From last year you should know:
- Classic models and types of memory
- How the evidence supports these models
- The methods used to test models of memory
Recap of 101/106:
Sperling - sensory memories are very brief memory that decay really quickly and are input into memory from the sense e.g. touch, sight
If sensory information is attended to it moves to short term memory
Short term memories are fairly brief memories kept in mind for a short period
If short term memories are rehearsed in working memory then they can be passed into long term memory - this is what this lecture focuses on
From PSY233 you should understand and be able to explain the following concepts
- Describe the Theory of long-term memory (Tulving, 1985) and Paradigms used to test this Theory
- Provide an account of the Predictions that are used to test the Theory (Tulving, 1985)
- Evaluate the evidence that supports Tulving’s (1985) model of long-term memory
- Describe memory consolidation as conceptualised by the two-step model (McClelland et al., 1995)
- Describe memory consolidation as conceptualised by the System Consolidation Hypothesis (Born & Wilhelm, 2012)
- Evaluate the evidence that supports the role of sleep in systems consolidation (Born & Wilhelm, 2012)
Preface to the lecture:
Representationalism is the idea that what we are doing in studies of memory is trying to explore mental representations existing in our mind, that we can’t see, that may be conscious or unconscious
The word “CAP” has a mental representation in semantic memory that is activated when the word is heard, becomes available to your conscious awareness, so you can use it. Can be drawn into an arena where you can use it.
Representations are drawn into working memory and rehearsed in order to keep them in your working memory
Representation may be stored in semantic memory – things like words and knowledge
Tulving 1985 said?
LO1
Tulving said there were 3 memory systems in long term memory…
- Procedural
Said this was the main memory system, the biggest circle with the others sitting inside it. Memory of how to do things, unconscious/implicit/non-declarative - Semantic
Within procedural memory, in theory sem. cannot operate independently of procedural memory (but procedural memory can operate independently). Things that we KNOW, extracted from procedural memories. Declarative - Episodic
Within semantic memory. Memory of episodes (the what, why, where, who etc), said to contain contextual information. Things we REMEMBER, extracted from semantic memories. Declarative
Does evidence from Clive Wearing support the idea that semantic memory can operate independently of episodic memory?
Clive cannot form new long-term memories BUT Clive can remember some things that would be stored in semantic (and procedural) memory such as language and piano
Episodic OR Semantic memory
Knowing vs Remembering: The distinction
LATE vs LAITE
Do you REMEMBER when you first saw this word, or do you KNOW it but not remember when you saw it?
Of course you remember the word late, but Tulving used the word “remember” to indicate a memory of the event of learning information, whereas if you “know” something you can recall it but not when you learnt it.
The Theory
Declarative memory can be divided into semantic and episodic memory stores (Tulving, 1972)
What evidence is used to support this theory?
Paradigms and Predictions
(LO1+2)
*** LESIONS
Paradigm:
Lesion - People with damage to their brains that results in a loss of episodic memory
Prediction:
People with brain lesions will be able to encode information into semantic memory without forming an episodic memory
Findings:
Vargha-Khadem et al., (1997) brain injuries at ages from birth to 9 years old that affected the hippocampus resulted in severe episodic memory loss but language, literacy and factual knowledge acquired at school
***FMRI
Paradigm:
fMRI - fMRI or other imaging techniques can show a difference in brain activation when people process information from semantic or episodic memory
Prediction:
Scans will show different patterns of activation when participants recall information from semantic or episodic memory
Findings:
Scans will show different patterns of activation when participants recall information from semantic or episodic memory (Greenberg & Verfaellie, 2010)
***COGNITIVE
Paradigm:
Cognitive paradigms - Paradigms that test semantic and episodic information e.g. interference from semantic knowledge on an episodic memory test
Prediction:
Semantic information will interfere with episodic recall/ recognition (but not vice versa)
Findings:
Kan et al., (2009) congruency effect observed - participants correctly ‘remember’ items that present the ‘true’ cost compared to items where the cost is ‘false’
The Theory
Declarative memory can be divided into semantic and episodic memory stores (Tulving, 1972)
What evidence is used to support this theory?
The Experiment
Kan et al 2009 in more detail
Kan et al 2009
The Experiment:
In this task ppts are presented with a list of items and their prices. Asked to state whether the price is ‘true’ or ‘false’.
Test Phase:
In this task ppts should select the option that corresponds to the price they were presented with initially NOT the actual price of the item.
