnesic semantic intact, then amnesic e.g. HM should be able to add to his vocabulary

– HM couldn’t learn definition of new words that became current after his surgery

– so if amnesic semantic memory is also impaired the strict split is questionable


Hoscheidt et al (2010) – fMRI

– hippocampus involved in retrieval of both semantic & episodic spatial info

– sharp distinctions bet. semantic & episodic systems too simplistic


Greenberg & Verfaellie (2010) – review of studies, with compromise position

– studies show semantic & episodic influence each other in encoding & retrieval

– theories should allow for interdependencies between the 2 types of memory

– clear cut distinction too simplistic



Spiers (2001) research (above) is compelling

Vargha-Khadem et al (1997) – kids

  • kids with early amnesia due to bilateral hippocampal damage
  • poor episodic but normal semantic as they grew up


Verfaillie et al (2000) – PS

  • unique specific bilateral hippocampal damage
  • poor episodic but pretty normal semantic: semantic relies on other areas

– could be Gabrieli discrepancy occurred because of damage to other close areas

– kids studies show that semantic memory can be intact without episodic



– evidenced by changes in behaviour (e.g. learning to ride a bike)

– doesn’t involve conscious recollection (e.g. person couldn’t say what they learned)


priming: influencing response to target by showing a related stimulus beforehand

procedural memory: knowing how and the ability to perform skilled actions




– a subcategory of non-declarative memory

Spiers et al (2001) – 147 amnesia cases: procedural intact (declarative impaired)

Corkin (1968) – HM could learn mirror drawing & target tracking

Cavaco et al (2004) – real world type tasks (e.g. weaving, machine): amnesics comparable to controls (decl impaired)

Siegert et al (2006 ) – assoc with striatum: Parkinsons (damage to striatum) poor procedural

Debaere et al (2004) – also cerebellum activated in neuroimaging with learning skill task



Brown & Robertson (2007) – declarative & procedural interactions: not separate

  • showed interference between procedural and declarative learning
  • challenge concept of fixed independent memory systems: dynamic interaction




Schacter (1987) uses concept to differentiate between different types of memory testing:

Explicit memory: tests where subject recalls a specific learning episode e.g. free recall

Implicit memory: tests not referencing a previous event and so accesses event indirectly.

Explicit memory processing is conscious; Implicit memory processing automatic/unconscious


Verfaellie et al (1991)

Studied implicit and explicit memory in amnesic patient ‘TR’

– using galvanic skin response in implicit tasks showed TR’s implicit memory intact

– explicit memory poor (as expected)


Gabrielli et al (1995) – patient ‘MS’ (right occipital lobe removed for epilepsy)

– amnesics usu. have poor explicit memory (free recall); intact implicit (repetition priming)

– MS had good explicit but poor implicit memory

– concludes separate systems are behind these 2 types of memory (ROcortex)




Summary of Condition

– showed a severe anterograde amnesia

– resected H.M.’s medial temporal lobe structures for intractable epilepsy

– Scoville and Milner believed lesion of the hippocampus caused H.M.’s amnesia

– also that severity of amnesia relates to size of hippocampal removal

– S&M: was H.M.’s anterograde amnesia global & severely impaired on all tests? Yes.

– regardless of stimulus material (words, digits, paragraphs, pseudowords, faces, shapes, clicks, tones, tunes, sounds, mazes, public events, personal events)

– regardless of the sensory modality (vision, audition, somatosensory system, olfaction)


Specifics of condition

– can’t acquire episodic & semantic (inc. new word meanings) knowledge

– MTL structures removed crucial for long-term declarative memory (facts & events)



– H.M.’s short-term memory is OK [& so doesn’t depend on these MTL structures]

– so he can encode new information, but can’t use this info to make LTM trace

– H.M.’s memory and language capacities are dissociable.

– Milner et al (1968) – his comprehension of language is undisturbed

– motor learning showed dissociation bet. declarative memory & other kinds of learning

– Milner: his time & error scores decreased across 3 days training on mirror-tracing task


Episodic & Semantic Memory

– sometimes has meagre conscious recollections of information encountered postoperatively

– poor episodic recall of events; mostly semantic (factual)

– debate over which type of memory should be most affected: episodic more or both equally

– shows some semantic learning for famous people since operation (esp. given prompts)


Repetition Priming

– Repetition priming occurs incidentally when stimuli encountered in a study list influence performance on a subsequent test, without conscious awareness of that influence

– e.g. deciding whether the word ‘episode’ in a study list is something you can or cannot touch will increase the likelihood that in a subsequent test, you will complete the stem ‘epi-’ as ‘episode’ (the primed word) rather than as ‘epic’,‘epicure’ or ‘epilepsy’ (unprimed words)

– HM failed repetition priming for words coming into use after 1965 (12 years after surgery)

– he primed normally with premorbid words (but could not recall or recognize them explicitly)

– after 1953 couldn’t acquire new vocabulary words: [lacked the representations to support lexical retrieval procedures]


What learning remains?

Other preserved learning is believed to be independent of the MTL memory system:

–  perceptual after-effects, prism adaptation, perceptual learning & most repetition priming

– HM could draw floor plan of house where he stayed after he had amnesia

– topographical memory intact; some personal semantic memory – i.e. some declarative mem

– but he is impaired on spatial memory tests

– Corkin: maybe HM didn’t get time to learn in lab tests vs. walking through house for years

– atypical declarative mem: could recognise studied complex coloured mag pics up to 6mths


Another dissociation?

– some argued these 2 kinds of recognition are 2 diff processes, supported by different neural substrates.

– one hypothesis is that conscious recollection of the learning episode depends on the hippocampus, whereas familiarity judgements without episodic content rely on perirhinal cortex

– some neuro evidence: amnesic patient with extensive damage to MTL structures (including perirhinal cortex, amygdala, hippocampal formation and parahippocampal cortex) doesn’t show the benefit of extended study of pics