The Results:
- Patients with hippocampal damage performed worse overall due to impaired episodic memory and semantic knowledge does not impact on episodic recall
- Semantic knowledge impacts on episodic recall for healthy controls
- Semantic knowledge intact for both groups
Conclusions:
- If there was just one kind of memory store then semantic memory would be impaired as well as episodic memory.
- Episodic recall would be the same regardless of whether the items were congruent or not
- The existing schema in semantic memory has been activated, supports recognition of new information stored in episodic stores
What evidence is used to support Tulving’s theory?
Double dissociation
This kind of lesion study is called a double dissociation (lesions are already present, brain function is assessed using fMRI scans) - when two related mental processes are shown to function independently of each other
Imaging evidence of different parts of the brain being active when retrieving information from ‘semantic’ or ‘episodic’ information
Cognitive paradigms here test differences in recall/recognition for information stored in semantic and episodic memory (tests alongside scans).
What is memory consolidation?
Some theorists believe you must store info in episodic memory before consolidating it into long term semantic memory e.g. the word cap, you know it but can’t recall when you first heard it. Over time you form lots of episodic memories for the word cap, including info on the where, who, smells, environments etc. Gradually the core information from these memories is extracted, and just the word cap is stored in LTM sematically, so you’ve turned an episodic memory into long term semantic memory. This is consolidation.
- Rapidly encoded information in episodic memory is bound with existing information in Long Term Memory
- Links between memory systems managed by hippocampal systems
- Long Term Memory (LTM) stores in neocortex
- Abstraction of info into LTM = consolidation
Theories of declarative memory consolidation…
LO4 - Two-step computer model
e.g. McClelland et al (1995)
Computational Models
Two- stage model of memory consolidation.
- Events are stored in the hippocampal systems and transferred to neocortex over time
Said: We initially encode information really quickly in the hippocampal system, have a representation in semantic LTM in the cortex, initially there are lots of links between the neocortex and the hippocampal systems, but over time these links begin to die out as the core information becomes more extracted and eventually we are left with the core info without the contextual information held in mind.
McClelland et al said we had to have this two step model because we want to ignore something called catastrophic interference e.g.
- Penguins are birds
- Can move, grow, swim
- Therefore all birds can swim and all things that swim are birds XXX
Without the degredation of the links between hipp. and neoxortex the contextual information attached to core information would begin to interfere with new information and incorrect assumptions are formed.
Theories of declarative memory consolidation…
LO5 - System Consolidation Hypothesis
Theory
Born and Wilhelm, 2012
Whilst you are asleep, there is a sort of replay of learned information in the brain (neuronal replay), neurons that are active during the learning of new information are reactivated in the same patterns but very quickly during sleep.
Shown in studies with rodents, during waking hours the animal learns a maze, neurons reactivate during sleep.
Theories of declarative memory consolidation…
LO6 - System Consolidation Hypothesis
Study 1 - Atienza & Cantero, 2008
Asked people to look at pictures and classify them as neutral, positive or negative and the intensity of the feeling e.g. elephant, watering can, train
Half the ppts were then sleep deprived after encoding
Test phase 1 week later - ppts asked whether they remembered (episodic) or knew (semantic) those pictures
Findings:
Those who were sleep deprived had poorer memory of the pictures, but this effect was specific to episodic memory
Conclusions:
- Semantic store not affected by lack of sleep as stable store of existing information
- Episodic memory affected by lack of sleep – ability to consolidate new information is impaired
- If information is not stored in Episodic memory it will not make it into to long term memory
Theories of declarative memory consolidation…
LO6 - System Consolidation Hypothesis
Study 2 - Rasch et al., (2007)
Evidence that during sleep it’s the reactivation of the memory that makes the difference in recall
Participants learned location of pairs of cards whilst exposed to the scent of roses. Exposure to the scent during sleep resulted in better memory for card locations after sleep compared to absence of exposure.
Importantly no effect was seen if the participants were exposed to the scent whilst awake.
Not everyone accepts these theories of consolidation…
Multiple Trace Theory (Nadel & Moscovitch, 1997)
Not everyone accepts these theories of consolidation…
Nadel and Moscovitch state that you don’t completely extract information from episodic memory/ hippocampus. This theory states that while we do still extract information from episodic to semantic memory that the links with the hippocampus remain intact.
- Hippocampal complex encodes all information that is attended
- The ‘facts’ of an episode are extracted & stored in the neocortex.
- BUT links with the hippocampus are maintained
Final Question…
If normal memory function requires information to be stored in episodic memory and then transferred to semantic memory how do people without episodic memories consolidate new information?
